Prevalence, Predictors and Prognosis of Depressive

Recent Publications in this Series
NIINA MARKKULA
5/2016 Liisa Korkalo
Hidden Hunger in Adolescent Mozambican Girls: Dietary Assessment, Micronutrient Status,
and Associations between Dietary Diversity and Selected Biomarkers
6/2016 Teija Ojala
Lactobacillus crispatus and Propionibacterium freudenreichii: A Genomic and Transcriptomic
View dissertationes scholae doctoralis ad sanitatem investigandam
7/2016 César Araujo
universitatis helsinkiensis 24/2016
Prostatic Acid Phosphatase as a Regulator of Endo/Exocytosis and Lysosomal Degradation
8/2016 Jens Verbeeren
Prevalence, Predictors and Prognosis of Depressive Disorders in the General Population

Regulation of the Minor Spliceosome through Alternative Splicing and Nuclear Retention of the
U11/U12-65K mRNA
9/2016 Xiang Zhao
HMGB1 (Amphoterin) and AMIGO1 in Brain Development NIINA MARKKULA
10/2016 Tarja Pääkkönen (Jokinen)
Benign Familial Juvenile Epilepsy in Lagotto Romagnolo Dogs
11/2016 Nora Hiivala Prevalence, Predictors and Prognosis of Depressive
Patient Safety Incidents, Their Contributing and Mitigating Factors in Dentistry
12/2016 Juho Heinonen
Disorders in the General Population
Intravenous Lipid Emulsion for Treatment of Local Anaesthetic and Tricyclic Antidepressant
Toxicity
13/2016 Riikka Jokinen
Genetic Studies of Tissue-Specific Mitochondrial DNA Segregation in Mammals
14/2016 Sanna Mäkelä
Activation of Innate Immune Responses by Toll-like Receptors and Influenza Viruses
15/2016 Mari Hirvinen
Immunological Boosting and Personalization of Oncolytic Virotherapies for Cancer Treatment
16/2016 Sofia Montalvão
Screening of Marine Natural Products and Their Synthetic Derivatives for Antimicrobial and
Antiproliferative Properties
17/2016 Mpindi John Patrick
Bioinformatic Tools for Analysis, Mining and Modelling Large-Scale Gene Expression and Drug
Testing Datasets
18/2016 Hilla Sumanen
Work Disability among Young Employees Changes over Time and Socioeconomic Differences
19/2016 Oyediran Olulana Akinrinade
Bioinformatic and Genomic Approaches to Study Cardiovascular Diseases
20/2016 Prasanna Sakha
Development of Microfluidic Applications to Study the Role of Kainate Receptors in
Synaptogenesis
21/2016 Neha Shrestha
Mesoporous Silicon Systems for Oral Protein/Peptide-Based Diabetes MENTAL HEALTH UNIT
Mellitus Therapy
22/2016 Tanja Holopainen
NATIONAL INSTITUTE FOR HEALTH AND WELFARE AND
Targeting Endothelial Tyrosine Kinase Pathways in Tumor Growth and DEPARTMENT OF PSYCHIATRY
Metastasis FACULTY OF MEDICINE
23/2016 Jussi Leppilahti DOCTORAL PROGRAMME IN CLINICAL RESEARCH
Variability of Gingival Crevicular Fluid Matrix Metalloproteinase -8 Levels
in Respect to Point-of-Care Diagnostics in Periodontal Diseases UNIVERSITY OF HELSINKI
24/2016
Helsinki 2016 ISSN 2342-3161 ISBN 978-951-51-2070-0

Department of Psychiatry
Faculty of Medicine
University of Helsinki
Mental Health Unit

National Institute for Health and Welfare
PREVALENCE, PREDICTORS AND

PROGNOSIS OF DEPRESSIVE DISORDERS IN
THE GENERAL POPULATION
A LONGITUDINAL POPULATION STUDY
Niina Markkula
ACADEMIC DISSERTATION
To be presented, with the permission of the Faculty of Medicine of

the University of Helsinki, for public examination in the Christian Sibelius auditorium,
Psychiatry Centre, on 3rd June 2016, at 12 noon.
Helsinki 2016
Supervised by Research Professor Jaana Suvisaari, MD, PhD
Helsinki, Finland
Docent Samuli Saarni, MD, PhD

Department of Psychiatry, University of Helsinki,
Helsinki, Finland

Helsinki, Finland
Turku University Hospital and University of Turku,

Turku, Finland
Reviewed by Docent Soili Lehto, MD, PhD

Institute of Clinical Medicine / Psychiatry and
Clinical Research Centre
University of Eastern Finland
Kuopio, Finland
Professor Jouko Miettunen, PhD

Center for Life Course Health Research
University of Oulu
Oulu, Finland
Opponent Professor emeritus Raimo K. Salokangas, MD, PhD

University of Turku
Turku, Finland
Cover image by Marcelino Martínez González
Dissertationes Scholae Doctoralis Ad Sanitatem Investigandam Universitatis

Helsinkiensis 24/2016
ISSN 2342-3161 (print)
ISSN 2342-317X (online)
Hansaprint
Helsinki 2016
To my family
ABSTRACT
Depressive disorders are a major public health concern worldwide due to

their pervasiveness, often chronic or recurrent course and serious adverse
outcomes. These include psychosocial disability, reduced quality of life,
physical health problems and increased mortality. This study examined
prevalence, predictors and different adverse outcomes of depressive
disorders (major depressive disorder and dysthymia) in a general population
setting. Specifically, the study aimed to establish the prevalence of depressive
disorders in Finland in 2011 and assess possible changes over the past
decade; to examine risk factors for new-onset depressive disorders; to
investigate the long-term prognosis of depressive disorders and its
determinants; and to assess excess mortality associated with depressive,
anxiety and alcohol use disorders. A large longitudinal study of the Finnish
population, consisting of the Health 2000 and Health 2011 Surveys, was used
to investigate these questions. The survey data was complemented with data
from the Care Register for Health Care and the Finnish Causes of Death
Statistics.
The results show that one in 10 adults in Finland suffered from a depressive
disorder in 2011, and the prevalence increased from 2000 to 2011,
particularly among women. Methods to account for non-participation also
showed that the non-participation of people with depressive disorders in
population studies significantly biases prevalence estimates. People who
were younger, had a history of multiple childhood adversities, lower trust
axis of social capital, or an anxiety disorder or subclinical depressive
symptoms at baseline, had a higher risk of developing depressive disorders.
In addition, having three or more physical diseases was a risk factor for
dysthymia. Among people with depressive disorders at baseline, 34-43% still
had some depressive, anxiety or alcohol use disorder after 11 years, and 48-
61% had clinically significant depressive symptoms. Unmarried people and
those with more severe initial symptoms had a higher risk of persistent
course. People with depressive disorders had a twofold mortality risk,
whereas the risk was 1.7-fold in alcohol use disorders and was not increased
in anxiety disorders, when adjusted for other risk factors.
This study confirms that depression is an increasing public health concern in

Finland, and that depressive disorders have serious long-term consequences.
To reduce the burden of depressive disorders, it is of key importance to
develop primary prevention efforts and measures to reduce the negative
health and social consequences of depression.
4
TIIVISTELMÄ
Masennushäiriöt ovat merkittävä kansanterveydellinen ongelma

maailmanlaajuisesti. Ne ovat yleisiä, usein pitkäkestoisia ja niillä on suuri
taipumus uusia. Niihin liittyy vakavia kielteisiä seurauksia, kuten
psykososiaalista haittaa, heikentynyttä elämänlaatua, fyysisen terveyden
ongelmia ja lisääntynyttä kuolleisuutta. Tässä tutkimuksessa selvitettiin
masennushäiriöiden (masennustilan sekä pitkäkestoisen masennuksen)
esiintyvyyttä, riskitekijöitä ja erilaisia kielteisiä seurauksia pitkittäisessä
väestötutkimuksessa. Tutkimuksen tavoitteina oli selvittää
masennushäiriöiden esiintyvyys Suomessa vuonna 2011 ja arvioida
esiintyvyyden mahdollisia muutoksia viime vuosikymmenen aikana, tutkia
masennushäiriöiden ilmaantuvuuden riskitekijöitä, selvittää
masennushäiriöiden pitkän aikavälin ennustetta ja siihen vaikuttavia
tekijöitä sekä arvioida masennus-, ahdistus- ja alkoholihäiriöihin liittyvää
ylikuolleisuutta. Tutkimusaineistona käytettiin Terveys 2000 - ja Terveys
2011 –tutkimuksia, jotka muodostavat yhdessä pitkittäisen
väestötutkimuksen. Väestötutkimuksen tietoja täydennettiin
hoitoilmoitusrekisterin ja kuolemansyyrekisterin tiedoilla.
Tulokset osoittavat, että yhdellä kymmenestä suomalaisesta aikuisesta oli

masennushäiriö vuonna 2011. Masennushäiriöiden esiintyvyys lisääntyi
vuosien 2000 ja 2011 välillä erityisesti naisilla. Väestötutkimusten
perusteella tehdyt arviot masennushäiriöiden esiintyvyydestä voivat olla
todellista pienempiä, koska niistä kärsivät henkilöt osallistuvat
väestötutkimuksiin muita harvemmin. Riskitekijöitä uusien
masennushäiriöiden ilmaantumiselle olivat nuori ikä, useat lapsuuden
vastoinkäymiset, ahdistuneisuushäiriö, lievät masennuksen oireet jo
tutkimuksen alussa sekä sosiaalisen pääoman ulottuvuuksista heikko kyky
tuntea luottamusta. Niistä seurantatutkimuksen henkilöistä, joilla oli
tutkimuksen alussa masennushäiriö, 34‒43 %:lla oli todettavissa masennus-,
ahdistus- tai alkoholihäiriö yhdentoista vuoden kuluttua
seurantatutkimuksessa, ja 48‒61 %:lla oli kliinisesti merkittäviä
masennusoireita. Riski häiriön pitkittymiselle oli suurempi niillä, joilla oli
muita vaikeampia oireita vuonna 2000, sekä niillä, jotka eivät olleet
naimisissa tai avoliitossa. Masennushäiriöihin liittyi kaksinkertainen
kuolleisuusriski, alkoholihäiriöissä riski oli 1.7-kertainen, kun taas
ahdistushäiriöissä riski ei ollut suurentunut.
Tämä tutkimus vahvistaa, että masennushäiriöt ovat kasvava

kansanterveysongelma Suomessa ja niillä on vakavia pitkäaikaisia
seurauksia. On erittäin tärkeää kehittää sekä ennaltaehkäisyä että
5
toimenpiteitä, joilla vähennetään masennushäiriöiden kielteisiä
terveydellisiä ja sosiaalisia seurauksia.
6
ACKNOWLEDGEMENTS
This study was conducted between 2010 and 2016 at the Mental Health Unit
of the National Institute for Health and Welfare. I sincerely thank the former
and the current Heads of Department, Professor Jouko Lönnqvist and
Research Professor Mauri Marttunen, for providing such excellent research
facilities. I wish to thank Professor Markku Heikinheimo, Head of the
National Graduate School of Clinical Investigation, and Professor Antti
Mäkitie, Head of the Doctoral Program in Clinical Research, for the
opportunity to study in the graduate school, and for its financial and
institutional support.
In addition to the National Graduate School of Clinical Investigation, this

study received financial support from the Matti and Maija Vaskio
Foundation, the Foundation for Psychiatric Research in Finland, and the
Jalmari and Rauha Ahokas Foundation.
It has been a privilege to learn from my supervisors. I am grateful to Docent

Samuli Saarni for generously accepting to guide me into the world of
research from my first steps, when I clearly had no idea of what I wanted to
do or how to do go about it. His advice has always been friendly, constructive
and concise. I am in awe of the way he balances his many responsibilities and
interests, and how he approaches research as one more piece in the puzzle to
improve health and well-being.
My deepest gratitude and admiration goes to Research Professor Jaana

Suvisaari. All my questions, big and small, received her immediate attention
– sometimes at 7 a.m. on a weekday or 11 p.m. on a weekend, despite her
busy schedule and many responsibilities. I am particularly grateful for her
unstinting commentary on this summary, which would have been
significantly worse without her motivation to expand its themes and search
for more updated references.
I am grateful to the reviewers of this thesis, Docent Soili Lehto and Professor
Jouko Miettunen, for their perceptive comments and suggestions, and for a
review process that was much easier than I had anticipated.
I thank my thesis committee members, Mika Gissler and Kristian Wahlbeck,

for their encouragement and advice. We did not have many meetings in
person, but having such experienced and kind people supporting me has
been reassuring.
7
The writing of this thesis has been solitary, but by no means lonely, work. I
have been fortunate to work with a large and talented team of people. Your
insightful comments on my manuscripts have taught me much and greatly
improved my work. I regret that most of our contact has been via email, and I
hope to see more of you in the future. Thank you Jonna Perälä, Krista Partti,
Sebastián Peña, Seppo Koskinen, Tommi Härkänen, Sami Pirkola, Suoma
Saarni, Kirsi Ahola, Aino Mattila, Jens Strehle, Satu Viertiö and Tarja
Nieminen. Tommi deserves a special mention for his incredible knowledge
and skills in statistics, and his endless patience in trying to impart some of
that wisdom to me. I could not have done this without you.
This work was made possible by the Health 2000 and Health 2011 Survey
teams, lead by Arpo Aromaa and Seppo Koskinen, whose work has allowed
me to analyse such exceptional data.
I am grateful to Lezak Shallat and Timo and Tiina Joenpelto for language
revision of this summary.
I thank my employers in Chile, Dr. Mauricio Gómez and Professor Báltica

Cabieses, for being so understanding concerning the requirements of
completing my PhD, including traveling to Finland for its defense. I am
especially grateful to Pedro Zitko for providing the stimulating academic
environment I had sorely missed in Chile.
I am so grateful to my friends, for their academic and, more importantly,

non-academic support for all these years. Eeva, Eliisa, Helka, Liisi, Kanerva,
Marianna, Miira, Sara, Suvi, Tiia, Ulla O., Ulla S., Venla and everyone else –
thank you!
I am grateful to my parents, Ritva and Pekka, and my sister Emilia for

teaching me the values of social justice, discipline and hard work, but also the
importance of love and support. I received a solid foundation that has
allowed me to venture out to the world, knowing I can always go back home,
no matter what. Due to the long-distance nature of this work, my parents
have also had a very hands-on relationship with this thesis, carrying
computers across the Atlantic and printing the draft thesis and and handing
it in to the faculty. I am eternally grateful for this support throughout my
years of study.
I thank my Chilean family, Yamile, Sebastián and Paulina, for their equally
concrete contribution, in the form of countless hours of baby-sitting, love and
support.
My husband Sebastián, I thank for his constructive, although sometimes

unnecessarily harsh, critique as a co-author. But more that that, I thank him
8
for having faith in me; having patience; for understanding the ups and downs
of it all, for being there for me. Thank you for everything. I love you.
My children Sebastián and Alvar, both born during the course of this thesis,
have guaranteed that my relationship to it was a healthy one: that it was “just
a job” I do between 9 a.m. and 5 p.m., while Life is what happens during the
rest of the hours of the day.
Niina Markkula
May 2016
Santiago, Chile
9
CONTENTS
Abstract .................................................................................................................. 4
Tiivistelmä .............................................................................................................. 5
Acknowledgements ................................................................................................ 7
Contents ............................................................................................................... 10
List of original publications ..................................................................................15
Abbreviations ....................................................................................................... 16
1 Introduction .................................................................................................... 18
2 Review of the literature .................................................................................. 20
2.1 Definition of depressive disorders .......................................................... 20
2.1.1 ICD-10............................................................................................... 20
2.1.2 DSM-IV ............................................................................................ 21
2.1.3 DSM-5 .............................................................................................. 22
2.1.4 Main differences between the diagnostic systems .......................... 23
2.2 Prevalence of depressive disorders......................................................... 24
2.2.1 Measurement of depressive disorders in population studies ......... 24
2.2.1.1 Methods of assessing depressive disorders in

population studies .............................................................................. 24
2.2.1.2 Methodological challenges in prevalence studies ............... 26
2.2.2 Global prevalence of major depressive disorder............................. 27
2.2.3 Prevalence of depressive disorders in Finland ............................... 31
2.2.4 Prevalence of dysthymia .................................................................. 31
2.2.5 The global burden of depressive disorders ..................................... 32
2.3 Prevalence trends of depressive disorders over time ............................. 33
2.4 Risk factors for depressive disorders...................................................... 33
2.4.1 Sociodemographic factors................................................................ 34
10
2.4.1.1 Gender .................................................................................. 34
2.4.1.2 Age ........................................................................................ 35
2.4.1.3 Socioeconomic position ....................................................... 35
2.4.1.4 Marital status ....................................................................... 37
2.4.2 Family history and early life experiences ....................................... 37
2.4.2.1 Family history ...................................................................... 37
2.4.2.2 Early life experiences........................................................... 37
2.4.3 Social capital .................................................................................... 38
2.4.4 Chronic somatic conditions and health behaviours ....................... 39
2.4.5 Other psychiatric disorders and personality traits ......................... 40
2.5 Prognosis of depressive disorders .......................................................... 42
2.5.1 Key concepts of prognosis ............................................................... 42
2.5.1.1 Remission, recovery and recurrence .................................... 42
2.5.1.2 Residual symptoms and psychosocial disability ................. 43
2.5.2 Remission, recovery and recurrence in population studies ........... 43
2.5.3 Quality of life in peOPLE recovered from depression .................... 45
2.5.4 Predictors of different negative outcomes ...................................... 45
2.5.4.1 Individual characteristics .................................................... 45
2.5.4.2 Disorder characteristics....................................................... 46
2.6 Excess mortality in depressive, anxiety and alcohol use disorders ....... 47
2.6.1 Excess mortality in major depressive disorder ............................... 47
2.6.2 Excess mortality in dysthymia ........................................................ 48
2.6.3 Excess mortality in anxiety disorders ............................................. 49
2.6.4 Excess mortality in alcohol use disorders....................................... 49
2.6.5 Possible mechanisms of excess mortality ....................................... 50
2.6.5.1 Suicide and other unnatural deaths .................................... 50
2.6.5.2 Depression and coronary heart disease .............................. 50
11
2.6.5.3 Hazardous health behaviours ...............................................51
2.6.5.4 Biological dysregulation ...................................................... 52
2.6.5.5 Antidepressants ................................................................... 53
2.6.5.6 Health care utilisation ......................................................... 53
2.7 Summary of the literature review and gaps in knowledge ..................... 54
3 Aims of the study ............................................................................................ 56
4 Materials and methods ................................................................................... 57
4.1 Data sources ............................................................................................ 57
4.1.1 Health 2000 Survey ......................................................................... 57
4.1.2 Health 2011 Survey .......................................................................... 57
4.1.3 Register data .................................................................................... 59
4.1.3.1 Care Register for Health Care (HILMO) .............................. 59
4.1.3.2 Mortality register (Sub-studies III-IV) ................................ 61
4.2 The CIDI Interview ................................................................................. 61
4.3 Study variables ........................................................................................ 63
4.3.1 Outcome variables ........................................................................... 63
4.3.1.1 Diagnostic status at follow-up (Sub-studies I-III) ............... 63
4.3.1.2 New-onset depression (Sub-study II) .................................. 64
4.3.1.3 Recovery and persistence (Sub-study III) ........................... 64
4.3.1.4 Depressive symptoms and psychological distress

(Sub-study III) .................................................................................... 64
4.3.1.5 Health-related quality of life (Sub-study III) ...................... 65
4.3.1.6 Self-rated health (Sub-study III) ......................................... 65
4.3.1.7 Mortality and causes of death (Sub-studies III-IV)............. 65
4.3.2 Exposure and confounder variables at baseline ............................. 65
4.3.2.1 Sociodemographic variables (all Sub-studies) .................... 65
4.3.2.2 Childhood adversity (Sub-studies II-III) ............................ 66
12
4.3.2.3 Social capital (Sub-studies II-III) ....................................... 66
4.3.2.4 Somatic comorbidity, smoking and obesity (Sub-

studies II-IV) ...................................................................................... 67
4.4 Statistical methods ................................................................................. 68
4.4.1 Multiple imputation......................................................................... 68
4.4.2 Weights ............................................................................................ 69
4.4.3 Sub-Study I ...................................................................................... 70
4.4.4 Sub-Study II .................................................................................... 70
4.4.5 Sub-Study III .................................................................................... 71
4.4.6 Sub-Study IV .................................................................................... 71
5 Results ............................................................................................................ 73
5.1 Prevalence of depressive disorders (Sub-study I) .................................. 73
5.2 Risk factors of depressive disorders (Sub-study II) ............................... 75
5.3 Prognosis of depressive disorders (Sub-study III)..................................77
5.4 Excess mortality in depressive, anxiety and alcohol use disorders

(Sub-study IV) ................................................................................................. 79
6 Discussion ...................................................................................................... 82
6.1 Summary of the main results .................................................................. 82
6.2 Potential risk factors and predictors of negative outcomes of

depression ........................................................................................................ 83
6.2.1 Gender and age ................................................................................ 83
6.2.2 Socioeconomic position .................................................................. 85
6.2.3 Childhood adversities and family history ....................................... 86
6.2.4 Marital status and social capital ..................................................... 87
6.2.5 Chronic somatic conditions............................................................. 88
6.2.6 Other psychiatric disorders and psychological symptoms ............. 89
6.3 Increasing prevalence of depressive disorders and impact of

non-participation ............................................................................................. 90
13
6.4 Mortality in depressive, anxiety and alcohol use disorders ................... 91
6.5 Methodological considerations – strengths and limitations ................. 93
6.5.1 Population survey data .................................................................... 93
6.5.2 The CIDI interview .......................................................................... 94
6.5.3 Non-participation and multiple imputation ................................... 94
7 Conclusions ..................................................................................................... 96
References ............................................................................................................ 99
14
LIST OF ORIGINAL PUBLICATIONS
This thesis is based on the following publications:
I Markkula N, Suvisaari J, Saarni SI, Pirkola S, Peña S, Saarni SE,

Ahola K, Mattila AK, Viertiö S, Strehle J, Koskinen S, Härkänen
T. Prevalence and correlates of major depressive disorder and
dysthymia in an eleven-year follow-up - results from the
Finnish Health 2011 Survey Journal of Affective Disorders 2015
(173) 73-80
II Markkula N, Marola N, Nieminen T, Koskinen S, Saarni SI,

Härkänen T, Suvisaari J. Predictors of new-onset depressive
disorders in the working-age population – results from the
longitudinal Finnish Health 2011 Study. Submitted for
publication
III Markkula N, Härkänen T, Nieminen T, Peña S, Mattila AK,

Koskinen S, Saarni SI, Suvisaari J. Prognosis of depressive
disorders in the general population– results from the
longitudinal Finnish Health 2011 Study. Journal of Affective
Disorders 190 (2016) 687–696
IV Markkula N, Härkänen T, Perälä J, Partti K, Peña S, Koskinen S,

Lönnqvist J, Suvisaari J, Saarni SI., Mortality in people with
depressive, anxiety and alcohol use disorders in Finland. Br J
Psychiatry, 2012. 200(2): p. 143-9.
The publications are referred to in the text by their Roman numerals.
15
Introduction
ABBREVIATIONS
AUD Alcohol use disorder

BDI Beck Depression Inventory
BMI Body Mass Index
CES-D Center for Epidemiologic Studies Depression Scale
CIDI Composite International Diagnostic Interview
CIS-R Clinical Interview Schedule (Revised)
DALY Disability Adjusted Life-Year
DEPS Depression Scale
DIS Diagnostic Interview Schedule
DPAX Depression and Anxiety Schedule
DSM Diagnostic and Statistical Manual
GAD Generalised anxiety disorder
GBD Global Burden of Disease
GHQ General Health Questionnaire
HADS Hospital Anxiety and Depression Scale
HPA Hypothalamus-pituitary-adrenal (axis)
HSCL Hopkins Symptom Checklist
ICD-10 International Classification of Diseases, 10th revision
ICPE International Consortium in Psychiatric Epidemiology
IPW Inverse probability weights
ISPI Iowa Structured Psychiatric Interview
M-CIDI Composite International Diagnostic Interview, Munich version
MDD Major depressive disorder
MDE Major depressive episode
MI Multiple imputation
MINI Mini-International Neuropsychiatric Interview
NCS-R National Comorbidity Survey, Replication
OR Odds ratio
PDD Persistent depressive disorder
PERI Psychiatric Epidemiological Research Instrument
PHQ-G Personal Health Questionnaire
PHQ Patient Health Questionnaire
PSE Present State Examination
WHO World Health Organization
WMHS World Mental Health Survey
SADS Schedule for Affective Disorders and Schizophrenia
SCAN Schedules for Clinical Assessment in Neuropsychiatry
SCID Structured Clinical Interview for DSM
SCL-90 Symptom Checklist 90 items
SSRI Selective serotonin reuptake inhibitors
YLD Years lived with disability
16
YLL Years of life lost
etc. et cetera
i.e. id est
e.g. exempli gratia
17
Introduction
1 INTRODUCTION
Depression can be used to describe everyday feelings of sadness. However, it

also refers to a psychiatric disorder whose key feature is a sad or depressed
mood, accompanied by other distinct symptoms, including loss of interest,
fatigue, memory problems, insomnia and recurrent thoughts of death.
Typically, the disorder causes suffering and renders the person affected
partly or fully unable to carry on with everyday life.
Every year, approximately one in 20 people suffers from major depressive

disorder (MDD) (Ferrari et al., 2013c). However, this rough estimate
depends on how the exact criteria of the disorder are defined, how it is
measured, and in which population, country and continent the study is
carried out (Kessler and Bromet, 2013). The annual prevalence of dysthymia
(another common depressive disorder with typically a more chronic course
than MDD) is much lower, estimated at 1-2% (Charlson et al., 2013, Gureje,
2011, Waraich et al., 2004). From 9-16% of the population may have a
depressive disorder severe enough to warrant specialised medical care over
the course of their lifetime (Pedersen et al., 2014) and many more will suffer
from a less severe form. Increased awareness of depression in society and
increased use of antidepressants has led many to believe that the prevalence
of depression has increased, but there have been no indications of this in
previous population studies (Ferrari et al., 2013c, Patten et al., 2015a).
Depression is more common among women, younger people, unmarried or

divorced individuals, and those in a socially disadvantaged position (Kessler
and Bromet, 2013). It is not clear, however, whether the association between
these factors and depression necessarily implies causation. Some factors may
increase the risk of developing depression, while others decrease the
likelihood of recovery.
The age of onset of depression is typically in the early 20s, but there is wide
variation (Kessler and Bromet, 2013). On average, an illness episode lasts a
few months (Eaton et al., 2008b, Kessler et al., 2003, Spijker et al., 2002),
but a chronic or recurrent course is regrettably common (Eaton et al., 2008b,
Spijker et al., 2002). Depression may negatively impact educational
attainment, marital relations, employment, parental functioning, as well as
general health status and mortality risk (Kessler and Bromet, 2013). The risk
of suicide increases as much as 20-fold compared to people without
depression, and 4-7% of depressed individuals die by suicide (Holmstrand et
al., 2015, Isometsä, 2014, Nordentoft et al., 2011).
18
The aim of this study is to examine prevalence, predictors and different
adverse outcomes of depressive disorders in a general population setting.
The study is based on two large health examination surveys of the Finnish
population, the Health 2000 study and its follow-up study, the Health 2011
Survey. In addition, register data on hospitalisations and mortality is utilised.
19
Review of the literature
2 REVIEW OF THE LITERATURE
2.1 DEFINITION OF DEPRESSIVE DISORDERS

Depressive disorders refer to conditions that are characterised by sad or
depressed mood and also entail other symptoms, such as inability to
experience pleasure, feelings of guilt, thoughts of death, cognitive symptoms
(such as diminished ability to think or concentrate) and physical changes
(such as insomnia and loss of appetite). These disorders cause emotional and
physical suffering, and often lead to reduction in functional capacity and the
ability to work.
The pathophysiology, or mechanism of how depressive disorders develop in

the human body, is not clear. This is why the diagnostic criteria rely on
describing a collection of symptoms considered essential and characteristic
of the disorder. Currently, two diagnostic systems are in place: the
International Classification of Diseases, 10th revision (ICD-10), and the
Diagnostic and Statistical Manual, fourth and fifth revisions (DSM-IV and
DSM-5). The ICD-10 is used by health care systems in many countries,
including Finland, whereas the DSM-IV and DSM-5 are frequently used in
research. Both criteria are presented here.
2.1.1 ICD-10
According to the ICD-10 Classification of Mental and Behavioural Disorders
(WHO, 1993), a depressive episode (code F32) is a condition that lasts a
minimum of two weeks and is characterised by at least two of three core
symptoms: depressed mood that is present for most of the day and almost
every day; loss of interest in activities that the person normally enjoys; and
fatigue or decreased energy. In addition, one or more of the following
symptoms must also be present, so that there are a minimum of four: loss of
confidence and self-esteem; excessive and inappropriate guilt; recurrent
thoughts of death, suicide or suicidal behaviour; diminished ability to think
or concentrate; agitation or retardation; sleep disturbance; change in
appetite and weight change. In addition to these symptom criteria,
hypomanic or manic episodes should be excluded, as well as the symptoms
attributable to psychoactive substance use or organic mental disorder.
Depressive episodes are further classified into mild (4-5 symptoms),

moderate (6-7 symptoms) or severe (8-9 symptoms). If the person has had
prior depressive episodes and recovered for a minimum of 2 months between
episodes, the disorder is called “Recurrent depressive disorder” (F33).
20
Dysthymia (F34.1) is a condition of depressed mood either constantly or
recurrently for a minimum of two years, in which periods of normal mood do
not last for longer than a few weeks. Few or none of the episodes are severe
enough to meet criteria for recurrent mild depressive disorder. Three of the
following symptoms should be present during some of the periods of
depression: reduction in energy or activity; insomnia; loss of self-confidence;
difficulty concentrating; tearfulness; loss of interest in sex and other
pleasurable activities; feelings of hopelessness; perceived inability to cope
with everyday responsibilities; pessimistic attitude toward the future or
brooding over the past; withdrawal from social activities; and being less
talkative than normal. The onset of dysthymia may be early (adolescence to
late 20s) or late, often following a depressive episode.
The ICD-10 also presents criteria for “Recurrent brief depressive disorder”
(F38.10), in which recurrent episodes of equal intensity but shorter duration
than depressive episode occur.
2.1.2 DSM-IV
The DSM-IV criteria for major depressive disorder (MDD) (code 296.xx)
include minimum duration of two weeks, presence of either depressed mood
or anhedonia most of the day and nearly every day, and other symptoms of
the following, for a minimum of five: significant change in appetite or weight;
insomnia or hypersomnia; psychomotor retardation or agitation observable
to others; fatigue or loss of energy; excessive or inappropriate guilt or
worthlessness; diminished ability to think or concentrate; and recurrent
thoughts of death or suicide. The disorder is classified into mild, moderate,
severe or with psychotic features.
In addition to symptom and duration criteria, DSM-IV criteria require that

the condition cause clinically significant distress or impairment in social,
occupational or other areas. Exclusion criteria of mania, hypomania,
substance use and somatic conditions are similar to ICD-10. In addition, the
DSM-IV does not establish a diagnosis of MDD if symptoms are related to
the loss of a loved one within the past two months, unless the symptoms
include psychomotor retardation, excessive or inappropriate guilt,
suicidality, psychotic symptoms, or are associated with a marked decrease in
functional capacity.
Dysthymia (300.4) is described as a disorder where the mood is depressed

for most of the day more than half of the time for two years or more, and two
or more of the following symptoms are present: change in appetite; insomnia
or sleeping too much; low energy; low self-esteem, poor concentration; and
hopelessness. The symptoms should cause significant impairment in social,
21
work or other important areas. Symptom-free periods must be no longer

than two months.
2.1.3 DSM-5
There were no changes to core symptoms or duration criteria for MDD
between DSM-IV and DSM-5. A new specifier “with mixed features” was
introduced to describe a depressive episode with a maximum of three manic
symptoms. The most notable change was the omission of the bereavement
exclusion. This was due to the fact that the time limit was considered
arbitrary and not in line with actual duration of grieving, and because
bereavement-related MDD does not differ from non-bereavement-related
MDD in terms of risk factors, previous history of depression, and course or
response to treatment (Kendler et al., 2008, Uher et al., 2014). In the DSM-
5, bereavement is likened to other stressors that commonly influence the
onset and course of MDD. In one study of individuals who experienced a
brief major depressive episode (less than two months), 38% reported the
disorder to be bereavement-related (McCabe and Christopher, 2015), with
certain distinct differences to the non-bereavement related episodes, such as
less suicidality and worthlessness.
In the DSM-5, dysthymia was reconceptualised as “persistent depressive

disorder” (PDD) (300.4). This category includes both chronic major
depressive disorder and former dysthymia. The change was made because no
significant differences between these two conditions had been established,
whereas significant differences appear to exist between chronic and non-
chronic depression. Some 29% of individuals with lifetime depression are
estimated to suffer from persistent depressive disorder (Murphy and Byrne,
2012).
“Premenstrual dysphoric disorder” (625.4) was moved from an appendix to

the main category of depressive disorders. “Disruptive mood dysregulation
disorder” (269.99) for children up to 18 years was also added to the category.
The changes to the diagnostic categories of depressive disorders in the DSM-

5 may result in changes in prevalence (Uher et al., 2014). Guidance on
whether PDD and MDD can be diagnosed at the same time is contradictory,
but according to PDD criteria: “If the symptom criteria are sufficient for a
diagnosis of a major depressive episode at any time during this period, then
the diagnosis of major depression should be noted, but it is coded not as a
separate diagnosis but rather as a specifier with the diagnosis of persistent
depressive disorder.” Therefore, it is possible that the prevalence of MDD will
decrease, as chronic cases will be diagnosed as PDD. On the other hand, the
omission of the bereavement exclusion may result in an increase in the
22
prevalence of MDD, given the large proportion of people who attributed a
depressive episode to bereavement (McCabe and Christopher, 2015).
2.1.4 MAIN DIFFERENCES BETWEEN THE DIAGNOSTIC SYSTEMS

The main differences in diagnostic criteria for MDD between the ICD-10 and
the DSM (IV and 5) systems are: 1) the different number of symptoms
required (four in ICD-10 and five in DSM); 2) different number of core
symptoms (three in ICD-10 and two in DSM); and 3) the DSM criterion of
“significant distress or functional impairment,” which the ICD-10 does not
consider (Table 1).
In the case of dysthymia, the differences are: 1) minimum number of

symptoms (four in ICD-10 and three in DSM); and 2) number of symptoms
listed (12 in ICD-10 and seven in DSM).
Table 1. Main differences in the diagnostic systems of depressive disorders: ICD-10,

DSM-IV and DSM-5
Core symptoms Number of Other criteria Exclusion

symptoms
(total)
Major depressive disorder

ICD-10 Minimum 2 of the 4 - Hypomania, mania,
following: depressed symptoms
mood, anhedonia, and attributable to
decreased energy or psychoactive
fatigue substance use or
organic mental
disorder
DSM-IV Either depressed 5 Must cause Same as in ICD-10,
mood or anhedonia clinically and also if symptoms
significant occur within 2 months
distress or of bereavement
impairment
DSM-5 As in DSM-IV 5 As in DSM-IV Same as in ICD-10,
no bereavement
exclusion
Dysthymia
ICD-10 Depressed mood 4 (not Mania
either constantly or necessarily at
recurrently for a the same time)
minimum of two years from a list of 12
DSM-IV Depressed mood 3 from a list of MDD in the first two
7 years, mania,
substance induced
23
2.2 PREVALENCE OF DEPRESSIVE DISORDERS
2.2.1 MEASUREMENT OF DEPRESSIVE DISORDERS IN POPULATION

STUDIES
There is wide variation in prevalence of depressive disorders across

countries. This is assumed to be due to real, existing variations and to
methodological differences in the studies.
One issue is the diverse definitions of depression. Some studies include MDD
only, whereas others measure MDE, including bipolar depression. Still others
include dysthymia or depression not otherwise specified. Wider inclusion of
diagnostic categories increases observed prevalence rates (Ferrari et al.,
2013c).
Diagnostic instruments also matter. When symptom scales are used as a

basis of diagnosis, measured prevalences tend to be higher than when
structured diagnostic interviews are used (Ferrari et al., 2013c). Some studies
have also found higher prevalences using ICD-10 instead of DSM-IV; in
others no significant differences existed (Andrews et al., 2001, Ferrari et al.,
2013c). Differences among studies in target population, sampling and
participation rates also contribute to inaccuracies in prevalence estimates.
Finally, conceptual models of depression as well as the presentation of

symptoms differ across cultures (Karasz, 2005, Ventevogel et al., 2013). This
limits the applicability of psychometric instruments designed for different
cultural contexts. To enable comparison and for practical reasons, however,
the same standardised and validated instruments are normally used in
epidemiological studies, irrespective of the context.
2.2.1.1 Methods of assessing depressive disorders in population

studies
Population studies on psychiatric disorders require tools that are reliable,
replicable and valid to the construct they are measuring, as well as feasibly
applicable to thousands of persons at a reasonable cost. The latter requisite
generally limits the time available and the type of professional who can
administer the tool.
Two methods that meet these requirements are structured interviews and
symptom scales. Structured interviews (or schedules) are designed to assess
the presence of a psychiatric disorder dichotomously – the disorder either is
or is not present. They were developed to enable psychiatric assessment in
24
population surveys using lay interviewer-administered tools (Brugha et al.,
1999). There are two types of structured interviews: semi-structured and
fully structured interviews. A semi-structured interview is performed by a
mental health professional, usually a psychiatrist or psychologist, and
resembles a clinical examination, where the questions have a standard
wording, but probing and further questions are flexible. In a fully structured
interview, a trained lay interviewer asks the questions and clarifying
questions or further details exactly as has been written, and codes answers
according to instructions. The usefulness of a fully structured interview
depends greatly on the exact wording and how understandable it is to the
participants. In practice, due to human resource costs, population surveys
always use fully structured interviews.
The most frequently used structured interview is the Composite

International Diagnostic Interview (CIDI) in its various forms (see Chapter
4.2 for a more detailed description of this instrument). The most relevant
structured interviews are listed in Table 2.
Table 2. Structured diagnostic interviews to assess depressive disorders in

epidemiological studies, ordered by decade of first publication of the instrument
Structured diagnostic interviews (schedules) used to establish diagnosis of Fully /

depression in epidemiological studies Semi-
structured
PSE Present State Examination 1960s Semi
CIS (-R) Clinical Interview Schedule (Revised) 1970s Fully
SADS Schedule for Affective Disorders and Schizophrenia 1970s Semi
DIS Diagnostic Interview Schedule 1980s Fully
DPAX Depression and Anxiety Schedule 1980s Fully
ISPI Iowa Structured Psychiatric Interview 1980s Fully
MINI Mini-International Neuropsychiatric Interview 1990s Fully
SCAN Schedules for Clinical Assessment in Neuropsychiatry 1990s Semi
(development of PSE)
CIDI Composite International Diagnostic Interview (UM-CIDI, 1990s Fully
WMH-CIDI, M-CIDI) (development of DIS)
PRIME- Primary Care Evaluation of Mental Disorders 1990s Semi
MD
SCID Structured Clinical Interview for DSM (III-R, IV and 5) 1990s Semi
(development of SADS)
Symptom scales are designed to measure symptoms of a disorder in a

dimensional scale. Sometimes these scales are used to identify people with
possible depressive disorder by establishing a cut-off point, above which the
25
disorder is possible or probable. They are also used in clinical practice to

monitor levels of symptoms, which is their original intended use. The most
relevant self-reported depressive symptom scales are listed in Table 3.
Table 3. Symptom scales to assess depressive symptoms, ordered by decade of first

publication of the instrument
Symptom scales used to assess level of symptoms and establish possible diagnosis of
depression
HSCL Hopkins Symptom Checklist 1950s

BDI Beck Depression Inventory 1960s
Zung Zung Depression Scale 1960s
CES-D The Center for Epidemiologic Studies Depression Scale 1970s
HADS Hospital Anxiety and Depression Scale 1980s
PERI Psychiatric Epidemiological Research Instrument 1980s
BDI-II Beck Depression Inventory II 1990s
SCL-90 Symptom Checklist 90 items (development of HSCL) 1990s
DEPS The Depression Scale 1990s
PHQ-G Personal Health Questionnaire 1990s
PHQ Patient Health Questionnaire 2000s
The advantage of symptom scales over structured interviews is the shorter

time needed, and the fact that they are self-administered. However, their
diagnostic validity is limited (Pettersson et al., 2015). In this literature
review, studies using valid diagnostic procedures based on structured
interviews are prioritised.
2.2.1.2 Methodological challenges in prevalence studies

2.2.1.2.1 Non-participation of people with mental disorders
When prevalence of mental disorders is assessed through population surveys,
participation in these surveys may influence estimates of prevalence. Many
studies indicate that people with psychiatric disorders have a lower
likelihood of participating in population studies (Eaton et al., 1992, Suvisaari
et al., 2009). Different types of psychiatric disorders have been found to
contribute to non-participation (Haapea et al., 2008). Disorder severity,
however, may increase risk of non-participation in both psychotic and non-
psychotic disorders (Haapea et al., 2007, Lamers et al., 2012). Conversely,
studies in the Netherlands have found that non-participation was associated
with correlates of mental disorders, such as low education, younger age and
unemployment, but not the disorders themselves (de Graaf et al., 2000, de
Graaf et al., 2013). These results have not been replicated in other studies.
26
Therefore, non-participation of people with psychiatric disorders may bias
prevalence estimates. It is thus likely that true prevalences are higher than
those estimated based on population surveys that do not correct for non-
participation (Haapea et al., 2008).
2.2.1.2.2 Lifetime prevalence estimates

Doubts on the accuracy of lifetime prevalence estimates in population
surveys have been raised for two main reasons: the high ratio of period and
lifetime prevalence rates; and the curvilinear association of lifetime
prevalence estimates with age (Parker, 1987). In the first case, a low ratio of
period to lifetime prevalence would be expected in the case of episodic
disorders, such as depression. However, period prevalences of 50-60% of
lifetime prevalence are often observed (Moffitt et al., 2010, Parker, 1987). In
the second case, one would assume an increasing prevalence of lifetime
disorders with age. However, the lifetime prevalence of depression peaks at
mid-life and then decreases.
Both observations are explained by recall bias, which is the systematic

inaccuracy of reporting events that occurred in the past. The influence of
recall bias on assessments of lifetime prevalence of mental disorders has
been ascertained in several studies (Kruijshaar et al., 2005, Moffitt et al.,
2010, Patten et al., 2012, Takayanagi et al., 2014). When comparing
prospective to retrospective study settings, the rate of lifetime mental
disorders was double in the prospective compared to retrospective
assessment (Moffitt et al., 2010). In the case of MDD, the increase is almost
threefold: the lifetime prevalence evaluated retrospectively was 4.5%,
whereas cumulative evaluation from several waves of a prospective survey
yielded a lifetime prevalence of 13.1% (Takayanagi et al., 2014). In another
prospective study, only 40% of persons reported a past depressive episode,
verified in a previous wave of the study, 10 years after it had occurred (Patten
et al., 2012). Finally, when the association of declining lifetime prevalence
with age was examined, it was explained by recall bias, and possibly
differential mortality rates, rather than cohort effect (Patten, 2003).
For these reasons, this literature review focuses on period prevalence only,
and 12-month prevalence is presented whenever available.
2.2.2 GLOBAL PREVALENCE OF MAJOR DEPRESSIVE DISORDER
The prevalence of major depressive disorder (MDD) varies widely depending

on the population studied and the diagnostic instrument and diagnostic
criteria used. A recent systematic review found the lowest prevalence
(0.05%) in Japanese males aged 65 and older in 1998, using a structured
27
clinical interview (SCID). This review found the highest prevalence (73%)
among women aged 15 and older in post-war Afghanistan in 2005, using a
symptom scale (HSCL-25) (Ferrari et al., 2013c).
When comparing representative adult population studies, differences still

exist between studies, countries and regions. These variations are not
explained by country income categories: applying standardised methodology
across 18 countries, the WHO World Mental Health Surveys, (WMHS) found
an average 12-month prevalence of 5.5% in high-income countries and 5.9%
in low-income countries (Bromet et al., 2011). In contrast, regional
differences do exist, so that low prevalences are found in East and Southeast
Asia and Sub-Saharan Africa, whereas higher prevalences are found in North
and South America, North Africa and some European countries (Bromet et
al., 2011, Ferrari et al., 2013c, Steel et al., 2014).
Some variations are related to the methodological issues discussed in

Chapter 2.1.1. But it is likely true that differences between countries do also
exist. The WMHS used the same structured interview and similar sampling
procedures in all participating countries, and still found fivefold differences
in 12-month prevalence of MDE, varying from 2.2% in Japan to 10.4% in
Brazil (Bromet et al., 2011). Previously, it was thought that the differences
might be due to different reporting thresholds. But analyses of the WMHS
data demonstrate increasing disability with increasing prevalence, pointing
at a true difference in prevalence rather than differences in interpretation or
measurement (Kessler and Bromet, 2013).
Recent high-quality reviews of epidemiological literature of MDD have

established a global point-prevalence (current or past month) of 4.7% and a
“pooled period prevalence of mood disorder” (point or 12-month) of 5.4%
(Ferrari et al., 2013c, Steel et al., 2014). Earlier systematic reviews and meta-
analyses have placed the average 12-month prevalence in the range of 4.1% to
6.9% (Eaton et al., 2008a, Waraich et al., 2004, Wittchen et al., 2011).
Several nationally representative studies have found annual prevalences of
3.9% to 6.8% (Table 4). It can therefore be concluded that the global 12-
month prevalence of MDD is approximately 5%, but that significant
variations exist between countries and regions.
28
Table 4. Prevalence of depressive disorders in different general population studies
12-month
Publication Study setting Instrument prevalence
used of MDD
Bijl et al., 1998. Prevalence of Netherlands Mental CIDI 5.8%
psychiatric disorder in the Health Survey
general population: and Incidence Study-
results of the Netherlands (NEMESIS),
Mental Health Survey representative of the
and Incidence Study Dutch general population
(NEMESIS) (Bijl et al., 1998) aged 18-64
Murphy et al., 2000. A 40-year Stirling County Study, DPAX 1 and 2, 5.3% in 1952
Perspective on the Prevalence longitudinal general DIS and 1970,
of Depression (Murphy et al., population study in 2.9-5.7% in
2000) Canada (general 1990
population aged 18 and depending on
older; in 1952 heads of instrument
households only)
Andrews et al., 2001. Australian National CIDI 2.1 6.7% (ICD-
Prevalence, comorbidity, Mental Health Survey, 10), 6.3%
disability and service utilisation. representative population (DSM-IV)
(Andrews et al., 2001) study of persons 18 years
and older
Kessler et al., 2003. The National Comorbidity WHO-CIDI 6.6%
Epidemiology of Major Survey Replication (NCS-
Depressive Disorder. R), representative
(Kessler et al., 2003) population survey of
household residents aged
18 and older in the US
Hasin et al. 2005. Epidemiology National Epidemiologic Alcohol Use 5.3%

of Major Depressive Disorder Survey on Alcoholism Disorder and
(Hasin et al., 2005) and Related Conditions, Associated
nationally representative Disabilities
population survey of Interview
household residents aged Schedule–
18 and older in the US DSM-IV
Version
(AUDADIS-IV)
de Graaf et al., 2012. Netherlands Mental CIDI 3.0 5.2%
Prevalence of mental disorders Health Survey and
and trends from 1996 to 2009. Incidence Study 2
Results from the Netherlands (NEMESIS-2),
Mental Health Survey representative of the
and Incidence Study-2 Dutch general population
(de Graaf et al., 2012) aged 18-64
Patten et al. 2015 Descriptive Canadian Community WHO-CIDI 3.9%
Epidemiology of Major Health Study—Mental
Depressive Disorder in Canada Health (CCHS-MH),
in 2012 nationally representative
(Patten et al., 2015b) study of persons 15 years
and over
Jacobi et al., 2015. Twelve- German Health Interview DEGS-CIDI 6.8%
months prevalence of mental and Examination Survey
disorders in the German Health for Adults
Interview and Examination Mental Health Module
Survey for Adults – Mental (DEGS1-MH), nationally
Health Module (DEGS1-MH) representative study of
29
(Jacobi et al., 2015) the population aged 18-

79
Reviews and cross-national studies
Andrade et al., 2003. The Cross-national study of WHO-CIDI, Variation

epidemiology of major general populations with University of from 1.2%
depressive episodes: results 10 low- to high-income Michigan-CIDI in Japan to
from the International countries included and Munich-CIDI 10.0% in
Consortium of Psychiatric the US
Epidemiology (ICPE) Surveys. (MDE, not
(Andrade et al., 2003) MDD)
Waraich et al., 2004. Systematic review of 15 studies using Pooled
Prevalence and Incidence general population standardised or prevalence
Studies of Mood Disorders: A studies clinical diagnostic 4.1%
Systematic Review of the tools were
Literature. included
(Waraich et al., 2004)
Wittchen and Jacobi, 2005. Size Review of 27 studies in Varying, eg. CIDI, Median
and burden of mental disorders 16 European countries GHQ prevalence
in Europe—a critical review and 6.9%,
appraisal of 27 studies. variance
(Wittchen and Jacobi, 2005) 3.1-10.1%
(MDE)
Wittchen et al., 2011. The size Systematic literature 25 community 5.7% from
and burden of mental disorders reviews, re-analyses of studies on studies;
and other disorders of the brain existing data sets, depression were 6.9% when
in Europe 2010. national surveys and included taking into
(Wittchen et al., 2011) expert consultations account
expert
estimates
Bromet et al. 2011. Cross- World Mental Health WHO-CIDI Average
national epidemiology of DSM- Surveys, 18 countries prevalence
IV major depressive episode with nationally 5.5% in
(Bromet et al., 2011) representative adult high-
samples income and
5.9% in
low-income
countries,
variation
from 2.2%
to 10.4%
Ferrari et al, 2013. Global Systematic review and 116 studies of 4.7% (point
variation in the prevalence and meta-analysis of prevalence with prevalence)
incidence of major depressive community prevalence varying
disorder: a systematic review of studies methodology,
the epidemiological literature including
(Ferrari et al., 2013c) structured
interviews and
symptom scales
Steel et al., 2014. The global Random-effects meta- 174 studies from Pooled
prevalence of common mental analysis of prevalence 63 countries period
disorders: a systematic review studies of all mental published 1980- prevalence
and meta-analysis 1980–2013 disorders, structured 2013 of mood
(Steel et al., 2014) diagnostic interview disorder
methodology 5.4%
30
2.2.3 PREVALENCE OF DEPRESSIVE DISORDERS IN FINLAND
In Finland, there is a long tradition of extensive population health surveys

that include instruments to measure depression. In 1978-1980, the Mini-
Finland health survey included the Present State Examination (PSE) to
assess prevalence of mental disorders. Some 8000 individuals from the
general population, aged 30 years and over, were sampled. The prevalence of
“neurotic depression” was 4.6% (Lehtinen et al., 1990).
In 1996, the short form of the University of Michigan CIDI (UM-CIDI) was
used to assess depression in a random sample of adults (15-75 years,
n=5993). The 12-month prevalence of major depressive episode was 9.3%
(Lindeman et al., 2000). The short form of the UM-CIDI instrument,
however, does not apply exclusions for organic or somatic conditions, nor
does it distinguish between different types of depression.
In the Health 2000 Survey, the full version of the Munich CIDI (M-CIDI) was
used to assess prevalence in adults 30 years and over (n=8028). The 12-
month prevalence of MDD was 4.9%; dysthymia was 2.5%; and any
depressive disorder was 6.5% (Pirkola et al., 2005b). In these figures, non-
participation was accounted for by using poststratification weights.
Incidence of depression in the general population was studied in Southwest

Finland as part of the international ODIN study. The incidence rate for all
depressive disorders was 28.5 / 1000 per year; for first-time episodes, it was
20.5 / 1000 (Lehtinen et al., 2005).
In summary, the prevalence of depressive disorders in Finland is very close

to the global average.
2.2.4 PREVALENCE OF DYSTHYMIA
Compared with MDD, fewer studies have investigated the prevalence of

dysthymia. For the Global Burden of Disease 2010 study, a thorough
systematic review found 38 studies, and established a period (combined
current, 1-, 3-, 6- and 12-month) prevalence of 1.6% (Charlson et al., 2013).
Prevalences in different studies ranged from 0.0% to 8.5%. Data from many
regions was lacking, and the estimate was partly based on modelling. Other
reviews have found annual prevalences of 2.0% (Waraich et al., 2004) and
0.1-1.5% (Gureje, 2011). National studies have found 0.5-1.5% 12-month
prevalence rates in the United States (Blanco et al., 2010); 0.9-2.5% in the
Netherlands (Bijl et al., 1998, de Graaf et al., 2012); 1.1% in Australia
(Andrews et al., 2001); and 1.7% in Germany (Jacobi et al., 2015)
31
2.2.5 THE GLOBAL BURDEN OF DEPRESSIVE DISORDERS
When only mortality is assessed, the impact of mental disorders on disease

burden is largely underestimated. Historically, this has led to mental
disorders being neglected within different public health priorities. The World
Development Report of 1993: Investing in Health report introduced the
Disability Adjusted Life Year (DALY) measure that made it possible to
quantify, for the first time, the real impact of mental disorders on public
health (Bank, 1993). This new focus on disability in addition to mortality has
been of key importance in placing mental disorders on the global public
health agenda.
The global burden of disease (GBD) is calculated as a function of prevalence,

disability and deaths due to the disorder, and is expressed in DALYs. These
are a sum of years of life lost due to premature mortality (YLL) and years
lived with disability (YLD), a calculation that is based on the prevalence and
disability caused by the condition.
In the GBD 2010 study, prevalence and epidemiological modelling was done
for both MDD and dysthymia (Charlson et al., 2013, Ferrari et al., 2013a,
Ferrari et al., 2013c). For the disability weights, lay health state descriptions
were presented in an online survey to nearly 15.000 lay participants from
around the world. When 0 represents full health and 1 represents death,
disability weights were assigned as follows: mild depression 0.16; moderate
depression 0.41; severe depression 0.66; and dysthymia 0.16. Based on
population surveys, it was estimated that 14% of people with depressive
disorders are asymptomatic; 59% have a mild condition; 17% moderate; and
11% severe.
Globally, mental and substance use disorders are the leading cause of years
lived with disability (YLD), causing 21.2% of all YLDs (Global Burden of
Disease Study 2013 Collaborators, 2015). The burden of disability due to
MDD (measured in YLD) increased by 53% between 1990 and 2013, but the
age-standardised disability burden increased by only 4% (Global Burden of
Disease Study 2013 Collaborators, 2015). The burden due to dysthymia
increased by 55% without any significant change in age-standardised rate.
When taking into account mortality and disability, the age-standardised rate
of DALYs due to depressive disorders did not change significantly between
1990 and 2013 (Murray et al., 2015). MDD ranks second (after lower back
pain) as a contributor to the burden of years lived with disability (YLD), and
accounts for 8.1% of all YLDs. Dysthymia is the 16th contributor and
accounts for 1.3% (Global Burden of Disease Study 2013 Collaborators,
2015). Considering the impact of non-participation has on prevalence
estimates, it is likely that the global burden of depression is even higher than
currently estimated.
32
In the GBD 2010, mortality due to depressive disorders was modelled so that
MDD was considered a risk factor for suicide (odds ratio (OR) 19.9) and
ischeamic heart disease (OR 1.6) (Ferrari et al., 2013b). This meant that
MDD accounted for 16 million DALYs (46%) due to suicides, and 3.8 million
DALYs (2.9%) due to ischemic heart disease. However, these deaths were not
included in the calculation of burden caused by depressive disorders. Had
they been reattributed to MDD, the overall burden of MDD would have
increased from 2.5% to 3.4% of global DALYs, ranking as the eighth instead
of the eleventh cause (Ferrari et al., 2013b).
There is substantial regional variation in the contribution of depressive

disorders to the overall disease burden. For example, in the North African
and Andean South American regions, depressive disorders are the third
leading contributor to DALYs, whereas in different parts of Sub-Saharan
Africa they are the 13th to 19th contributor, respectively, and in South Asia
they are the 14th contributor (Ferrari et al., 2013b). However, in the burden
of disability (YLD) category, MDD is among the three most important
contributors in all but a few countries globally, and in all countries it ranks
among the top 10 (Global Burden of Disease Study 2013 Collaborators, 2015).
In Finland, depressive disorders are the 6th leading contributor to DALYs.
2.3 PREVALENCE TRENDS OF DEPRESSIVE

DISORDERS OVER TIME
Through heightened awareness of depression among the general population,
increased help-seeking and use of services, and certain topical issues (such as
increased disability pensions due to depression), there is a widespread
perception that depression has increased. Contrary to his perception
however, no studies indicate this. Globally, the point prevalence of MDD has
remained stable at 4.4% in 1990 and 2010 (Ferrari et al., 2013c). There are
few national studies where this has been assessed, but general population
studies from the Netherlands, Canada, US and the UK have found no
increase in the prevalence of depressive disorders from the 1950s to 2010s
(Brugha et al., 2004, de Graaf et al., 2012, Kessler et al., 2005, Murphy et al.,
2000, Patten et al., 2015b). During this time, however, there have been
changes in both the diagnostic criteria and study methodologies, which
complicates the comparison of the studies.
2.4 RISK FACTORS FOR DEPRESSIVE DISORDERS

Depressive disorders are multifactorial, and both genetic and environmental
factors increase the risk of developing depression. With age, the importance
33
of environmental factors increases (Nivard et al., 2015). There is evidence of

a genetic risk, but efforts to identify the specific loci involved have so far
failed. It is possible that the genetic contribution is smaller than expected, or
that the disorder represents either different symptom clusters or different
genetic pathways to the same symptomatic manifestation (Colman and
Ataullahjan, 2010, Flint and Kendler, 2014, Kendler et al., 2013).
Environmental correlates of depressive disorders, in particular MDD, have

been studied extensively in cross-sectional studies. However, a factor or
characteristic may be associated with current depression via two ways: by
being a risk factor of incidence of the disorder; or by contributing to longer
duration and thus increasing prevalence of depression in people with that
characteristic (Lorant et al., 2003). To understand the mechanisms of
etiology and target preventive efforts, knowledge about the first type of
associations would be essential. This can only be reliably assessed in
longitudinal studies, which are more rare than cross-sectional studies.
In this literature review, primarily risk factors of incidence are reviewed.

When correlates of prevalence are discussed, this will be specifically
mentioned. The most important risk factors are summarised in Table 5.
2.4.1 SOCIODEMOGRAPHIC FACTORS
2.4.1.1 Gender
Depression is more common among women than men (Kessler and Bromet,
2013, Seedat et al., 2009) and this is due to its higher incidence among
women: women have an approximately 1.5 to 2-fold risk of developing
depression compared with men (Anthony and Petronis, 1991, De Graaf et al.,
2002, Eaton et al., 2001, Eaton et al., 2008c, Klein et al., 2013, Stegenga et
al., 2013, Wang et al., 2010a). However, the reverse is true in prepubertal
children, and the higher risk of women is observed only from puberty (12
years) onwards (Wesselhoeft et al., 2015). The incidence of depression in
women peaks at 20 years of age; and after 40 years of age, the incidence is
close to that of men (Pedersen et al., 2014).
Twin studies have been used to model the risk factors of major depression in
men and women. They suggest that the risk of major depression results from
three pathways in both men and women: internalising symptoms;
externalising symptoms; and psychosocial adversity. In men, losing a parent
as a child, low self-esteem and genetic risks were more important
explanatory variables than in women. The authors pointed out, however, that
the pathways in both genders were similar, and that the differences are of
less importance than the similarities (Kendler et al., 2002, 2006). In a
34
further study, the same group noted that for women, problems in caring
relationships and interpersonal issues (such as lack of parental warmth,
divorce, social support and marital dissatisfaction as well as neuroticism)
were more important risk factors than for men (Kendler and Gardner, 2014).
For men, stressful life events of financial, occupational or legal nature, as
well as childhood sexual abuse, conduct disorder, drug abuse, and a prior
history of depression were more important factors than for women.
Notably, the World Health Survey shows that the relative risk of women to
men increases with increasing country income (Rai et al., 2013). On the other
hand, in the World Mental Health Survey, the difference between men and
women decreased in the younger cohorts (Seedat et al., 2009). They also
found gender differences to be smaller in countries and cohorts with greater
equality between male and female roles. An exception among countries is
China, where some studies have found no gender differences in prevalence of
depression (Bromet et al., 2011, Lee et al., 2009). A recent meta-analysis,
however, did show a typical pattern of female dominance in depression (Gu
et al., 2013).
2.4.1.2 Age
Depression is more prevalent in younger age groups (Kessler and Bromet,
2013). In longitudinal studies, younger age has been established as a risk
factor for developing depression (Eaton et al., 2008c, Stegenga et al., 2013,
Wang et al., 2010a). This association, however, is less clear, or sometimes
even reversed, in low-income countries (Kessler and Bromet, 2013). When
incidence of depression was examined in a register-based study in Denmark,
it increased steeply from 10 years of age until 20 and then reduced, until
there was another peak at 80-90 years of age (Pedersen et al., 2014).
However, these findings should be interpreted with some caution, as there

are challenges related to diagnosing depression among older persons using
structured clinical interviews. Comparison of the CIDI, GHQ and K-10 in the
Australian Mental Health Survey demonstrated that inconsistencies between
the instruments rose with age (O'Connor and Parslow, 2009). It was
interpreted that the complex questions of the CIDI may confuse older people
and thus lead to an underestimation of mental disorders in the elderly
(O'Connor and Parslow, 2010).
2.4.1.3 Socioeconomic position

There is a well-known cross-sectional association between lower
socioeconomic position (SEP) and depressive disorders (Fryers et al., 2003,
Lorant et al., 2003, Pulkki-Råbäck et al., 2012). A systematic review of 51
35
prevalence studies found an OR 1.8 for prevalence of depression among the

lowest versus the highest socioeconomic group.
However, the causality of the association is a subject of debate, and both

causal pathways have evidence and a credible theory behind them. These are
called the “causation” theory, where low SEP is a risk factor for depression,
and the “selection” theory, where depression either prevents upward social
mobility or causes downward movement (Dohrenwend et al., 1992). In line
with the “causation” theory, the stress hypothesis suggests that individuals
with high SEP have better personal resources to cope with stress, such as
self-esteem, coping style and locus of control, and are less prone to develop
depression because of these resources (Lorant et al., 2003).
It is important to note that three central confounders of the relationship

between SEP and depression are gender, age and physical diseases. The latter
is often ignored in the literature (Lorant et al., 2003).
A meta-analysis of five longitudinal studies found only a slightly elevated risk

of depression in the lowest SEP group (OR 1.2), whereas the risk of prevalent
depression and persistence of disorder were higher (OR 1.8 and 2.0,
respectively) (Lorant et al., 2003). This was interpreted as supporting the
“causation” theory, and in particular the “stress” theory of higher SEP as a
protective factor, and in the event of falling ill with depression, a factor
promoting recovery. Many individual longitudinal studies have found
poverty to be associated only with persistence, not onset of depression
(Skapinakis et al., 2006, Weich and Lewis, 1998), although some have found
an association with incidence of depression (Kosidou et al., 2011).
It is most likely, however, that both processes occur simultaneously and

interact (Lorant et al., 2003). A further question, in context of the causation
theory, is whether the influence of lower social status on adult depression
results from lower SEP in adulthood, or whether the influence is a result of
adversity earlier in life (Muntaner et al., 2004). Again, it is likely that lower
SEP both during childhood and adulthood increase the risk of depression.
Of the different indicators of socioeconomic status, associations have been

found with educational status and employment status, as well as income or
material assets (Fryers et al., 2003). The relationship with education might
be weaker and in many cases non-existent (Andersen et al., 2009, Kosidou et
al., 2011). The Canadian NPHS study found very differing results according
to employment status: lower education was associated with increased risk of
developing depression among those who worked, but lower risk among those
who did not work (Wang et al., 2010b).
36
In addition, financial hardship, such as having to go without meals or
heating, is associated with current depression, independently of other
measures of SEP (Butterworth et al., 2009, Dijkstra-Kersten et al., 2015,
Lahelma et al., 2006, Wang et al., 2010b, Weich and Lewis, 1998). However,
the association is cross-sectional, and it is unclear whether it predicts new-
onset depression, with both positive findings (Skapinakis et al., 2006, Wang
et al., 2010b, Weich and Lewis, 1998) and negative ones (Butterworth et al.,
2009, Dijkstra-Kersten et al., 2015) from longitudinal studies. The causality
here may be in the other direction: people suffering from depression may be
more likely to experience either true or perceived financial strain.
2.4.1.4 Marital status

Depression is more prevalent among those who are divorced, separated or
never-married (Kessler and Bromet, 2013). In the WMHS, being married was
associated with reduced risk of onset of depression in men, but not in women
(Scott et al., 2010). Being separated or divorced was associated with
increased risk in both genders. Some other studies have also found an
increased risk of developing depression among the unmarried (Anthony and
Petronis, 1991), or specifically, those who separated during the follow-up
period (Lorant et al., 2007).
However, many studies have not found a significant association between

marital status and the risk of developing depression. Therefore, it may be
that marital status influences the course of illness, or that the causal pathway
is reversed.
2.4.2 FAMILY HISTORY AND EARLY LIFE EXPERIENCES
2.4.2.1 Family history

A meta-analysis of the genetic epidemiology of depression found a 2.8-fold
risk in first-degree relatives of people with major depression. The heritability,
or the proportion of variation in the population explained by genetic
variation, is estimated at 37% (Sullivan et al., 2000).
2.4.2.2 Early life experiences

Many factors and experiences during childhood and adolescence have been
established as risk factors for developing depression (Kendler et al., 2002)
(Kendler et al., 2006). The risks start accumulating from the very beginning
of life: low birth weight, but not premature birth, is a risk factor for
depression (Loret de Mola et al., 2014).
37
Childhood adversities – such as low parental education (Park et al., 2013,

Ritsher et al., 2001), parental depression (Eaton et al., 2001, Klein et al.,
2013) or any parental mental disorder (Pirkola et al., 2005a, Stegenga et al.,
2013), exposure to bullying (Bowes et al., 2015, Sourander et al., 2015) and
family discord (Pirkola et al., 2005a), or physical, emotional or sexual abuse
(Chen et al., 2014, Stegenga et al., 2013) – are well-known risk factors for
depression. The accumulation of several adversities increases the risk
(Elovainio et al., 2015, Pirkola et al., 2005a). The impact of childhood
adversity might be more apparent in women than in men (Veijola et al.,
1998).
However, childhood adversities and other early life experiences are

frequently ascertained retrospectively, and people often report them
inconsistently. Recently, it was demonstrated that current depression and
stress were associated with remembering childhood adversities that the
individual had not reported in an earlier interview, and not forgetting the
adversities previously reported, both indicating memory bias related to
current depression (Colman et al., 2015).
Finally, threatening and other stressful events during adult life increase the
risk of incident depression (Lehtinen et al., 2005), and adult and childhood
stressful events are also associated with each other (Kendler et al., 2002).
When these events, such as financial problems, robbery or work problems,
are separated into those dependent on the individual’s behaviour and those
independent of that behaviour, both types predict depression (Kendler et al.,
2002, 2006).
2.4.3 SOCIAL CAPITAL
Social capital refers to the collective value of the social networks possessed by
an individual. In a Finnish study, factor analysis was used to condense
several variables indicating aspects of social capital into three dimensions:
social support; social participation and networks; and trust and reciprocity
(Nieminen et al., 2008). The same dimensions have also been identified in
other studies and expert groups (Zukewich and Norris, 2005). There was a
strong association between trust and participation dimensions and
psychological well-being (Nieminen et al., 2010).
There is extensive literature on the association between social capital and

mental health, and in particular depression, at both the individual (De Silva
et al., 2005, Forsman et al., 2012, Jones et al., 2014, Nyqvist et al., 2013) and
population level (Smith and Kawachi, 2014). A summary measure of low
social integration correlated with increased suicide risk among women in the
38
US (Tsai et al., 2015). With few exceptions, these associations have been
established cross-sectionally.
One study looked specifically at social capital and risk of depression in a

prospective setting and found that the increased risk of depression among
people with low social trust was explained by baseline depression (Fujiwara
and Kawachi, 2008). The Dutch NESDA study also did not find any measure
of social support (having a partner, network size, perceived emotional
support) to predict new depressive episodes when controlled for personality
features and clinical characteristics (Noteboom et al., 2015). However, other
longitudinal studies have identified social isolation (Kaplan et al., 1987), lack
of social support (Stegenga et al., 2013) and decreased social participation
(Kivelä et al., 1996) to predict the onset of depression.
2.4.4 CHRONIC SOMATIC CONDITIONS AND HEALTH BEHAVIOURS
Many chronic diseases have high comorbidity with depression. Having a

chronic physical disease predicts the onset of depression (Kaplan et al., 1987,
Patten, 2001, Stegenga et al., 2013, Wang et al., 2010a). In addition to
chronic diseases, traumas, such as traumatic brain injuries, also increase the
risk of depression (Perry et al., 2016). People with chronic somatic disorders
may also have a longer course of illness. This, together with increased
incidence, results in a higher prevalence of depression among people with
chronic somatic disorders (Patten, 2005). The association may be directly
causal, as in the case of hypothyroidism, or it may be that the experience of
living with a chronic illness contributes to developing depression. Common
causal pathways, such as inflammation, have also been investigated (Kiecolt-
Glaser and Glaser, 2002). The association between chronic somatic
conditions and depression is bidirectional, so that depression is also a risk
factor for many somatic conditions, and worsens their prognosis (see 2.6.5).
In a recent meta-analysis, both obesity (adjusted OR 1.6) and being

overweight (adjusted OR 1.08) predicted the onset of depression (Luppino et
al., 2010). High consumption of processed foods is also a risk factor for later
development of depressive symptoms, whereas a diet rich in vegetables, fruit
and fish is a protective factor (Akbaraly et al., 2009). The Mediterranean diet
and consumption of fruits and nuts, unsaturated fatty acids and legumes, in
particular, seems to be protective against depression (Sanchez-Villegas et al.,
2009). In addition, leisure time physical activity unrelated to occupational
activity appears to protect against psychological distress in men (Wiles et al.,
2007). A systematic review of diet and depression confirmed these findings,
but methodological limitations of the studies and their limited evidence base
raise cautions against making firm conclusions (Rahe et al., 2014).
39
Both alcohol use and alcohol use disorder (AUD) are associated with an
increased risk of depression, of up to twofold in diagnosed AUDs (Boden and
Fergusson, 2011). Cannabis use is also associated with an increased risk of
depression, and there is a dose-response where heavy use (OR 1.6) implies a
higher risk than light use (OR 1.2) (Lev-Ran et al., 2014).
Finally, the association between smoking and depression is strong and well-
known. A systematic review of longitudinal studies concluded that the risk of
developing depression among adolescent smokers was 1.7-fold, and the risk
of taking up smoking among depressed adolescents 1.4-fold (Chaiton et al.,
2009). Also among adult smokers, the risk of developing depression is
twofold, and smoking cessation could reduce this risk (Bakhshaie et al., 2015,
Pasco et al., 2008). A twin modelling study concluded that there is both a
causal relationship between nicotine dependence and development of
depression and a shared genetic risk (Edwards and Kendler, 2012).
2.4.5 OTHER PSYCHIATRIC DISORDERS AND PERSONALITY TRAITS
Because there is some diagnostic overlap and common risk factors between
depressive and other psychiatric disorders, it is understandable that the
latter are also risk factors for developing depression. Anxiety disorders in
particular are a risk factor for later onset of depression (Eaton et al., 2008c,
Klein et al., 2013, Stegenga et al., 2013). Comorbidity with anxiety disorders
may vary according to the diagnostic thresholds used, with lower diagnostic
thresholds for MDD increasing comorbidity (van Loo et al., 2015b).
Personality traits such as neuroticism also increase the risk of depression (De
Graaf et al., 2002, Noteboom et al., 2015). A low sense of coherence, harm-
avoidance and low self-esteem are associated with increased incidence of
depression (Lehtinen et al., 2005, Luutonen et al., 2011, Miettunen et al.,
2012, Miettunen and Raevuori, 2012).
Subclinical depressive symptoms and general psychological distress (Ernst et

al., 1992, Horwath et al., 1992, Klein et al., 2013, Skapinakis et al., 2006), as
well as sleep problems (De Graaf et al., 2002, Riemann and Voderholzer,
2003) indicate a risk of developing depression. It can be questioned whether
these represent an independent risk factor or a marker of a premorbid state
of depressive disorder.
Finally, it is important to note that accumulation of several risk factors

further increases the risk of depression (Meng and D'Arcy, 2014).
40
Table 5. Established risk factors of depressive disorders in longitudinal population studies
Genetic and early life

(Anthony and Petronis, 1991,
De Graaf et al., 2002, Eaton et
Biological factors Female sex al., 2001, Eaton et al., 2008c,
Klein et al., 2013, Stegenga et
al., 2013, Wang et al., 2010a)
Genetic risk (Flint and Kendler, 2014)
Early life and Low parental education and occupational (Park et al., 2013, Ritsher et al.,
family position 2001)
Physical or sexual abuse (Stegenga et al., 2013)
(Eaton et al., 2001, Klein et al.,
Family history of MDE 2013, Stegenga et al., 2013,
Wang et al., 2010a)
(Eaton et al., 2008c, Stegenga
Younger age
et al., 2013, Wang et al., 2010a)
Adulthood
Neuroticism (De Graaf et al., 2002,
Personality
Noteboom et al., 2015)
Harm avoidance (Miettunen et al., 2012)
Socio-economic (Anthony and Petronis, 1991,
Unemployment
position Stegenga et al., 2013)
(Lorant et al., 2007, Skapinakis
et al., 2006, Stegenga et al.,
Financial strain
2013, Wang et al., 2010b,
Weich and Lewis, 1998)
(Wang et al., 2010b)(in the
Low education working population) (Stegenga
et al., 2013)
Low income (among non-working (Wang et al., 2010b)
population and working men)
(Anthony and Petronis, 1991,
Social capital and
Separated or divorced Lorant et al., 2007, Stegenga et
relationships
al., 2013)
Social isolation (Kaplan et al., 1987)
Low social participation (Kivelä et al., 1996)
(De Graaf et al., 2002, Eaton et
Negative life events
al., 2001)
Problems with neighbourhood (Stegenga et al., 2013)
Low social support (Stegenga et al., 2013)
(Ernst et al., 1992, Horwath et
Other psychiatric al., 1992, Klein et al., 2013,
Depressive symptoms
disorders Skapinakis et al., 2006,
Stegenga et al., 2013)
Sleep problems (De Graaf et al., 2002)
Anxiety disorders, substance use and e.g. (Eaton et al., 2008c,

other psychiatric disorders Stegenga et al., 2013)
Poor physical (Kaplan et al., 1987)
Poor perceived health
health
(Kaplan et al., 1987, Patten,
Chronic somatic conditions 2001, Stegenga et al., 2013,
Wang et al., 2010a)
41
2.5 PROGNOSIS OF DEPRESSIVE DISORDERS

The average duration of a major depressive episode in population studies is
3-4 months (Eaton et al., 2008b, Kessler et al., 2003, Spijker et al., 2002),
but 15-20% have a chronic course of illness, lasting years (Eaton et al.,
2008b, Spijker et al., 2002). Based on an extensive literature review, the
GBD study modelled the average episode duration of 37.7 weeks (8.7
months) for MDD (Ferrari et al., 2013a).
Dysthymia is a long-term illness. Although its duration is less well known

than MDE, the median time to recovery is counted in years, not months
(Klein et al., 2006). In its review, the GBD study found insufficient data to
model duration and remission (Charlson et al., 2013).
2.5.1 KEY CONCEPTS OF PROGNOSIS
2.5.1.1 Remission, recovery and recurrence
In a landmark paper from 1991, a task force of the MacArthur Foundation

Research Network on the Psychobiology of Depression (Frank et al., 1991),
proposed definitions of key concepts of major depressive disorder as follows:
Remission: a brief period (x to z days) during which the individual no

longer meets syndromal criteria for the disorder and has only minimal
symptoms.
Recovery: Remission that has lasted z+1 days or longer. The concept
refers to recovery from the episode, not the illness itself.
Relapse: Return of symptoms that meet the syndrome criteria during the
period of remission, but before recovery. Conceptually a relapse
represents the return of the symptoms of a still ongoing episode.
Recurrence: Appearance of a new episode of the disorder during recovery.
The concept of “asymptomatic” can be operationalised as having 8 points or

less on the BDI (21 questions) scale; 7 or less on the Hamilton Rating Scale
for Depression; or a maximum of 2 symptoms on the Schedule for Affective
Disorders and Schizophrenia (Frank et al., 1991).
Frank et al. (1991) provide various suggestions for the time limits of
remission and recovery. Suggested minimum duration for remission is 2 to 3
weeks, whereas the maximum duration, and limit to recovery, could be 2-6
42
months. In a later analysis, 2 months was considered too short, as more than
half of those in remission at 2 months still experienced a relapse within 18
months. 4 or 6 months was recommended as the limit between remission
and recovery (Furukawa et al., 2008).
In the case of dysthymia, the GBD 2010 study defined remission as no longer
meeting diagnostic criteria over a follow-up period of minimum 2 years
(Charlson et al., 2013), but the concepts are less clearly defined than for
MDD.
2.5.1.2 Residual symptoms and psychosocial disability

Residual symptoms or subthreshold symptoms refer to symptoms of
depressive disorder that persist after an episode, when either the number or
intensity of symptoms no longer meet diagnostic criteria. Sub-threshold
symptoms are also associated with psychosocial disability (Judd et al.,
2000).
Lowered psychosocial functioning is frequently found after a depressive

episode, which could, in theory, result from three sources: residual
symptoms; a permanent scarring effect; or a trait vulnerability that pre-
existed the depressive episode. Ormel et al. (2004) tested these hypotheses
and found evidence both for trait effect, meaning that the premorbid
psychosocial functioning was already lower among persons who later
developed depression, and state effect, meaning that psychosocial
functioning worsened significantly during the depressive episode and was
related to symptom severity. Scarring, a low psychosocial functioning in the
absence of depressive symptoms, only occurred in severe recurrent
depression.
In addition, cognitive impairment in the fields of executive function and

attention often persists after remission from depression (Rock et al., 2014),
although the cognitive impairment resulting from depression may be smaller
in population samples than in clinical studies (Castaneda et al., 2008).
Impaired cognition may in turn cause problems in psychosocial functioning
(Papakostas and Culpepper, 2015).
2.5.2 REMISSION, RECOVERY AND RECURRENCE IN POPULATION

STUDIES
Many of the longitudinal studies of general population samples that examine

prognosis of depressive disorders were carried out several decades ago with
now outdated methodology. In the Stirling County study, which began in
43
1952 and was followed up for 17 years, 75% of cases with baseline depression
had a poor outcome or subsequent episodes, and 26% had a chronic course of
illness (Murphy et al., 1986). The Zurich cohort study followed up a cohort of
19 to 20-year-olds, and found that 47% of those with baseline depression
received no diagnosis at any of the follow-ups during seven years (Vollrath
and Angst, 1989). The Upper Bavarian Longitudinal Community Study
followed a rural population in Bavaria, Germany, and found that, after 25
years, 20% of those with pure depression at baseline still had depression, and
73% had no depression or anxiety disorder (Fichter et al., 2010). A Swedish
study with a very long follow-up of 30-49 years, found a recurrence rate of
40%, and transition to other diagnoses in 21% of the sample (Mattisson et
al., 2007).
In the more recent population studies, the prognosis seems somewhat better.
In the Netherlands, 21% of persons with pure (not comorbid) depressive
disorder, and 35% of those with comorbid disorder, still had a depressive
disorder at the seven-years follow-up (Rhebergen et al., 2011). In Canada,
77% of persons with MDE had recovered by the 2-year follow-up (Fuller-
Thomson et al., 2014).
Recovery rates in clinical studies have generally been lower than in

population studies, with 45-50% at 2 to 5-year follow-up. (Holma et al.,
2008, Penninx et al., 2011, Stegenga et al., 2012). In a study of Finnish
primary care patients with MDD, the median time to remission was 20
months (Riihimäki et al., 2014).
Also, when symptom scores such as the BDI are used as outcome measures,
recovery rates are lower. Two studies in general populations in Finland found
recovery, defined as non-symptomatic in the BDI scale, in 35% of the sample
over a 2-year follow-up and in 46% over 9 years (Dowrick et al., 2011,
Viinamäki et al., 2006b).
It is commonly believed that approximately 50% of persons with a first

depressive episode will have a recurrent episode. This figure may be lower in
the general population and higher when specialised mental health care
settings are studied (Hardeveld et al., 2010). In a Dutch general population
sample, 13% had a recurrent episode within five years; 23% in 10 years; and
42% in 20 years (Hardeveld et al., 2013).
Dysthymia is considered a chronic, as opposed to episodic, disorder. In a 10-

year follow-up of patients with dysthymia, 74% achieved recovery, but 71% of
them had a relapse. Median time to recovery was 52 months (Klein et al.,
2006). In a general population study, 65% of persons with initial chronic
depression had recovered at three years (Agosti, 2014).
44
2.5.3 QUALITY OF LIFE IN PEOPLE RECOVERED FROM DEPRESSION
During the symptomatic phase, persons with depressive disorders experience
significant reductions in health-related quality of life (Saarni et al., 2007). It
is unclear how much of this reduction persists after symptomatic recovery
from the disorder. In a 6-year follow-up study, patients with initial
depression achieved normal mood, functional capacity and life satisfaction
(Koivumaa-Honkanen et al., 2008). On the other hand, in another 6-year
follow-up study, women who recovered from depression continued to have
lower health-related quality of life in the fields of social functioning and pain.
Some of the association was mediated by sleep disturbance, which was an
independent predictor of low health-related quality of life (Joffe et al., 2012).
Finally, in an elderly population, there was a surprising temporal association
between reduction in HRQoL and subsequent depressive symptoms, but not
in the opposite direction (Hajek et al., 2015).
2.5.4 PREDICTORS OF DIFFERENT NEGATIVE OUTCOMES
Knowledge of factors that predict negative outcomes, such as chronicity,

recurrence, higher disability or mortality in depressive disorders, could in
theory be used to target more intensive treatment interventions to these
high-risk groups. Attempts have been made to find markers that predict
treatment response to antidepressants, but have so far produced few results
that are applicable to clinical practice (Kuk et al., 2010).
2.5.4.1 Individual characteristics

The evidence on the impact of age and gender on prognosis of depressive
disorders is inconclusive. Some studies have found younger persons to have a
higher risk of non-recovery or recurrence (Eaton et al., 2008b, Holma et al.,
2008) while older age appears to be an indicator of chronicity in dysthymia
in particular (Agosti, 2014, Klein et al., 2008, Penninx et al., 2011). Some
studies have found women to have a worse prognosis (Spijker et al., 2001),
but most have found no difference between genders. Being single (divorced,
separated, never-married or widowed) is a consistent predictor of
unfavourable outcomes (Agosti, 2014, Eaton et al., 2008b, Mueller et al.,
1999). Worsening of the economic situation during the follow-up is also
associated with persistence of depressive symptoms (Viinamäki et al.,
2006a).
Childhood adversities, in particular different forms of abuse, have been

associated with longer duration of depressive symptoms, persistence and
recurrence (Agosti, 2014, Dowrick et al., 2011, Fuller-Thomson et al., 2014,
Gilman et al., 2013, Klein et al., 2008, Nanni et al., 2012, Rhebergen et al.,
45
2011). Possible explanations for this, as suggested by Gilman et al. (2013), are
that psychosocial stressors and their sequelae increase vulnerability to the
impact of stress on psychopathology, and may also reduce the effectiveness of
psychological treatments. Accumulation of several traumatic events during
adult life is also associated with persistence of depressive symptoms
(Tanskanen et al., 2004). Family history of depression, which may indicate
both genetic risk and impact on childhood experiences, is another risk factor
of persistence of depression (Dowrick et al., 2011).
Despite the well-documented cross-sectional relationship between social

capital and depression, few studies have investigated the impact of social
capital on prognosis of depression. In a clinical setting, low social support
and adverse life events were associated with more depressive symptoms at
follow-up (Leskelä et al., 2006).
2.5.4.2 Disorder characteristics

The severity of the disorder is closely related to its outcomes, with more
severe disorders (measured with symptom scales or according to DSM-
defined diagnostic criteria) and those with psychotic features being more
persistent (Penninx et al., 2011, Spijker et al., 2001, Spijker et al., 2002,
Viinamäki et al., 2006a). Psychiatric comorbidity, especially with anxiety and
personality disorders, is also related to worse outcomes (Agosti, 2014, Holma
et al., 2008, Klein et al., 2008, Murphy et al., 1986, Patten et al., 2010,
Penninx et al., 2011, Rhebergen et al., 2011, Spijker et al., 2001, Spijker et al.,
2002, Viinamäki et al., 2006a). Personality traits such as neuroticism also
predict worse outcomes (Rhebergen et al., 2009, Spijker et al., 2001).
Finally, cognitive symptoms of depression, such as memory impairment and
attention difficulties, have an adverse impact on the course of illness
(McIntyre et al., 2013, Papakostas, 2014, Trivedi and Greer, 2014).
Specifically, autobiographical memory deficits have been associated with
recurrent depression (Talarowska et al., 2016).
In terms of recurrence, the number of earlier episodes is the strongest

predictor of a recurrent episode (Bulloch et al., 2014, Dowrick et al., 2011,
Hardeveld et al., 2010). Regarding subtypes of depression (melancholic and
atypical), it appears that severity is a more important predictor of outcomes
than the subtype, although suicidal thoughts may persist longer in the
melancholic subtype (Lamers et al., 2016).
46
2.6 EXCESS MORTALITY IN DEPRESSIVE, ANXIETY
AND ALCOHOL USE DISORDERS
Most psychiatric disorders have increased mortality, and the pooled all-cause
mortality risk for all mental disorders is twofold compared to the general
population (Walker et al., 2015). Two-thirds of these deaths are due to
natural causes, and one-third to unnatural or unknown causes. With 14% of
all global deaths attributable to mental disorders, they rank among the most
important causes of mortality, contrary to what was previously believed
(Walker et al., 2015). Most of the evidence is from high-income countries,
but identical findings in terms of mortality risk and years of life lost are
reported from low-income countries (Fekadu et al., 2015).
In a meta-analysis, the median of years of life lost due to mental disorders is

10 years (Walker et al., 2015), but higher figures, from 10-20 years, have
been reported for any psychiatric disorder (Chang et al., 2011, Wahlbeck et
al., 2011) and up to 28.5 years for schizophrenia (Olfson et al., 2015). In
Finland, the gap in schizophrenia is somewhat smaller, at 16-17 years
(Laursen et al., 2013).
In a recent meta-analysis, the risk was highest for psychotic disorders (RR
2.5) and lowest for anxiety disorders (RR 1.4) (Walker et al., 2015).
Significantly higher risks have been reported for instance in personality
disorders (SMR 5-6) (Björkenstam et al., 2015) and schizophrenia (SMR 3.7)
(Olfson et al., 2015). However, the increased mortality is observed in all
categories of mental disorders (Harris and Barraclough, 1998).
Fortunately, there are indications that, in the Nordic countries, the mortality
gap between people with mental disorders and the general population is
decreasing (Gissler et al., 2013, Wahlbeck et al., 2011).
2.6.1 EXCESS MORTALITY IN MAJOR DEPRESSIVE DISORDER
Four recent extensive reviews have assessed excess mortality in major

depressive disorder. First, the Global Burden of Disease study, based on its
literature review, used an excess mortality risk of 1.9 in its epidemiological
modelling (Ferrari et al., 2013a). Second, a meta-analysis of 21 population-
based studies found an RR of 1.91 in MDD (Baxter et al., 2011). Third, in a
meta-analysis of 203 studies covering different mental disorders, the
mortality risk for depression was 1.7-fold (Walker et al., 2015).
A fourth large meta-analysis of 293 studies (Cuijpers et al., 2014a) found a

relative mortality risk of 1.64, which decreased to 1.52 when corrected for
publication bias. This was lower than the RR 1.8 reported in an earlier meta-
47
analysis of the same research team (Cuijpers and Smit, 2002). This study
found that the longer the follow-up and the better the quality of the study,
the lower the reported mortality risk. Cuijpers et al. therefore concluded that
there is an elevated mortality risk in MDD, but this may have been
exaggerated by publication bias and low-quality studies.
Many good-quality individual studies and earlier reviews have found hazard
ratios close to those reported in the recent meta-analyses, which are 1.5 to 1.9
(Eaton et al., 2008a, Leinonen et al., 2014, Mykletun et al., 2009, Penninx et
al., 1999). However, a large study of the Veterans Health Administration in
the US found only a 17% increased risk, which was considered partially
explained by the higher overall mortality of the studied group (Zivin et al.,
2015). Moreover, two large nationally representative community studies
from North America did not find any independent risk after adjusting for
sociodemographic differences, health behaviour and somatic health status
(Everson-Rose et al., 2004, Patten et al., 2011).
Most of the deaths are due to natural causes, but the relative risk is higher for
unnatural deaths (Hiroeh et al., 2008, Walker et al., 2015). Increased risk
has been found for circulatory diseases, respiratory diseases, diabetes,
influenza and septicaemia (Leinonen et al., 2014, Nabi et al., 2010a, Zivin et
al., 2015). Of unnatural deaths, the risk is increased not only for suicide
(Leinonen et al., 2014, Nock et al., 2009, Zivin et al., 2015), but also other
unnatural causes, such as homicide and accidents (Crump et al., 2013a, b).
Cancer mortality appears not to be increased among the depressed (Leinonen
et al., 2014, Zivin et al., 2015).
It is unclear whether some subgroups of depressed people have higher

mortality than others. However, a recent meta-analysis showed a higher risk
among men (HR 2.0) than women (HR 1.6) (Cuijpers et al., 2014b). This was
also found in a meta-analysis of studies among the elderly (Saz and Dewey,
2001).
2.6.2 EXCESS MORTALITY IN DYSTHYMIA
There is a paucity of information about excess mortality in dysthymia. A

literature review from 1998 found four studies reporting mortality risk in
dysthymia. Combining the results, the overall risk of death was not
increased, but the risk of unnatural causes was 4.5-fold, whereas the risk of
natural death was 0.8-fold (Harris and Barraclough, 1998). More recently,
the literature review carried out for the Global Burden of Disease 2010 study
found only one meta-analysis, which consisted of two studies. No excess
mortality in dysthymia was found (RR 1.37 95% CI 0.93–2.00) (Baxter et al.,
2011).
48
2.6.3 EXCESS MORTALITY IN ANXIETY DISORDERS
Even though anxiety disorders are as common as depressive disorders, their

impact on mortality is much less studied. A meta-analysis of mental
disorders and mortality found a 1.4-fold risk of all-cause mortality in anxiety
disorders, the lowest among the mental disorders studied (Walker et al.,
2015). Another meta-analysis reports an RR 1.9 for panic disorder based on
four studies, and insufficient data for other anxiety disorders (Eaton et al.,
2008a). An earlier systematic review found two studies that reported no
increase in all-cause mortality, but significant increase in unnatural deaths in
anxiety neurosis (Harris and Barraclough, 1998). In panic disorder, all-cause
mortality was twofold.
Individual studies have found more mixed results. A population-based study

in Norway found no increase in mortality after controlling for confounders.
Interestingly, comorbid anxiety reduced the mortality risk in depression
(Mykletun et al., 2009). Similarly, a study of elderly with anxiety and
depression found no increased mortality associated with anxiety (Holwerda
et al., 2007). In the Netherlands, however, anxiety was associated with 1.7-
fold mortality risk in men but not in women (van Hout et al., 2004) in one
study, and 1.8-fold risk in women in another study (Denollet et al., 2009). In
the US, the risk was elevated for both men and women (Ostir and Goodwin,
2006). Of the specific anxiety disorders, generalised anxiety disorder (GAD)
is associated with an increased cardiovascular mortality risk, even when
controlled for comorbid depressive disorder (Martens et al., 2010).
2.6.4 EXCESS MORTALITY IN ALCOHOL USE DISORDERS
According to earlier meta-analyses, mortality in alcohol use disorders

(alcohol abuse and alcohol dependence) is 1.8 to 2-fold (Eaton et al., 2008a,
Harris and Barraclough, 1998). A recent meta-analysis of 81 studies found a
twofold risk in general population studies, and a 3.4 to 4.6-fold increase in
clinical samples (Roerecke and Rehm, 2013) Similarly, another new meta-
analysis of 39 studies found a high mortality risk ratio, of 3.45, compared to
the general population, without distinguishing between different samples
(Laramée et al., 2015). Few general-population studies have investigated
mortality in alcohol use disorders controlling for confounders. A 14-year
follow-up of a general population sample in Germany found 4.6-fold
mortality in women and 1.9-fold in men (John et al., 2013). In Finland,
mortality was 1.6-fold among persons with alcohol dependence and 20-fold
in alcohol-induced psychosis (Perälä et al., 2010).
49
2.6.5 POSSIBLE MECHANISMS OF EXCESS MORTALITY
Several mechanisms have been proposed to explain the excess mortality in

depression, such as increased suicide rates and other unnatural deaths,
unhealthy behaviours, biological dysregulation, poor access to or non-
compliance with medical treatment, and iatrogenic effects of psychiatric
medications. However, the quality of evidence for these mechanisms varies,
and even the causal directions of some of them are unclear (Cuijpers and
Schoevers, 2004, Thornicroft, 2011).
2.6.5.1 Suicide and other unnatural deaths

In a Swedish longitudinal study with 50 years follow-up, the lifetime suicide
risk for persons with affective disorders was 6%, much higher than the 0.3%
of the population without mental disorders (Holmstrand et al., 2015). In
men, the risk was 8.4-fold and in women 2.6-fold compared to the
population without mental disorders. In a Danish register-based study with
median follow-up of 18 years, 6.7% of depressed men and 3.8% of depressed
women committed suicide during the follow-up (Nordentoft et al., 2011).
Among people with depression, men, those with more severe depression,
hopelessness, previous suicide attempts and psychiatric comorbidities have a
higher suicide risk (Hawton et al., 2013). Suicides account for 12-13% of the
excess deaths among depressed outpatients (Moustgaard et al., 2013).
Suicide attempts are much more frequent than completed suicides. In a
follow-up of depressed patients treated in secondary care, there were 10
attempts per 100 patients annually, although many patients had more than
one attempt (Holma et al., 2010).
In addition to suicide, the risk of homicidal death is 2.6-fold and the risk of
accidental death 2.2 to 2.5-fold (Crump et al., 2013a, b). However, even
though the risk of unnatural deaths is increased, in absolute terms they
account for a minority of all deaths.
2.6.5.2 Depression and coronary heart disease

As discussed in 2.4.4, links between depression and many somatic diseases
are bidirectional. Depressed persons have a 30% increased risk of both
coronary heart disease and myocardial infarction (Gan et al., 2014). Among
patients with acute coronary heart syndrome, depression is an independent
risk factor for adverse outcomes such as all-cause and cardiac mortality
(Lichtman et al., 2014, Whooley and Wong, 2013). Both behavioural (e.g.
physical inactivity, medication nonadherence, smoking, poor diet and social
isolation) and biological mechanisms (autonomic nervous system activation,
systemic inflammation, activation of the HPA axis, stress-induced ischemia,
50
platelet activation, endothelial dysfunction and common genetic risk factors)
are listed as potential mechanisms. According to some trials, antidepressant
medication potentially improves cardiac outcomes (Whooley and Wong,
2013).
In addition to heart disease, MDD is a risk factor for other physical diseases.
In the WMHS study, MDD predicted onset of all of the ten somatic
conditions studied, including arthritis, diabetes, asthma, peptic ulcer and
cancer (Scott et al., 2015).
2.6.5.3 Hazardous health behaviours

The association with many hazardous health behaviours and depression is bi-
directional. These behaviours increase the risk of depression, as described in
2.4.4, and depression increases the risk of taking up or continuing such
behaviours.
As noted before, there is a two-way association between smoking and

depression. Depressed persons have a higher risk of taking up smoking
(Chaiton et al., 2009), and a 2-3 fold risk of relapse after quitting smoking
(Zvolensky et al., 2015). In addition, there might be risk factors that increase
the likelihood of both smoking and depression (Boden et al., 2010, Dierker et
al., 2002, Edwards and Kendler, 2012). Alcohol consumption is also higher
among depressed persons (Cuijpers and Schoevers, 2004). In fact, in a study
of 40-64-year old Finnish people, alcohol-related causes of death accounted
for 50% of the excess mortality in depressed men, and 30% in women
(Moustgaard et al., 2013).
Besides smoking, depression is associated with unhealthy diets, higher body-

mass index, metabolic syndrome, lower HDL cholesterol and reduced
physical activity, all of which directly contribute to mortality (Cuijpers and
Schoevers, 2004, Lehto et al., 2008a, Lehto et al., 2008b). Those with more
severe symptoms and comorbidities are more likely to experience major
weight gain (Heiskanen et al., 2013). All of these health behaviours are also
risk factors for depression (Luppino et al., 2010, Wiles et al., 2007).
It has been suggested that the link between depression and many of its
behavioural and other physical risk factors is a systemic inflammatory
process (Berk et al., 2013, Kiecolt-Glaser et al., 2015). Psychosocial stressors,
for example, increase levels of pro-inflammatory cytokines (Berk et al., 2013,
Kiecolt-Glaser et al., 2015). In addition, sleep deprivation, poor diet and lack
of exercise impact both immune function and systemic inflammation, and
obesity itself is a state of systemic inflammation (Berk et al., 2013). These
pathways are closely related to increased mortality, as discussed below.
51
2.6.5.4 Biological dysregulation

Three main biological mechanisms have been proposed to explain the higher
somatic morbidity and mortality among persons with depression: neuro-
immune dysregulation, hyperactivity of the HPA axis, and
sympathoadrenergic dysregulation. However, the causal and temporal
relationship between these mechanisms and depression is not entirely clear
(Cuijpers and Schoevers, 2004).
First, depression can stimulate the production of proinflammatory cytokines,

which may in turn increase the risk of cardiovascular disease, diabetes,
osteoporosis, arthritis and some cancers (Cuijpers and Schoevers, 2004,
Kiecolt-Glaser and Glaser, 2002). The relationship between inflammation
and depression is bidirectional (Kiecolt-Glaser et al., 2015) and is more
closely related to the physical symptoms of depression: tiredness, sleep
problems and lack of appetite (Jokela et al., 2015). Interventions that reduce
inflammation markers include cognitive-behavioural therapy, exercise and
meditation, and selective serotonin reuptake inhibitors (SSRIs) (Irwin and
Cole, 2011, Vogelzangs et al., 2012).
Second, hyperactivity of the hypothalamic pituitary adrenal (HPA) axis is

observed in 50% of depressed individuals, and in 80% of those severely
depressed (Anacker et al., 2011). The feedback regulation of the axis is
impaired, likely due to the abnormal function of the glucocorticoid receptors
(Anacker et al., 2011). As a result of the malfunction of this axis,
hypercortisolemia in the blood is associated with insulin resistance,
accumulation of intra-abdominal fat and lowered bone density (Cuijpers and
Schoevers, 2004). The causal relationship between abnormalities of the HPA
axis and depression has been unclear, and it has been suggested that both are
caused by another factor, such as stress or trauma (Cuijpers and Schoevers,
2004). However, increasing evidence suggests that malfunction of the HPA
axis may play a causal role in the development of depression: for example,
polymorphisms of the glucocorticoid receptors have been associated with
increased incidence of affective disorders (Cuijpers and Schoevers, 2004).
Glucocorticoid treatment sometimes induces depressive symptoms, and
elevated glucocorticoid levels result may reduce hippocampal neurogenesis
(Anacker et al., 2011). Differences between subtypes of depression have been
documented, showing that persons with atypical depression might suffer
from hypocortisolism, as opposed to the hypercortisolism seen in
melancholic depression (Gold, 2015).
Third, depression is associated with increased noradrenaline levels. The

sympathoadrenergic dysregulation and autonomic hyperactivity can lead to
reduced heart rate variability and increased risk of cardiovascular events
(Cuijpers and Schoevers, 2004).
52
2.6.5.5 Antidepressants
Concerns exist regarding the impact of antidepressants on mortality, and, in
particular, the tricyclic antidepressants that have potentially dangerous side
effects, including cardiac conduction abnormalities (Somberg and Arora,
2008). At the population level, mortality among antidepressant users is
higher compared to non-users, and increases further with use of other
psychotropic medications (Palmaro et al., 2015, Sundell et al., 2011).
However, in these register-based studies, users and non-users have different
rates of depression, which makes interpretation of the results difficult.
Among people with a previous suicide attempt, the use of selective serotonin
reuptake inhibitors (SSRI) is associated with decreased all-cause mortality
(RR 0.59) because of lower risk of cardiovascular and cerebrovascular deaths
(Tiihonen et al., 2006). On the other hand, among elderly depressed persons,
antidepressant users had a 1.7-fold mortality risk compared with non-users
(Coupland et al., 2011). Again, in this naturalistic study, use of
antidepressants could be an indicator of disorder severity, and the mortality
risk could be attributable to the disorder itself. In summary, methodological
challenges limit the conclusions that can be made regarding the impact of
antidepressants on mortality.
2.6.5.6 Health care utilisation

Even though people with depression have higher prevalence of many chronic
somatic conditions, they do not always receive adequate treatment for these
conditions (Fagiolini and Goracci, 2009, Lawrence and Kisely, 2010). People
with psychiatric disorders have fewer medical visits than the general
population (Cradock-O'Leary et al., 2002), and their use of primary care and
preventive services is particularly low (Salsberry et al., 2005). People with
mental problems report more problems accessing general medical care than
psychiatric care (Zeber et al., 2009). Reasons for not seeking help are related
to lack of recognition of symptoms and the reluctance to seek care, lack of
social support and isolation, and the cost of care, especially in settings where
access to health care is associated with employment (Fagiolini and Goracci,
2009, Zeber et al., 2009).
The care received may be of worse quality than that provided to the general
population. For example, patients with mental disorders and diabetes were
less likely to receive appropriate laboratory and eye exams, and were more
likely to be in poor glycaemic control than people without mental health
problems (Frayne et al., 2005). People with mental illness are also less likely
to receive invasive coronary interventions, and have increased cardiac
mortality following cardiac events (Kisely et al., 2007, Mitchell and
Lawrence, 2011).
53
The inequities in medical care received by people with mental disorders

occur partly due to lack of information and communication problems with
general healthcare workers, who might misattribute physical symptoms to
mental disorder (Fagiolini and Goracci, 2009). Negative stereotypes may also
lead to treating people with mental disorders less thoroughly (Thornicroft,
2011).
Finally, depressed persons are three times more likely to be nonadherent to

their medication (DiMatteo et al., 2000). Three possible explanations for this
are: 1) depressed persons are so hopeless that they do not develop positive
expectations of the treatment, which would be important to achieve
treatment adherence; 2) social isolation and the lack of family support that
often follow from depression may negatively influence adherence; 3)
problems in cognitive functioning and memory impairment in particular
might be obstacles to adhering to treatment.
2.7 SUMMARY OF THE LITERATURE REVIEW AND

GAPS IN KNOWLEDGE
While some issues regarding the prevalence, risk factors, prognosis and
mortality in depressive disorders are relatively well-established and
documented, the evidence is contradictory regarding others, and in some
cases, there are apparent gaps in knowledge.
Although it is clear that the annual prevalence of major depressive disorder is

approximately 5% with wide global variations, it is not known to what extent
non-participation in surveys influences prevalence estimates. There is little
reliable and recent information on the prevalence of dysthymia, another
important contributor to the burden of mental illness. It is also unclear
whether there have been changes in prevalence rates in the past decade, as
most studies on prevalence changes compare 1990s with the 2000s, or even
earlier time points.
Established risk factors of new-onset depressive disorders include female

gender, childhood adversities (although this may be influenced by memory
bias), chronic somatic conditions and other psychiatric disorders. The
findings are inconsistent regarding the impact of marital and socioeconomic
status, and very limited regarding social capital.
Prognosis of MDD in population surveys appears rather good, with three

quarters experiencing recovery. Less is known about the prognosis of
dysthymia. Some factors – single marital status and disorder severity – have
been consistently associated with worse outcomes, but results are mixed
regarding the impact of age, gender, socioeconomic position and other
54
factors. The broader spectrum of outcomes, including consequences
involving self-rated health, health-related quality of life and residual
symptoms, is less well documented than remission and recovery.
Finally, while the mortality risk in depressive disorders is estimated to be

twofold, more recent and better studies report lower estimates. It is unclear
whether the mortality risk has been exaggerated by poor-quality studies, and
whether the risk differs by gender. Much less is known about excess mortality
in anxiety and alcohol use disorders, even though both disorder groups are
major public health concerns.
55
Aims of the study
3 AIMS OF THE STUDY
The overall aim of this study was to examine prevalence, predictors and
different adverse outcomes of depressive disorders in a general population
setting.
The specific aims were to:

1. Establish the prevalence of depressive disorders (MDD and
dysthymia) in the Finnish population in 2011, and assess possible
changes in prevalence from year 2000 to year 2011 (Sub-study I)
2. Examine risk factors for new-onset depressive disorders in an 11-year

follow-up (Sub-study II)
3. Investigate the long-term prognosis of depressive disorders in a

population sample using various outcome measures, and examine
different predictors of adverse outcomes (Sub-study III)
4. Assess excess mortality associated with depressive, anxiety and

alcohol use disorders and psychological distress and depressive
symptoms, and describe the principal causes of death in a general
population (Sub-study IV)
56
4 MATERIALS AND METHODS
4.1 DATA SOURCES

Data from the following sources were used in the study: the Health 2000 and
Health 2011 population health surveys; the Finnish Care Register for Health
Care; the Finnish Causes of Death Statistics held by Statistics Finland.
4.1.1 HEALTH 2000 SURVEY
The Health 2000 Survey is a nationally representative survey of the Finnish

population (Heistaro, 2008). It was conducted in 2000-2001. A total of
8,028 adults aged 30 years and over (the adult sample) were sampled using
stratified two-staged cluster sampling. The sample frame was the population-
wide insurance database held at the Social Insurance Institution (Kela). In
the first stage, 80 health centre districts (clusters) out of a total of 249 were
sampled, 16 from each of the five main hospital districts of Finland. Out of
these 80 clusters, 15 represented the largest cities in Finland, and the
remaining 65 were sampled using systematic probabilities proportional to
size design. In the second stage, individuals from these 80 clusters were
sampled by systematic random sampling. Persons aged 80 and over were
oversampled (2:1). In addition, a sample of young adults aged 18-29
(n=1894) took part in the survey of young adults, which had a shortened
version of the study protocol (Figure 1, Table 6).
The adult sample study protocol consisted of a home interview, self-

administered questionnaires and a health examination. In total, 7419 persons
(93%) participated in at least some part of the study.
The study had the approval of the Ethics Committee of the Hospital District
of Helsinki and Uusimaa. Written informed consent was obtained from the
participants.
4.1.2 HEALTH 2011 SURVEY
The Health 2011 Survey is a follow-up study of the Health 2000 (Koskinen,
2012). The whole sample of the Health 2000 Survey, consisting of people
who were alive, living in Finland, and had not refused to participate, were
invited to take part. The young adults sample from Health 2000, of people
who were 29-40 years of age in 2011, was included. In addition, there was a
new young adults sample (18-29 years) in the survey, but this was not
57
Materials and methods
included in the current study. Data were collected between August 2011 and
June 2012.
The adult sample study population was 7,885 individuals, of whom 5,806
(74%) individuals participated in at least one part of the study (Figure 1,
Table 6). The study had the approval of the Ethics Committee of the Hospital
District of Helsinki and Uusimaa. Participants provided written informed
consent.
Figure 1 Participation and non-participation in the Health 2000 and Health 2011 Surveys
58
Table 6. Study population and register data used in the four sub-studies
Health 2000 Health 2011 Register data
Sub-study I 30 years and over, 30 years and over, Care Register for Health
n=8028 n=7885 Care (multiple
imputation)
Sub-study II 30-65 years without 41 years and over, Care Register for Health
current or past n=3862 Care (multiple
depressive disorder, imputation)
n=4057
Sub-study III 30 years and over with 41 years and over, Care Register for Health
12-month depressive n=5733 Care (multiple
disorder, n=392 imputation), Finnish
Causes of Death
Statistics (vital status)
Sub-study IV 30 years and over, - Finnish Causes of
n=8028 Death Statistics (vital
status and cause of
death)
4.1.3 REGISTER DATA
4.1.3.1 Care Register for Health Care (HILMO)

The Finnish Care Register for Health Care, a continuation of the Hospital
Discharge Register, is managed by the National Institute for Health and
Welfare. It covers all hospitalisations in public and private hospitals in
Finland. Since information on specialised outpatient care was available only
since 1994, it was not utilised in this study. In this study, Sub-studies I-III
used information from the HILMO register for multiple imputation.
Information on lifetime psychiatric hospitalisations of both participants and
non-participants of the Health 2000 sample was obtained from the
registerand used in the studies as presented in Table 7. Information was
available from 1969 onwards. Lifetime instead of past-year hospitalisations
were used because of the low number of past-year hospitalisations.
59
Table 7. Care Register for Health Care information used in Sub-studies I-III.
Variable Details Sub- Sub- Sub-

study study Study
I II* III
Any psychiatric Hospitalisations for any psychiatric disorder (F04- X X
disorder by 2000 F99; ICD-8 and ICD-9: 291-319) during 1969-2000
Schizophrenia Hospitalisations for schizophrenia (ICD-10: F20; X X
by 2000 ICD-8 and ICD-9: 295.0-295.3, 295.5, 295.6, 295.8,
295.9) during 1969-2000
Non-affective Hospitalisations for non-affective psychoses (ICD- X X
psychoses by 10: F22, F23, F24, F25, F28, F29; ICD-8 and ICD-
2000 9: 295.4, 295.7, 297, 298, 299) during 1969-2000
Depressive Hospitalisations for depressive disorders (ICD-10: X X
disorders by F32, F33, F341; ICD-9 (1987-1995) 2961, 3004A;
2000 ICD-8: 2960, 2962) during 1969-2000
Anxiety Hospitalisations for anxiety disorders (ICD-10: F40, X
disorders by F41, F42, F430, F431; ICD-9 and ICD-8: 3000,
2000 3002, 3003) during 1969-2000
Alcohol use Hospitalisations for alcohol use disorders (ICD-10: X X
disorders by F10; ICD-9: 303, 3050, 291; ICD-8: 303, 291)
2000 during 1969-2000
Dementia by Hospitalisations for dementia (ICD-10: F00-F03; X X
2000 ICD-9 and ICD-8 290) during 1969-2000
Any psychiatric Hospitalisations for any psychiatric disorder (F04- X X
disorder by 2011 F99; ICD-8 and ICD-9: 291-319) during 1969-2011
Schizophrenia Hospitalisations for schizophrenia (ICD-10: F20; X X
by 2011 ICD-8 and ICD-9: 295.0-295.3, 295.5, 295.6, 295.8,
295.9) during 1969-2011
Non-affective Hospitalisations for non-affective psychoses (ICD- X X
psychoses by 10: F22, F23, F24, F25, F28, F29; ICD-8 and ICD-
2011 9: 295.4, 295.7, 297, 298, 299) during 1969-2011
Depressive Hospitalisations for depressive disorders (ICD-10: X X X
disorders by F32, F33, F341; ICD-9 (1987-1995) 2961, 3004A;
2011 ICD-8: 2960, 2962) during 1969-2011
Anxiety Hospitalisations for anxiety disorders (ICD-10: F40, X X
disorders by F41, F42, F430, F431; ICD-9 and ICD-8: 3000,
2011 3002, 3003) during 1969-2011
Alcohol use Hospitalisations for alcohol use disorders (ICD-10: X X
disorders by F10; ICD-9: 303, 3050, 291; ICD-8: 303, 291)
2011 during 1969-2011
Dementia by Hospitalisations for dementia (ICD-10: F00-F03;x X X
2011 ICD-9 and ICD-8 290) during 1969-2011
*In Sub-Study II persons with hospitalisations due to depressive disorders prior to 2000 were
excluded from the analyses and only hospitalisations between 2000-2011 were utilised in MI
60
4.1.3.2 Mortality register (Sub-studies III-IV)
Mortality data, including dates and causes of death, was obtained from the
Causes of Death register from Statistics Finland (Finland, 2014). The data are
compiled from death certificates, which are completed by the treating
physician, or in the case of an unexpected or unnatural death, a medico-legal
officer based on a forensic autopsy. This is the case in 25-30% of deaths,
which is why Finnish mortality statistics are considered very reliable (Ylijoki-
Sorensen et al., 2014).
4.2 THE CIDI INTERVIEW

Psychiatric disorders in the Health 2000 and Health 2011 surveys were
diagnosed with the Composite International Diagnostic Interview, Munich
version (M-CIDI) (Wittchen and Pfister, 1997). When designing the Health
2000 Survey, this was considered the best computerised version of the CIDI
available. The Munich version was based on CIDI 2.1, the latest version at the
time.
The predecessor of CIDI is the Diagnostic Interview Schedule (DIS), the first
fully structured psychiatric diagnostic interview that could be administered
by trained lay interviewers (Andrews and Peters, 1998, Kessler and Ustun,
2004). The CIDI was developed by an international task force, supported by
the WHO and led by Dr. Robbins, the developer of the DIS. It was created in
response to a need for reliable psychiatric instruments based on ICD criteria
that could be used in population surveys to allow international comparison.
The CIDI became available in 1990 and was used in several large population
studies in its first years. The International Consortium in Psychiatric
Epidemiology (ICPE) was created by the WHO to compare the results. To
improve comparability of not only diagnoses, but also of risk factors, access
to treatment and other important aspects of mental illness, the ICPE formed
the WHO World Mental Health Survey Initiative and developed an improved
and expanded version of the CIDI, the WMH-CIDI. The new version included
modules covering functioning, treatment, risk factors, socio-demographic
correlates and methodological issues, in addition to psychiatric diagnoses
(Kessler and Ustun, 2004).
The Munich version of the CIDI (M-CIDI) was developed parallel to CIDI 2.1,
as there was a need to update the interview to correspond to DSM-IV
diagnostic criteria, as well as some technical revisions. The development was
done in collaboration with the WHO-CIDI consortium, and the M-CIDI
corresponds closely to the WMH-CIDI. The M-CIDI has shown good test-
retest reliability (κ values 0.68 for major depressive disorder and 0.70 for
dysthymia) in a sample of 60 persons aged 14-28 years, with a mean interval
between interviews of 39 days (Wittchen et al., 1998).
61
The different versions of the CIDI are commonly used in general population
mental health surveys (Kessler and Ustun, 2004), and international
comparability is its major strength. It has shown good concordance with
more thorough semistructured psychiatric interviews, such as the Structured
Clinical Interview for DSM (SCID) (Andrews and Peters, 1998, Haro et al.,
2006). In an international validity study of the CIDI 3.0 (Haro et al., 2006),
the sensitivity for any 12-month mood disorder was 69.1 and specificity was
97.2, positive predictive value 49.6 and negative predictive value 98.7, with
the SCID as comparison. In the National Comorbidity Survey (NCS-R),
validity of the Major Depressive Episode module of the CIDI was assessed,
with sensitivity of 54.6, specificity 94.7, total classification accuracy 90.7,
positive predictive value 64.1 and negative predictive value 97.5 (Kessler et
al., 2003).
The M-CIDI has been compared to clinician diagnoses with excellent results:
the sensitivity and specificity of the depression module were 95 and 100 for
lifetime single and 93 and 100 for recurrent depressive episode, and 100 and
85 for dysthymia, respectively, with corresponding Kappa values 0.96, 0.95
and 0.54 (Reed et al., 1998). The lower Kappa value for dysthymia was due
the M-CIDI not applying an optional hierarchy rule where a history of long
unremitted MDD outweighs diagnosis of dysthymia.
However, in the case of population studies, it can be argued that, in addition

to individual-level concordance, a measure of aggregate-level concordance
(i.e. agreement of prevalence rates of two different instruments) is important
(Kessler et al., 2004). In the case of CIDI 3.0, there was no statistically
significant difference between the 12-month prevalence rates produced by
SCID or CIDI interviews, whereas lifetime prevalence of MDD was 33%
higher when using SCID (Haro et al., 2006).
The full M-CIDI interview lasts about 90 minutes, and had to be cut down for
the Health 2000, as it was part of a more extensive general health survey. In
2000, six sections were included: anxiety disorders, depressive disorders,
mania, schizophrenia and other psychotic disorders, alcohol use disorders
and other substance-related disorders. In addition, interviewer observations
were recorded. In 2011, sections on mania and other substance-related
disorders were not included, and the section on psychotic disorders was
shortened. Altogether eight disorders were covered in both in 2000 and
2011: panic disorder, agoraphobia, social phobia, generalised anxiety
disorder, dysthymia, major depressive disorder, and alcohol abuse and
dependence. Only 12-month prevalence was determined, except for alcohol
use disorders, for which also lifetime prevalence was assessed. Diagnostic
criteria of the DSM-IV were used.
62
The M-CIDI was translated from its English version into Finnish for the
Health 2000 and pilot-tested. Test-retest reliability was assessed for the
depression and dysthymia modules, and the inter-rater agreement was
excellent with κ values 0.88 for both disorders and percentage of agreement
94-98% (Heistaro, 2008).
In Health 2011, the time available for the mental health interview was
reduced further, and modules had to be prioritised based on their public
health importance and experience from Health 2000. The sections assessing
manic symptoms and substance use other than alcohol were omitted. The
translation from Health 2000 was revised against the original German
version and the English translation. The new CIDI was piloted and all errors
were corrected before beginning the study. The interviewers, non-psychiatric
health professionals, received a two-day training. The interview was carried
out at the end of the health examination or, in some cases, during the home
interview. The mean duration was 21 minutes, ranging from 2.6 to 176
minutes.
In 2000, a total of 6,005 CIDI interviews (75% of the sample) were

conducted and in 2011, the total was 4,478 (57%). Altogether 3,584 people
were interviewed both in 2000 and 2011.
The Psychoses in Finland study was carried out parallel to the Health 2000
Survey to identify persons with psychotic disorders (Perälä et al., 2007).
Psychotic disorders were screened using the Composite International
Diagnostic Interview, self-reported diagnoses, medical examination, and
national registers on hospitalisations, medication and disability pensions.
Lifetime diagnosis of psychotic disorders was established with the Research
Version of the Structured Clinical Interview for DSM (SCID) and a review of
medical records (First, 1997). In this study, people with a psychotic disorder
were excluded from Sub-study II, and the results of Sub-studies III and IV
were adjusted for psychotic disorders. In Sub-study III, the impact of
comorbid psychotic disorder on the prognosis of depressive disorder was
studied.
4.3 STUDY VARIABLES
4.3.1 OUTCOME VARIABLES
4.3.1.1 Diagnostic status at follow-up (Sub-studies I-III)

The presence of depressive disorders in 2011, MDD and dysthymia, was used
as the primary outcome measure in sub-studies I-III. For both studies, these
63
were combined as a group, depressive disorders, containing people with

MDD, dysthymia or both.
Also, the presence of any depressive, anxiety or alcohol use disorder was
considered as an outcome in Sub-study III.
4.3.1.2 New-onset depression (Sub-study II)

To analyse risk factors for incident or new-onset cases of depressive
disorders only, people with a history of depressive disorder at baseline were
excluded to the extent possible. People were excluded who in the Health
2000 survey had received a 12-month diagnosis of MDD or dysthymia in the
CIDI interview, had ever been hospitalised for a depressive disorder, or
reported having received a depression diagnosis from a physician. People
with any psychotic disorder in 2000 were also excluded, as they were
considered to likely differ from the general population in terms of risk factors
for onset of depression.
4.3.1.3 Recovery and persistence (Sub-study III)

Recovery was defined as not meeting diagnostic criteria for either MDD or
dysthymia in the past 12 months. This is consistent with the definition of
recovery presented by Frank et al. (Frank et al., 1991) where recovery is
defined as having no symptoms or minimal symptoms for 4-6 months. We
did not, however, take into account residual symptoms measured with the
BDI, to allow comparison to other similar studies.
Persistence of disorder was defined as having a depressive disorder both at

baseline and follow-up. Since there was no information on the health status
between these two timepoints, it could not be established whether this was a
chronic, ongoing episode or a recurrent episode. The term “persistence” was
chosen to reflect both options.
4.3.1.4 Depressive symptoms and psychological distress (Sub-study

III)
The Beck Depression Inventory (BDI) was used to assess current depressive
symptoms both at baseline and at follow-up (Beck et al., 1961). The 21-item
version was used in 2000, and a shorter 13-item version, validated by Aalto
et al., was used in 2011 (Aalto et al., 2012). The total score was divided into
three categories, indicating no significant depressive symptoms, moderate,
and severe depressive symptoms as follows: 0-9, 10-18 and 19 points or more
in 2000, and 0-4, 5-8 and 9 or more points in 2011.
64
Current psychological distress was measured using the General Health
Questionnaire (GHQ), where a score of 4 or more indicated psychological
distress (Goldberg et al., 1997).
It is noteworthy that the CIDI diagnoses covered the 12 months preceding the
study, meaning that some persons had already recovered by the time of the
interview. In contrast, the BDI and GHQ measured current symptoms. Both
were part of a larger self-administered questionnaire sent to the participants
to be comleted prior to the physical health exam.
4.3.1.5 Health-related quality of life (Sub-study III)

Health related-quality of life at follow-up was measured with the EQ-5D
instrument, which was converted into an index score based on time-trade-off
values elicited in a UK general population sample (Kind et al., 1999).
4.3.1.6 Self-rated health (Sub-study III)

Self-rated health at follow-up was measured by asking the respondents to
assess their current health status by rating it as good, rather good, moderate,
rather poor or poor.
4.3.1.7 Mortality and causes of death (Sub-studies III-IV)

Mortality data was obtained from the Statistics Finland Causes of Death
Register, as described above. In Sub-study III, only information on vital
status at the beginning of Health 2011 data collection (4 July 2011) was used.
In Sub-study IV, date and cause of death were included in the analyses. The
causes of death were categorised into natural deaths (ICD-10 codes A00-
R99), suicides (X60- X84), homicides (X85-Y09) and other unnatural
deaths, such as accidents, injuries and poisonings (S00-T98, V01-X49 and
Y40-Y98). The natural causes of death were further categorised into four
classes: tumours (C00-D48), cardiovascular diseases (I00-I99), pulmonary
diseases (J00-J99), and other (A00-B49, D50-H95, K00-N99, O00-R99).
4.3.2 EXPOSURE AND CONFOUNDER VARIABLES AT BASELINE
4.3.2.1 Sociodemographic variables (all Sub-studies)

Sociodemographic information was obtained in a structured interview.
Educational level, based on the highest level completed, was divided into
65
three categories: basic (no high school or vocational training); secondary

(high school or completed vocational school); and higher (degree from a
higher vocational institution, polytechnic or university). Marital status was
categorised into married and cohabiting; and single, which included those
never-married, divorced or widowed. Family income was obtained from
registers of the Finnish Tax Administration, adjusted for family size, and
divided into quintiles.
4.3.2.2 Childhood adversity (Sub-studies II-III)

Childhood adversities were elicited using an 11-item questionnaire “When
you think about your growth years, i.e. before you were aged 16, did you…?”,
with answer options “yes”, “no” or “cannot say” (Pirkola et al., 2005a). The
items inquired were
1. Did your family have long-term financial difficulties?

2. Were your father or mother often unemployed although they wanted to
work?
3. Did your father or mother suffer from some serious disease or
disability?
4. Did your father have alcohol problems?
5. Did your mother have alcohol problems?
6. Did your father have any mental health problem, e.g., schizophrenia,
other psychosis, or depression?
7. Did your mother have any mental health problem, e.g., schizophrenia,
other psychosis, or depression?
8. Were there any serious conflicts within your family?
9. Did your parents divorce?
10. Were you yourself seriously or chronically ill?
11. Were you bullied at school?
The total number of reported adversities was counted and categorised into 0,
1-2 and 3 or more reported adversities, as in Kananen et al. (2010). In sub-
study II, parental mental health problems (items 6 and 7) were included in
the models separately, and not included in the summary variable. In sub-
study III, they were first explored separately in the unadjusted logistic
models, and then included in the summary variable in the adjusted logistic
models.
4.3.2.3 Social capital (Sub-studies II-III)

Social capital was conceptualised as a three-dimensional measure, assessing
social support, social participation and trust, based on the work by Nieminen
et al. (Nieminen et al., 2008). These dimensions concur with those proposed
66
by a broad international consensus (Zukewich and Norris, 2005). The
indicator was created using factor analysis to divide 39 variables into the
three chosen dimensions. The items included in the three dimensions are
listed in Table 8.
The dimension of trust included questions about feeling safe in the

neighbourhood, disappointments caused by people close to you, and cynical
mistrust, which was measured on an eight-item scale, shortened from the
Cook-Medley hostility scale (Nieminen et al., 2008).
Table 8. Items in the three dimensions of the social capital measure (adapted from
Nieminen et al. (2008))
Social support Social networks and Trust

participation
People on whose help you Club and society activities Feeling safe in neighbourhood
can count when you feel
exhausted
People you think really care Theatres, movies Feeling safe walking out alone
about you no matter what after 10 p.m.
People that can really make Studying Cynical mistrust (8 questions)
you feel better when you feel
down
People from whom you get Church and religious activities Been surprised by the
practical help when needed behaviour of people you
thought you knew well
Exercise, fishing, gardening People you counted on
etc. outdoor activities disappointed you
Handicrafts, playing music,
singing
Visiting family, friends or
neighbors
Having family, friends or
neighbours visit you
Talking on the phone
Joining in any health
promotion/discussion group
(11 questions about health
promotion)
4.3.2.4 Somatic comorbidity, smoking and obesity (Sub-studies II-IV)

Somatic comorbidity was assessed based on self-reported diseases that had
been diagnosed by a physician. At baseline, participants were asked whether
they had received diagnosis from a doctor for 43 different conditions, and in
some cases, to provide details on treatment and intensity of the condition.
For this study, 24 conditions were chosen based on relevance to mortality
and comorbidity, and the reliability of self-report (Saarni et al., 2006). These
were categorised into eight groups: cardiovascular diseases (heart failure,
67
myocardial infarction, coronary heart disease, hypertension and stroke);

pulmonary diseases (chronic obstructive pulmonary disease, chronic
bronchitis and asthma); neurological diseases (migraine, Parkinson's
disease); musculoskeletal disorders (rheumatoid arthritis, osteoarthritis,
back or neck disease requiring a visit to a physician in the past 12 months);
vision and hearing disorders (unoperated cataract, glaucoma, macular
degeneration, hearing loss, tinnitus), and other diseases (disturbing allergy
requiring a visit to a physician in the last 12 months, psoriasis, inflammatory
bowel disease and urinary incontinence). Cancer and diabetes were included
separately. In Sub-studies II-III, somatic comorbidity was categorised into 0,
1-2 and 3 or diseases, whereas in Sub-study IV, the eight categories were
included as such.
Participants were classified into three categories based on smoking status:

current smokers (smoked at least 100 times in their lifetime, smoked
regularly for the past year, and most recently during the past month), ex-
smokers, and non-smokers.
Body-mass index (BMI) was calculated based on height and weight

measurements at the health examination.
4.4 STATISTICAL METHODS

In all sub-studies, sociodemographic and other baseline characteristics were
compared among different subgroups using the chi-square test for
categorical and t-test for continuous variables. All tests were two-tailed, and
p-value of <0.05 was chosen to denote statistical significance. The stratified
two-stage cluster sampling was accounted for by using survey procedures of
the statistical software. Weights were used to adjust for the oversampling of
individuals aged 80 and over, where the analytic sample was not limited to
younger individuals.
The R statistical software (version 3.1 for Linux) was used to carry out
multiple imputation. Other analyses, including analysis of the imputed data
in Sub-studies I-III, were done using the Stata statistical software package
(version 11.2 for Windows).
4.4.1 MULTIPLE IMPUTATION
Participation in the Health 2011 was much lower than in Health 2000, at
73% vs. 93%, and participation in the CIDI, in particularl was low, at 75% vs.
57%. From previous studies, it was known that attrition is selective, and
persons with mental disorders have a lower likelihood of participating.
68
Therefore, this non-participation needed to be accounted for. Two methods
were compared: inverse probability weights (IPW) and multiple imputation
(MI). IPW are technically easier to apply and are more frequently used, but
the MI is more efficient and flexible (Härkänen et al., 2016, Li et al., 2015,
Mackinnon, 2010, Rubin, 1978).
Multiple imputation replaces the missing values by numbers based on

associations among the observed values in the data set (Li et al., 2015). This
process of generating the replacement values is repeated many times,
producing several copies of the original data sets, but with no missing values.
Then, the imputed data sets are analysed separately according to the desired
statistical analysis, such as regression, and finally the results are combined
into a single estimate. As opposed to other forms of imputation (such as
means or single imputation), it does not create false precision in the data.
In this study, register data on hospitalisations for psychiatric disorders, as

presented in Table 7, and information from the Health 2000 interview was
utilised in the multiple imputation. In Sub-study I, the criteria for including
variables into the imputation model were that they had to have more than 5%
points more observed values than the outcomes and a Pearson correlation
higher than 0.20 with the outcomes. In Sub-studies II-III, the register
variables with highest correlation, as well as all the variables from the
regression models were included. The exact variables used for each of the
Sub-studies I-III are reported in the Supplementary tables of the studies.
The MI was based on chained equations (ICE) (van Buuren and Groothuis-
Oudshoorn, 2011) and regression trees (Therneau et al., 2015) suitable for
imputing categorical variables, such as the needed outcome variables: MDD
and dysthymia diagnoses. In Sub-study I, both baseline and follow-up
datasets were imputed simultaneously, by groups defined by age and gender.
A total of 20 imputed data sets were constructed in Sub-study I; 24 in Sub-
study III; and 35 in Sub-study II.
In Sub-Study I, the MI was based on the total sample of Health 2000 (a total
of 8,028 people) whereas in Sub-Study II, it was based on the sample of
4,057 participants without depression at baseline, and in Sub-Study III, on
the 7,112 people who participated in at least some part of the study. It was
considered that imputing all the missing values would not have reflected true
associations in the group of nonrespondents.
4.4.2 WEIGHTS
Multiple imputation was compared to the more traditional method of

accounting for non-response, inverse probability weights, which were applied
69
in the Health 2011 Survey. Their purpose is to generalise the data to

represent the survey’s target population, and adjust for the effect of non-
participation. Inverse probability weights were created using the
participation probabilities estimated by a logistic regression model, which
were then calibrated with respect to known population distributions, such as
age and gender distributions. Sociodemographic information from Statistics
Finland and data from the baseline survey were used for this purpose,
whereas disease-specific information from the Care Register for Health Care
was not used.
In the Health 2000 Survey, poststratification weights were applied. They

were based on the design weights, which are simply the inverse of the
inclusion probability of an individual. They were then calibrated using
poststratification according to known population characteristics: health care
district, university hospital district, age, gender and mother tongue
(Heistaro, 2008).
Design weights to account for the oversampling of individuals aged 80 and

over were also used in the MI analyses (Sub-studies I and III)
4.4.3 SUB-STUDY I
In Sub-study I, age and gender adjusted prevalences were estimated using

the predictive margins method (Lee, 1981). Prevalences in 2000 and 2011
were compared by adjusting for changes in the population so that the
standard population was chosen to be the 2011 population. This allowed for
analysis of time trends taking into account the changes that had occurred in
the age and gender distribution of the Finnish population. Logistic bivariate
and multivariate regression was used to analyse associations between
independent and dependent variables.
We used two methods to adjust for non-response: inverse probability

weighting (IPW) and multiple imputation (MI).
4.4.4 SUB-STUDY II
Logistic bivariate and multivariate regression models were used to analyse

risk factors of new-onset depression (MDD or dysthymia) in the working-
aged population (30-65 years at baseline). Six regression models were built
by subsequently adding new risk factors and confounders to those of the
previous model: 1) age and sex; 2) education, income and marital status; 3)
childhood adversities and paternal mental disorder; 4) social capital; 5)
somatic diseases; 6) mental health (anxiety and alcohol use disorders,
70
baseline depressive symptoms). Separate multivariate regression models
were built to analyse risk factors for MDD only and dysthymia only, where
only models 1 and 6 were used. Multiple imputation was used to handle
missing data.
4.4.5 SUB-STUDY III
In Sub-study III, prevalences with their corresponding confidence intervals

were used to describe diagnostic status at follow-up in the baseline MDD and
dysthymia groups. Proportions and means were used to describe depressive
symptoms, psychological distress, quality of life and self-rated health at
follow-up in the two groups. These were further stratified into those who had
recovered (no depressive, anxiety or alcohol use disorder at follow-up) and
those who had not.
Logistic multinomial regression models were built to analyse associations

between the predictors and two outcomes: death and persistence of
depressive disorder. This was done to account for the impact of right
censoring, where participants with the most severe disorder possibly also
have the highest mortality risk. In the multinomial regression, the two
outcome categories were compared against the reference category “alive and
without depressive disorder.”
Linear regression was used to analyse depressive symptoms at follow-up

based on the BDI-13 score. Upon analysing the linear regression model, 12
possibly outlying individuals were found and the analyses were performed
excluding them, which did not impact the results.
Multiple imputation was used to account for missing data.
4.4.6 SUB-STUDY IV
In Sub-study IV, analyses were limited to age group 30-70 years, as

participation in the CIDI interview in the older age groups was low, which
could have distorted the results. Cox’s proportional hazards regression model
was used to examine the risk of all-cause mortality. Four regression models
were built to examine the impact of different confounders: 1) adjusting only
for psychiatric comorbidity; 2) adjusting for marital status, education and
income; 3) adjusting for health and health behaviour; 4) combining all
confounders from models 1-3. All models were stratified for age and sex.
The models were tested to verify the proportional hazards assumption. This
did not hold for the variable describing history of cancer diagnosis (p-value
71
0.044), because the mortality risk of persons with diagnosed cancer was high
at the beginning of the follow-up and then decreased to level of risk close to
the cancer-free population. Therefore, cancer diagnosis was added to the
models through stratification, not direct adjustment.
72
5 RESULTS
5.1 PREVALENCE OF DEPRESSIVE DISORDERS (SUB-

STUDY I)
In 2011, the 12-month prevalence of MDD was 7.4% (95% CI 5.7-9.0); higher
in women (10.0%, 95% CI 8.2-11.8) than men (4.4%, 95% CI 2.1-6.7); and
higher in the 30-44 year age group (9.6%, 95% CI 7.1-12.1). The prevalence of
dysthymia was 4.5% (95% CI 3.1- 5.9) in the general population, with no
significant gender or age differences (Table 9).
In adjusted logistic regression, female gender (adjusted OR 2.3, 95% CI 1.58-

3.44), younger age, and being single (adjusted OR 1.5, 95% CI 1.19-2.01) were
associated with a higher prevalence of MDD. Education was not associated
with prevalence of MDD (Table 10).
Correlates of dysthymia in the adjusted logistic regression were low

education and being single (adjusted OR 2.5, 95% CI 1.55-3.95) (Table 10,
Markkula et al., previously unpublished results).
The prevalence of MDD, adjusted for sociodemographic changes that

occurred in the population, increased in women by 2% points and in men by
0.9% points. This was significant in women (p=0.049), but not in men (p=
0.777) (Figure 3). The prevalence of dysthymia did not change significantly
(p 0.426 for total population). When combined, the prevalence of depressive
disorders increased by 2.3% points in the total population, which was
significant (p =0.014). The increase was significant in women (2.9% points,
p=0.018), but not in men (1.6%, p= 0.323).
Non-participation was selective, so that non-participants had higher

prevalence of lifetime hospitalisations for depressive disorders than CIDI
participants (3.8% vs. 1.7%, respectively). The difference was more marked in
the age groups 65-74 years (3.6% vs. 1.6%) and 75 years and older (4.6% vs.
1.4%). A total of 71% of persons with lifetime psychiatric hospitalisations and
72% with lifetime hospitalisations for depressive disorders did not
participate in the CIDI interview. Therefore, it is understandable that the
prevalence rates based on MI utilising register data were higher than
prevalence rates based on weights only. A comparison is presented in Table
9. The increase was, on average, 2% points. The increasing trend in
prevalence was similar when using weights: the increase in prevalence of
MDD from 2000 (4.6%) to 2011 (5.3%) was not statistically significant
(p=0.057), but still observed.
73
Results
Table 9. 12-month prevalence of depressive disorders in Health 2011 by gender and age
group, comparing multiple imputation (MI) and weights
Prevalence 95% CI with MI Prevalence 95% CI with

with MI with weights
weights
Depressive disorders 9.6 7.9-11.3 6.8 6.0-7.5
Women 12.2 10.1-14.3 8.4 7.3-9.6
Men 6.7 3.8- 9.6 4.9 3.9-5.9
Age 30-44 years 12.5 10.1-14.9 9.4 7.6-11.3
Age 45-54 years 12.0 9.7-14.2 8.9 7.1-10.6
Age 55-64 years 7.0 4.1- 9.8 6.5 5.1- 8.0
Age 65-74 years 5.8 3.4- 8.1 3.2 2.0- 4.5
Age 75 and over 8.0 1.1-14.9 1.9 0.6- 3.2
MDD 7.4 5.7- 9.0 5.4 4.7-6.1
Women 10.0 8.2-11.8 7.0 5.9-8.0
Men 4.4 2.1- 6.7 3.5 2.7-4.3
Age 30-44 years 9.6 7.1-12.1 7.8 6.1- 9.4
Age 45-54 years 9.2 7.2-11.2 6.9 5.3- 8.5
Age 55-64 years 5.1 3.3- 6.9 5.1 3.7- 6.4
Age 65-74 years 4.5 2.3- 6.6 2.6 1.5- 3.7
Age 75 and over 6.7 1.0-12.5 1.0 0.1- 1.9
Dysthymia 4.5 3.1- 5.9 2.0 1.6-2.4
Women 4.3 3.0- 5.7 2.2 1.6-2.9
Men 4.7 2.4- 7.0 1.7 1.1-2.4
Age 30-44 years 4.5 2.6- 6.4 2.8 1.7- 3.9
Age 45-54 years 4.3 2.7- 5.9 2.7 1.8- 3.7
Age 55-64 years 4.2 2.0- 6.5 1.8 1.0- 2.5
Age 65-74 years 4.0 1.9- 6.1 0.8 0.2- 1.5
Age 75 and over 6.3 0.5-12.0 0.9 0.0- 1.9
Table 10. Correlates of MDD and dysthymia in the Health 2011 Study
MDD, adjusted OR Dysthymia, adjusted

(95% CI)* OR (95% CI)*
Men 1 1
Women 2.33 (1.58-3.44) 1.02 (0.62-1.70)
Age 30-44 years 1 1
Age 45-54 years 0.84 (0.63-1.13) 0.92 (0.55-1.54)
Age 55-64 years 0.48 (0.34-0.70) 0.65 (0.33-1.31)
Age 65-74 years 0.29 (0.17-0.51) 0.39 (1.14-1.02)
Age 75 years and over 0.26 (0.07-0.91) 0.49 (0.13-1.83)
Basic education 1 1
Intermediate education 0.78 (0.55-1.11) 0.58 (0.34-0.99)
High education 0.81 (0.57-1.17) 0.48 (0.28-0.83)
Married or cohabiting 1 1
Separated, widowed or unmarried 1.54 (1.19-2.01) 2.47 (1.55-3.95)
*Adjusted for sex, age, educational level, marital status and region
74

Figure 3 12-month prevalence of depressive disorders in the Finnish population in 2000 and
2011
5.2 RISK FACTORS OF DEPRESSIVE DISORDERS

(SUB-STUDY II)
At the follow-up, 104 new cases of MDD and 31 of dysthymia were diagnosed,
9 of who had both diagnoses. Incidence of depressive disorders differed
between subgroups, being highest among persons with anxiety disorders
(13.7%) and those with high BDI score (11.5%). In the complete sample,
incidence of new-onset depression was 4.4% (95% CI 3.6-5.2%).
In the final adjusted logistic regression models, significant risk factors of

new-onset depressive disorders were younger age, female gender, having
three or more childhood adversities, low trust axis of social capital, anxiety
disorder and depressive symptoms (Table 11). The predictors were different
for the two depressive disorders. Risk factors for MDD were younger age (OR
0.97, 95% CI 0.95-0.99 per year), female gender (OR 1.68, 95% CI 1.11-2.55),
anxiety disorder at baseline (OR 2.53, 95% CI 1.17-5.46) and subclinical
depressive symptoms (OR 1.64, 95% CI 1.00-2.68 for moderate and OR 3.03,
95% CI 1.48-6.19 for severe depressive symptoms). Risk factors for
dysthymia were younger age (OR 0.93, 95% CI 0.89-0.98 per year), three or
more childhood adversities (OR 3.36, 95% CI 1.22-9.27), low trust axis of
social capital (OR 0.30, 95% CI 0.09-0.97 for high trust) and chronic somatic
diseases (OR 3.37 for 1-2 diseases, 95% CI 1.19-9.52).
75
Results
Table 11. Incidence of depressive disorders by subgroup and adjusted odds

ratio for new-onset depressive disorders in the Health 2011 Survey
Incidence of P for Adjusted OR for

depressive difference depressive disorders
disorders (%) (95% CI)
Sociodemographic characteristics
Age group: 30-44 5.4 0.033 0.97 (0.95-0.99)
Age group: 45-54 4.2 (per year)
Age group: 55-64 2.9
Sex: male 3.5 0.018 1

Sex: female 5.3 1.46 (1.01-2.12)
Education: primary 3.6 0.420 1

Education: secondary 4.7 1.27 (0.75-2.15)
Education: tertiary 4.7 1.23 (0.73-2.09)
Income: 1st quintile 4.3 0.974 1

Income: 2nd quintile 4.8 1.23 (0.54-2.80)
Income: 3rd quintile 4.0 0.99 (0.42-2.34)
Income: 4th quintile 4.4 1.06 (0.46-2.47)
Income: 5th quintile 4.5 1.17 (0.51-2.67)
Marital status: married 4.4 0.838 1

Marital status: unmarried 4.5 0.94 (0.56-1.61)
Childhood adversities
No adversities 3.5 0.001 1
1-2 4.3 1.15 (0.78-1.70)
3 or more 7.6 1.76 (1.10-2.83)
Parental mental disorder 7.3 0.068 1.29 (0.66-2.53)

Social capital
Axis 1: support low 4.1 0.825 1
Axis 1: support medium 4.4 1.06 (0.66-1.71)
Axis 1: support high 4.6 1.07 (0.66-1.73)
Axis 2: participation low 5.2 0.329 1

Axis 2: participation medium 3.7 0.66 (0.41-1.07)
Axis 2: participation high 4.5 0.74 (0.45-1.19)
Axis 3: trust low 6.0 0.006 1

Axis 3: trust medium 4.6 0.85 (0.55-1.30)
Axis 3: trust high 2.9 0.58 (0.36-0.96)
Somatic morbidity
No somatic diseases 3.8 0.369 1
1-2 somatic diseases 4.8 1.32 (0.88-1.97)
3 or more somatic diseases 4.4 1.26 (0.60-2.90)
Psychiatric morbidity
BDI 0-9 3.8 <0.001 1
BDI 10-18 6.7 1.65 (1.04-2.61)
BDI 19 or more 11.5 2.49 (1.20-5.17)
Anxiety disorder 13.7 <0.001 2.75 (1.36-5.56)
Alcohol use disorder 6.0 0.375 1.32 (0.60-2.90)
76
5.3 PROGNOSIS OF DEPRESSIVE DISORDERS (SUB-
STUDY III)
A total of 392 people had a depressive disorder at baseline, out of whom 245
had MDD only, 94 dysthymia only and 53 both MDD and dysthymia. Of
them, 40.2% had current severe depressive symptoms, measured with the
BDI.
At the follow-up, 20.9% of people with baseline MDD and 27.0% of people
with baseline dysthymia still had a depressive disorder, and 33.8% and
42.6% some depressive, anxiety or alcohol use disorder, respectively.
Depressive symptoms were more common than diagnosis of depressive
disorder: 47.6% of those with baseline MDD and 61.2% of those with baseline
dysthymia still had significant depressive symptoms after eleven years (BDI-
13 5 points or higher) (Table 12).
Table 12. Diagnostic status and depressive disorders after eleven years follow-up by
baseline diagnosis of depressive disorder (MDD or dysthymia)
Diagnosis at No Any MDD in Dysthymia Any BDI-13

baseline diagnosis depressive, 2011 in 2011 depressive score 5
in 2011 anxiety or % % disorder or more
% alcohol use % %
disorder in
2011
%
MDD in 2000 66.2 33.8 16.0 8.1 20.9 46.5
(n=298)
Dysthymia in 57.4 42.6 16.4 13.4 27.0 61.1

2000 (n=147)
Also health-related quality of life, measured as the EQ-5D score, was lower
than in the general population (Table 13). However, when separated by
diagnostic status at follow-up, those who had recovered from MDD had
values close to the general population, whereas those recovered from
dysthymia still had lower quality of life and more depressive symptoms.
Also poor or rather poor self-rated health was more common than in the
general population: 16.7% of persons with baseline MDD and 28.6% of
persons with baseline dysthymia rated their health as poor or rather poor,
compared with 10.8% of the total sample.
77
Results
Table 13. Depressive symptoms, quality of life and self-rated health of persons with MDD
and dysthymia after eleven years follow-up
Baseline MDD Baseline dysthymia

All
Recovered Any CIDI Recovered Any CIDI
(no CIDI diagnosis (no CIDI diagnosis
diagnosis in 2011 diagnosis in 2011
in 2011) in 2011)
No significant depressive 75.4 65.4 27.3 53.3 18.0
symptoms (BDI-13 score
0-4) %
Moderate depressive 12.0 16.2 15.7 24.7 12.4
5-8) %
Severe depressive 12.5 18.4 57.1 22.0 69.6
9-39) %
Quality of life, EQ-5D (UK 0.81 0.80 0.72 0.73 0.65

index, mean)
Self-rated health rather 10.8 11.6 26.6 22.5 37.3
poor or poor %
Single marital status (adjusted OR 1.91, 95% CI 1.05-3.56) and symptom

severity at baseline (adjusted OR 1.05, 95% CI 1.01-1.09 per one point
increase in BDI) were associated with persistence of disorder. Symptom
severity at baseline also predicted more depressive symptoms at follow-up.
No other studied factors, such as age, gender, education, history of childhood
adversity, social capital, somatic or psychiatric comorbidity, were statistically
significant predictors of outcomes of depression.
A total of 29 individuals with baseline MDD and 25 individuals with baseline

dysthymia died during the follow-up. Older age, male sex and symptom
severity as measured with the BDI were associated with higher mortality risk.
The analyses were also carried out using weights to account for non-
participation, instead of MI. There were no important differences, but in the
weighted models, also childhood adversity and low social capital were
statistically significant predictors of poor outcomes of depressive disorders.
All in all, the estimates were similar in direction and magnitude.
78
5.4 EXCESS MORTALITY IN DEPRESSIVE, ANXIETY
AND ALCOHOL USE DISORDERS (SUB-STUDY IV)
Altogether 323 deaths were observed over the eight-year follow-up period in
the studied age group of 30-70-year-olds. Majority of deaths (82.9%) were
natural, 11.7% were accidents, 5.1% (n=16) suicides, one homicide and one
unknown. Of the 16 persons who committed suicide, five (38.5%) did not
have a known psychiatric diagnosis, but many of them had more than one,
the most common being MDD (5 persons), social phobia (5) and alcohol
dependence (3).
Depressive disorders were associated with an increased mortality risk (fully

adjusted HR 1.97, 95% CI 1.15-3.39) (Table 14). The unadjusted HR was 2.4,
and was more affected by adjustment to sociodemographic factors (HR 1.7)
than for health status and smoking (HR 2.2).
Among persons with anxiety disorders, a large share (50%) of deaths were
unnatural, and in some disorders this figure was 80-100%. The mortality risk
was increased (HR 2.32, 95% CI 1.35-3.96), but became statistically non-
significant when adjusted for psychiatric comorbidity, and remained non-
significant in further adjustments.
Alcohol use disorders also had a high rate of unnatural deaths (47.6%), and
were associated with significantly increased mortality risk (adjusted HR 1.72,
95% CI 1.10-2.71).
Also depressive symptoms, measured with the BDI, were associated with
increased mortality, without controlling for psychiatric disorders and
controlling for other confounders (HR 1.53, 95% CI 1.11–2.11 for BDI score
10–18 and HR 1.77, 95%CI 1.09–2.86 for BDI score 19 or more).
Figures 4-5 present survival curves for men and women with depressive,
anxiety and alcohol use disorders, and without these disorders, adjusted for
other risk factors (age, education, income, marital status, smoking status,
BMI and somatic diseases (Markkula et al., previously unpublished results).
79
Results
Table 14. Mortality risk by psychiatric disorder and gender in unadjusted and adjusted
models after 8-year follow-up (results by gender previously unpublished)
Adjusted for socioeconomic

Unadjusteda factors, health status and
smokinga
Hazard ratio p-value Hazard ratio (95% p-value
(95% CI) CI)
Any depressive disorder 2.40 (1.52-3.78) <0.001 1.97 (1.15-3.39) 0.01

Men 2.78 (1.56-4.98) 0.001 1.84 (0.89-3.81) 0.099
Women 1.95 (0.97-3.94) 0.062 2.13 (0.92-4.92) 0.078
Any anxiety disorder 2.32 (1.35-3.96) 0.002 1.20 (0.58-2.45) 0.62
Men 2.82 (1.50-5.30) 0.001 1.39 (0.60-3.18) 0.441
Women 1.46 (0.50-4.26) 0.490 0.83 (0.21-3.29) 0.791
Any alcohol use disorder 2.55 (1.65-3.93) <0.001 1.72 (1.10-2.71) 0.02
Men 2.40 (1.52-3.76) 0.000 1.52 (0.94-2.46) 0.088
Women 5.43 (1.42-20.6) 0.013 3.48 (0.68-17.7) 0.133
a. All models are stratified for age and sex. The adjusted model is also stratified for lifetime
psychosis and history of cancer, and adjusted for education (low/intermediate/high), income
quintile, marital status (married or cohabiting/other), smoking status (smoker/ex-
smoker/never-smoker), BMI (<25/25-30/ >30), and somatic diseases (pulmonary,
cardiovascular, musculoskeletal, vision and hearing, neurological, diabetes and other
(psoriasis, inflammatory bowel disease, urinary incontinence, disturbing allergy).
80
Figure 4 Survival curves of men with depressive, anxiety and alcohol use disorders
Figure 5 Survival curves of women with depressive, anxiety and alcohol use disorders
81
Discussion
6 DISCUSSION
6.1 SUMMARY OF THE MAIN RESULTS

This longitudinal population study of depressive disorders examined
prevalence, predictors and different adverse outcomes of depressive
disorders in a general population setting. Specifically, the aims were to
establish the prevalence of depressive disorders (major depressive disorder
and dysthymia) in the Finnish population in 2011, and to assess possible
changes in prevalence since 2000; to examine risk factors for new-onset
depressive disorders; to investigate the long-term prognosis of depressive
disorders in the population and different predictors of adverse outcomes;
and to assess excess mortality in depressive, anxiety and alcohol use
disorders.
The 12-month prevalence of depressive disorders in the Finnish general

population aged 30 years and over was 9.6% (95% CI 7.9-11.3) in 2011. There
was a significant increase in the prevalence from 2000 to 2011. Accounting
for non-participation by multiple imputation increased the prevalence
estimate by approximately two percentage points.
Risk factors for new-onset cases of depressive disorders were younger age,
female gender, childhood adversities, lower trust axis of social capital, and
having an anxiety disorder or subclinical depressive symptoms at baseline. In
addition, having somatic diseases was a risk factor for dysthymia.
Of the persons with MDD at baseline, 34% still had some depressive, anxiety
or alcohol use disorder after eleven years, and 48% had clinically significant
depressive symptoms. Among those with dysthymia at baseline, the
prognosis was worse: 43% had some depressive, anxiety or alcohol use
disorder, and 61% had depressive symptoms at follow-up. Both groups also
had worse health-related quality of life and self-rated health, indicating
greater health needs. The difference to the general population was
particularly noticeable among persons with baseline dysthymia. Unmarried
persons and those with a more severe initial disorder had a higher risk of
persistent course.
Persons with depressive disorders had a twofold mortality risk, whereas the
risk was 1.7-fold in alcohol use disorders and not increased in anxiety
disorders, when adjusted for age, sex, socioeconomic status, physical health
and health behaviours. The mortality risk was higher among depressed men
and in those with a more severe disorder.
82
6.2 POTENTIAL RISK FACTORS AND PREDICTORS OF
NEGATIVE OUTCOMES OF DEPRESSION
6.2.1 GENDER AND AGE
In this study, women had a higher prevalence of MDD but not dysthymia,
and their adjusted odds for current (prevalent) MDD, but not dysthymia,
were increased. The risk for new-onset MDD was 1.5-fold in women.
Prognosis of depressive disorders did not differ by gender except for the fact
that depressed men had a higher mortality risk than women. In the separate
mortality analyses, however, the hazard ratios were of similar magnitude in
both genders, suggesting that the excess mortality in depressed men reflected
the gender gap in life expectancy in the general population.
These findings are largely consistent with previous literature. Most studies
have found women to have higher incidence of depression, approximately 1.5
to 2-fold (De Graaf et al., 2002, Eaton et al., 2008c, Klein et al., 2013,
Stegenga et al., 2013, Wang et al., 2010a). However, incidence of depression
decreases in both genders after 30 years of age, and in the studied age group
of 41-77 years, the gender difference is smaller than in young adulthood
(Patten et al., 2016, Pedersen et al., 2014). Therefore, it is understandable
that the risk difference was slightly smaller than in many studies examining
younger populations.
Both the World Mental Health Surveys and a US national population survey
found a higher prevalence of dysthymia in women (Blanco et al., 2010,
Seedat et al., 2009). This differs from our finding, but there are few other
studies to compare with. According to a meta-analysis, the mortality risk in
depression is twofold in depressed men compared with depressed women.
Also, as compared with the general population, depressed men have a higher
mortality risk (RR 2.0) than women (RR 1.6) (Cuijpers et al., 2014b). The
higher mortality risk of men was observed in sub-study III (depressed men
vs. depressed women) but not sub-study IV (depressed men and women vs.
general population).
Why is the prevalence of MDD higher in women than in men? The question
can be approached by examining which factors influence the magnitude of
the risk difference. The risk difference between men and women peaks at 20
years; after 40 years, it is very small (Pedersen et al., 2014). Women’s higher
risk thus appears to be related to different life stressors in adolescence and
early adulthood. It has been suggested that the difference in early
adolescence is related to: the differing ages at which boys and girls develop
emotional maturity; the differences in the way boys and girls are brought up;
and differing reactions to maternal depression and other negative life events
83
Discussion
(Jenkins and Curwen, 2008, Zahn-Waxler et al., 2000). Biological

differences also influence the risk of depression and other mental disorders
in later life through epigenetic mechanisms (Kigar and Auger, 2013).
Hormonal changes in a woman’s life, such as entering puberty, pre-
menstrual symptoms, perinatal depression and menopause, might also
explain part of the higher prevalence (Buttner et al., 2013, Schiller et al.,
2015, Toffol et al., 2015). Estrogens have been suggested to increase the risk
of depression via two pathways – interactions with neurotrophic factors,
such as the brain-derived neurotrophic factor (BDNF), and their influence on
the serotonergic system (Borrow and Cameron, 2014). Early menarche, for
example, has been associated with increased risk of depressive symptoms in
early adolescence (Joinson et al., 2013). On the other hand, longer duration
of the reproductive period may protect against postmenopausal depression
(Georgakis et al., 2016).
As previously noted, risk factors for depression differ between genders, with
personality and interpersonal factors being more important among women,
and externalising psychopathology, history of depression and stressful events
related to financial and occupational problems more important among men
(Kendler and Gardner, 2014). In early adulthood, women report more life
events prior to the onset of depression, but this difference disappears later in
life (Harkness et al., 2010). It has been suggested that acute stressors
increase the risk of MDD in women and substance use disorders in men, but
the evidence to support this is mixed, and most likely the pathways are more
complex than previously thought (Slopen et al., 2011).
Between countries, the risk difference is higher in high-income countries (Rai

et al., 2013). On the other hand, Seedat et al. (Seedat et al., 2009) found that
the male-female difference decreased over time and was smaller where
gender roles were more equal. They hypothesize that the increase in female
employment opportunities and birth control has promoted women’s mental
health and reduced exposure to stressors.
Some of the higher prevalence of depression among women may be due to

ways that depressive symptoms are conceptualised. For example, crying and
loss of sexual desire are more frequent among women (Salokangas et al.,
2002). Consistent with this idea of different diagnostic thresholds, men who
receive a diagnosis of a depressive disorder report more disability (Scott and
Collings, 2010). The diagnostic criteria of depressive disorders, however, do
not include these contentious symptoms, and it is unlikely that measurement
bias would explain all of the gender difference in depression.
The prevalence and incidence of depressive disorders decreased with age,

with the exception of prevalence of dysthymia, which remained stable. This is
understandable, given the long-term course of the disorder, and in line with
84
earlier literature (Eaton et al., 2008c, Kessler and Bromet, 2013, Stegenga et
al., 2013, Wang et al., 2010a). Some of the decrease may be due to
methodological issues: O’Connor and colleagues have shown that the rate of
positive responses to the rather complicated CIDI screening questions shows
a much steeper decline with age than responses to the simpler symptom-
related items of the interview or symptom scales (O'Connor and Parslow,
2009, O'Connor and Parslow, 2010). Another consideration is the lower
participation rate in the mental health interview by older individuals.
Consistent with this, the prevalence rates in older age groups were influenced
by multiple imputation more than the younger ones, and, at the same time,
point estimates were less precise.
Nonwithstanding the limitations of psychometric instruments, it is possible

that the impact of stressors in different stages in life influences the incidence
of depressive disorders. Also, some personality features that predispose to
depression may become subtler with increasing age, such as issues with
interpersonal relationships and self-image. When comparing variation of
depressive symptoms over the life course, both mild and severe depressive
symptoms are less frequent in older age groups relative to younger ones,
whereas symptoms classified as “despondent” (or deeply hopeless) did not
decrease with age (Mezuk and Kendler, 2012).
6.2.2 SOCIOECONOMIC POSITION
In this study, neither income nor education as measures of socioeconomic

position (SEP) was associated with prevalence, incidence or outcomes of
depressive disorders. The only exception was dysthymia, which was more
prevalent among people with low education. A meta-analysis found a 1.8-fold
prevalence and 1.2-fold incidence of depression among those with lowest SEP
(Lorant et al., 2003), but many individual studies from high-income
countries have found no association. It is possible that the lack of association
in incidence in our study is explained by thorough controlling for
confounders, such as physical diseases and childhood adversities. In our
study population, educational attainment was strongly correlated with age, as
the educational level of the Finnish population has risen significantly over
the past decades (Statistics Finland, 2015). This may also partly explain our
lack of significant associations with education, since depressive disorders
were more common among the younger age groups.
These results do not support the theory of direct causation from adult SEP to
depression. On the other hand, childhood adversities, some directly and
some indirectly linked to socioeconomic position, were a strong risk factor
for depression. Similarly, in a study of Finnish adolescents, parental
85
Discussion
socioeconomic position was strongly correlated with current depressive

symptoms (Torikka et al., 2014).
In another analysis of the working-age participants of the Health 2000

Survey, low income, but not education or manual occupation, was associated
with a 1.7-fold odds of current depressive disorders, when adjusted for age,
gender and marital status (Pulkki-Råbäck et al., 2012). This finding, together
with the fact that dysthymia was more prevalent among people with less
education, could be interpreted to support the theory of selection: that
depressive disorders cause downward movement on the socioeconomic scale,
or prevent moving upwards. On the other hand, there could be common
factors leading to both low SEP and depressive disorders. Finally, we did not
specifically assess the impact of financial hardship, which has been a stronger
predictor than objective measures of SEP (Skapinakis et al., 2006, Wang et
al., 2010b, Weich and Lewis, 1998).
6.2.3 CHILDHOOD ADVERSITIES AND FAMILY HISTORY
In this study, an accumulation of three or more childhood adversities

predicted onset of depressive disorders, particularly dysthymia. People with
three or more childhood adversities also had a 1.9-fold risk for persistence of
depressive disorder, but this association reduced and became non-significant
after controlling for other confounding factors. Parental mental disorder was
not associated with onset of depressive symptoms, but we were not able to
examine the impact of parental depressive disorder specifically. Also, the
information was based on self-report and only two questions, one on each
parent’s mental health, and therefore is not as reliable as prospective studies
with standardised psychiatric evaluations of the parents. In earlier studies,
parental mental disorder has been a risk factor for depression (Eaton et al.,
2001, Klein et al., 2013, Stegenga et al., 2013). However, in one study, the
impact was mediated through childhood maltreatment (Plant et al., 2015).
The impact of childhood adversity on the risk of depression is well-known in

the literature (Bowes et al., 2015, Park et al., 2013, Pirkola et al., 2005a,
Ritsher et al., 2001, Sourander et al., 2015, Stegenga et al., 2013) (Elovainio
et al., 2015), even though there is controversy over the reporting bias
introduced by current depressive symptoms (Colman et al., 2015). In this
study, this was not a problem, as in Sub-study II, the individuals did not have
depression at baseline, when childhood adversities were inquired. However,
there may have been memory bias, as the study participants were middle-
aged or older at the time of the interview.
In two studies examining the link between childhood adversities and chronic
depressive disorder or psychological distress, the association was mediated
86
through personality pathology, such as avoidant or generally maladaptive
personality (Klein et al., 2015, Spinhoven et al.). In another study, the
negative impact of childhood maltreatment on the course of depression was
mediated by personality characteristics (Hovens et al., 2016). This could
explain how the risk remains increased in midlife and even later.
Furthermore, according to the interpersonal theory of depression, negative
interpersonal events in childhood, such as exposure to bullying, represent a
type of trauma which may lead to development of depression similarly to
other traumatic events in life (Sourander et al., 2015). The cognitive theory
suggests that peer victimisation, particularly verbal or relational, contributes
to the development of negative self-cognitions, which predisposes a person to
depression (Cole et al., 2014).
It is worthwhile to note that we did not have information on childhood abuse,

which has been found to be a particularly strong risk factor for depression. A
meta-analysis of consequences of different types of abuse found the strongest
influence on risk of depression in emotional abuse, followed by neglect and
physical abuse (Norman et al., 2012). Our results show that an accumulation
of other forms of adversity, perhaps considered more benign in nature, does
considerably increase the risk of depression. It may be encouraging news to
families who face adversities that according to our study, it was the
accumulation of three or more adversities that significantly increased the risk
of depressive disorders.
6.2.4 MARITAL STATUS AND SOCIAL CAPITAL
In this study, single marital status (i.e. being never married, divorced,
separated or widowed) was associated with a higher prevalence, but not a
higher incidence, of depression, in addition to an increased risk of
persistence of the disorder. It seems that single people do not have a higher
risk of developing depression, but that the higher prevalence is instead due to
reduced chance of recovery. Depressed individuals may also have a higher
likelihood of ending a relationship or not starting one, but this was not
examined in this study. Other studies have shown that mental disorders
reduce the chances of marrying and increase the risk of divorce (Breslau et
al., 2011).
The higher prevalence of depression among single people is in line with

earlier literature (Kessler and Bromet, 2013), and some studies have also
found higher incidence (Anthony and Petronis, 1991, Scott et al., 2010).
Similar to our findings, other studies have found longer episode duration and
a higher risk of recurrence among the unmarried (Eaton et al., 2008c, van
Loo et al., 2015a). The association between marriage and depression also
depends on customs and culture related to marriage: a study of married
87
Discussion
Pakistani women showed that depression was associated with teenage and
arranged marriage, and to different forms of abuse by the husband and his
family (Ali et al., 2009).
The increased risk of persistence of depression among the unmarried could

be related to the detrimental effect of loneliness. Social isolation or loneliness
is associated with depressive symptoms and to other negative health
outcomes (Hawkley and Cacioppo, 2010). Also, social support is reduced
among those who are not in a relationship (Joutsenniemi et al., 2006). It
seems that having a partner is of particular importance to mental health,
beyond the role of increased social capital that a relationship entails.
We found that social capital, and specifically its trust axis, was protective
against new-onset depressive disorders. Among people with baseline
depression, low participation and low trust were associated with having more
depressive symptoms at follow-up in the weighted-only analyses.
The questions measuring trust dealt with feeling safe in the neighbourhood;
cynical mistrust; and feeling disappointed or surprised by the behaviour of
people close to you. The construct of “trust” as a component of social capital
is closely related to personality features and the way an individual interprets
the neighbourhood and the people surrounding them. The features that allow
an individual to express trust in relation to their surroundings could also
make them more resilient to the life stressors that increase the risk of
depression. Interestingly, one study found no association between social
capital and incidence of depression, when controlling for personality features
(Noteboom et al., 2015). In addition to personal characteristics, however, the
concept of cynical mistrust also reflects the surrounding social environment
(Nieminen et al., 2008), and an environment of little trust between people
could thus predispose to depression. Another longitudinal study has also
shown that both cynical hostility and cynical mistrust were strong risk factors
for later depressive symptoms (Nabi et al., 2010b).
Low participation showed an association with new-onset depression that

disappeared when confounders were controlled for. Of note is the lack of
association between support and both incidence and persistence of
depression. This highlights the importance of individual characteristics and
actions over passively received support.
6.2.5 CHRONIC SOMATIC CONDITIONS
We found a higher incidence of dysthymia among persons with chronic

somatic conditions at baseline. Somatic comorbidity was not associated with
persistence with depressive disorders.
88
Having one or two chronic diseases increased the risk of new-onset
dysthymia, but not MDD. In other studies, chronic physical diseases have
been associated with incidence of depression (Kaplan et al., 1987, Patten,
2001, Stegenga et al., 2013, Wang et al., 2010a) and longer course of illness
(Patten, 2005, Patten et al., 2010). The Canadian NPHS, which is the largest
study to document association between chronic physical diseases and
depression, used an instrument that does not differentiate between types of
depression, and therefore it is not possible to assess whether it was the risk of
MDD or dysthymia that was increased.
The mechanism by which chronic physical diseases increase the risk of

dysthymia could be a common underlying factor or causal pathway such as
inflammation (Kiecolt-Glaser et al., 2015). Or it could be related to the
experience of living with a chronic illness and its consequences, such as loss
of the ability to work or the loss of functional capacity. Finally, detection of
depression among people with chronic somatic conditions may be
particularly challenging (Menear et al., 2015), and it is possible that
undetected depressive symptoms become chronic in this population. Most
likely, the higher risk of depression in this population is a combination of
these and other factors.
6.2.6 OTHER PSYCHIATRIC DISORDERS AND PSYCHOLOGICAL

SYMPTOMS
We found significantly higher incidence of MDD among people with anxiety

disorders. Psychiatric comorbidity was not associated with persistence of
depressive disorders, although this has been commonly found in other
studies (Kessler et al., 2008, Rhebergen et al., 2011). Baseline depressive
symptoms were predictive of later onset of depressive disorders among
people with no depression at baseline. Among people with baseline
depression, severity of depressive symptoms was a strong predictor of all
negative outcomes: persistence of disorder, level of depressive symptoms at
follow-up and mortality.
Similar to our findings, previous studies have found higher incidence of

MDD among people with anxiety disorders (Eaton et al., 2008c, Kessler et
al., 2008, Klein et al., 2013, Stegenga et al., 2013). Developmentally, the
genetic etiology in anxiety and depressive symptoms seems to be
independent in childhood, but shared from adolescence onwards, with an
overarching internalising genetic factor (Waszczuk et al., 2014). In addition
to common genetic risk factors, there can be other common underlying risk
factors, such as childhood adversity, personality features, negative life
experiences, or the direct impact of the anxiety disorder as a stressor, as well
89
Discussion
as its consequences, such as social isolation (Kessler et al., 2008, Moreno-

Peral et al., 2014, Moscati et al., 2015). Remarkably, alcohol use disorders
did not increase the risk of depressive disorders in this study. A meta-
analysis found a twofold risk of MDD among people with alcohol use
disorder (Boden and Fergusson, 2011).
The fact that subclinical depressive symptoms predicted later onset of

depression is to be expected, and is frequently found in other studies (Ernst
et al., 1992, Horwath et al., 1992, Klein et al., 2013, Skapinakis et al., 2006).
These could have been an indicator of a depressive episode about to begin
already close to baseline, or they may reflect a way of expressing emotions, or
a personality trait.
It is intuitive that the severity of the disorder also predicted persistence, and
this is line with previous findings (Lamers et al., 2016, Penninx et al., 2011).
However, it is notable that depressed people with more depressive symptoms
at baseline also had higher mortality. This may reflect a particular subtype of
depression with higher mortality risk (Ziegelstein, 2015), or simply that the
mechanisms leading to excess mortality in depression are more pronounced
in more severe disorders. A previous study found increased mortality risk
already present in persons with mild depressive symptoms, and an increase
in mortality risk with more severe symptoms (White et al., 2015).
6.3 INCREASING PREVALENCE OF DEPRESSIVE

DISORDERS AND IMPACT OF NON-PARTICIPATION
The annual prevalence of depressive disorders in the Finnish population
increased from 7.3% in 2000 to 9.6% in 2011. The increase was more notable
in women than men, and in MDD than in dysthymia. Unfortunately, no
direct explanation for the increase can be derived from this study. However,
based on the established risk factors for depression, some speculations can
be made. In 2000, Finland was experiencing steady economic growth,
whereas in 2011, there was an economic recession. Economic crisis is
associated with negative mental health impacts such as depression and
increase in suicide rates (Stuckler et al., 2009, Wahlbeck and McDaid, 2012).
The impact is mediated via increased unemployment and poverty, individual
financial difficulties, alcohol-related harm, social exclusion and family strain
(Wahlbeck and McDaid, 2012). It is possible that the economic downturn
contributed to the increase in prevalence of depressive disorders in Finland.
Considering that childhood adversities predict onset of depressive disorders,

some explanation to increasing rates of depression can also be found in the
previous recession of the early 1990s, when many families experienced
90
unemployment and financial difficulties as well as other adverse
consequences, such as parental alcohol use. In 2011, people who lived their
childhood during the recession of the early 1990s were in their early 30s, and
thus part of the Health 2011 sample. There is abundant evidence of the
psychological malaise of the generation born in 1987 (Paananen et al., 2013a,
Paananen et al., 2013b), and it is likely that the negative impact of the 1990s
recession extends beyond this cohort.
In keeping with our finding of increasing prevalence, during the same

observation period, 2001-2011, the prevalence of self-reported depression
increased among all adolescents, and doubled among adolescents whose
parents were unemployed (Torikka et al., 2014). This further highlights the
role of family social adversity as a contributor to the increasing prevalence of
depression.
We found significantly higher prevalence rates of depressive disorders when

accounting for non-participation by multiple imputation than by inverse
probability weights only. Based on this, it is evident that non-participation of
people with mental disorders in health surveys introduces a downward bias
in all population survey based prevalence estimates that do not correct for
non-participation (Haapea et al., 2008). This should be taken into account
when analysing and comparing studies on prevalence of mental disorders.
Based on our results, multiple imputation seems to be a more effective way of
correcting for non-participation than the more traditional weighting.
6.4 MORTALITY IN DEPRESSIVE, ANXIETY AND

ALCOHOL USE DISORDERS
We found excess mortality in all studied categories of mental disorders:
unadjusted, the hazard ratio was 2.4 for depressive; 2.3 for anxiety; and 2.6
for alcohol use disorders. When adjusted, however, the risk was twofold in
depressive disorders, and 1.7-fold in alcohol use disorders, but no longer
significant (HR 1.2) in anxiety disorders. Both anxiety and alcohol use
disorders had a high rate of unnatural deaths. Among depressed people,
older individuals, men, and people with more depressive symptoms had
higher mortality.
Our results regarding excess mortality in depressive disorders are in line with
recent reviews and meta-analyses that found 1.5 to 1.9-fold mortality risks
(Baxter et al., 2011, Cuijpers et al., 2014a, Ferrari et al., 2013a, Walker et al.,
2015). The mortality risk was increased despite extensive controlling for
confounders and long follow-up, factors that had been associated with
smaller risk in previous studies, and in a general population sample as
opposed to a clinical one, which also have higher mortality rates (Cuijpers et
91
Discussion
al., 2014a). Of note, our sample included only people with unipolar
depression. Most other studies and none of the recent four reviews mention
explicitly whether only unipolar depression was considered. It seems,
however, that this would not significantly impact the results, as mortality in
bipolar disorder is similar to depressive disorders (Walker et al., 2015; Hayes
et al., 2015).
Based on our results, the approximately twofold risk reported by the meta-
analyses does not appear to be exaggerated or inflated by low-quality studies.
This supports the saying that there truly is no health without mental health
(Prince et al., 2007).
A recent meta-analysis found a higher mortality risk among depressed men

(HR 2.0) than among women (HR 1.6) (Cuijpers et al., 2014b). We had
similar results in Sub-study III, where depressed men had a higher mortality
risk than did women. However, in Sub-study IV, there was little difference in
mortality risk between men and women in the magnitude of the risk. The
only exception was alcohol use disorders, where the risk was 1.5-fold for men
and 3.5-fold for women, although due to the small number of observations,
neither was statistically significant.
Unfortunately, there was not enough data to draw conclusions on the excess
mortality in specific disorders, and thus the mortality risk in dysthymia
remains unknown.
Mortality risk in anxiety disorders (HR 1.2) was smaller in magnitude than
reported in a recent meta-analysis (HR 1.4) (Walker et al., 2015) and not
significant. Separated by gender, the risk was larger in men than in women,
as in an earlier study in the Netherlands (van Hout et al., 2004), but not
significant in either sex. The mortality risk became non-significant when
controlled for comorbid depressive and alcohol use disorders. It is possible
that not all studies examining mortality in anxiety disorders have been able
to control for comorbid depressive disorders, which may have biased the
results.
Excess mortality in alcohol use disorders was similar to that found in earlier
studies (Eaton et al., 2008a, Harris and Barraclough, 1998), and similar to
the twofold risk found in a meta-analysis of general population studies
(Roerecke and Rehm, 2013), but lower than the 3.5-fold risk in the most
recent meta-analysis (Laramée et al., 2015). However, most of the studies in
the meta-analysis simply compared age and gender-controlled mortality
rates without further adjustments. The fact that we controlled carefully for
many physical, mental and behavioural risk factors might explain our lower
risk estimate. Also, mortality was lower in studies where cases were selected
from the general population and not from treatment centres, and those with
92
longer follow-up, both findings in line with our lower risk estimate. The
mortality risk was higher in women than in men, which is also consistent
with earlier studies (John et al., 2013, Roerecke and Rehm, 2013).
6.5 METHODOLOGICAL CONSIDERATIONS –

STRENGTHS AND LIMITATIONS
6.5.1 POPULATION SURVEY DATA
This was a longitudinal population-based study, which has several benefits

over other study settings. First, as opposed to studies based on clinical
samples or treatment-based registers, a population-based study includes
both subjects who have sought treatment and those who have not. Therefore,
conclusions regarding depressive disorders can be made without the bias
related to differing help-seeking behaviours (Susukida et al., 2015). Second,
unlike cross-sectional studies, we were able to draw conclusions on temporal
associations between potential risk factors and depressive disorders, and also
describe outcomes of the disorders. Third, as this was a general health
survey, we had extensive information on the subjects’ physical health status.
However, the longitudinal population setting also posed some limitations.

Even though the study sample was large, the number of people diagnosed
with depression at baseline or follow-up was not sufficient to carry out many
analyses by subgroups, such as gender, age group or type of depression.
Likewise, when analysing many potential risk factors simultaneously, the
statistical power was reduced, resulting in widening confidence intervals.
The fact that the survey covered a broad range of health indicators limited
the time available for questions regarding mental health. As a result, we did
not have information on, for example, lifetime disorders, and therefore could
not exclude certain conditions, such as bipolar disorder at follow-up. Also, we
did not have information on the participants’ mental health between the
baseline and follow-up interview, which could have been acquired using a
more detailed mental health interview or more detailed register data. The
question regarding childhood adversities could have been influenced by
recall bias. Due to the limited time available for both the mental health
interview and mental health instruments, we did not include instruments to
measure personality features, which are important risk factors of depression.
Finally, we were not able to analyse specific symptoms or clusters of
symptoms, and whether they had specific risk factors, which has been found
in other studies (Fried et al., 2014).
93
Discussion
6.5.2 THE CIDI INTERVIEW
The Composite International Diagnostic Interview, Munich version (M-CIDI)

(Wittchen and Pfister, 1997) was used to diagnose depressive, anxiety and
alcohol use disorders in the study. It is one of the most widely used
structured interviews in population surveys worldwide, including the WHO
World Mental Health Survey (Kessler and Ustun, 2004), and provides
excellent opportunities for cross-country comparison. The Munich version of
the CIDI has good test-retest reliability (Wittchen et al., 1998), and, when
compared with clinician diagnoses in a treatment setting, has excellent
sensitivity and specificity (95-100 and 93-100, respectively) (Reed et al.,
1998).
However, in a community sample, where disorders are less common than in

a clinical setting, the CIDI performs only moderately well. Two studies have
compared CIDI to the SCID, which is the gold standard of psychometric
instruments: the NCS-R and the WMHS. In these studies, the CIDI managed
to capture only 55-69% of persons with depression (sensitivity), and 95-97%
of the persons without depression (specificity). In the studied populations,
this led to rather low precision (positive predictive value), 50-55% (Haro et
al., 2006, Kessler et al., 2003).
Nevertheless, at the population-level, the CIDI neither inflates nor deflates

the annual prevalence of mental disorders, when compared with the SCID
(Haro et al., 2006). Therefore, our prevalence estimate can be considered an
accurate estimate of the prevalence of depressive disorders, as defined in the
diagnostic manuals. However, it is possible that some subjects identified as
“cases” of depressive disorders by the CIDI in this study would not have been
identified as such by a more detailed or semi-structured interview. The
opposite is also true: the CIDI may have missed some cases. It is difficult to
assess what, if any, impact this would have had on our results on risk factors
and prognosis.
6.5.3 NON-PARTICIPATION AND MULTIPLE IMPUTATION
Selective non-participation complicated the analyses in Sub-studies I-III.

Previous studies have shown that psychiatric disorders are associated with
non-participation (de Graaf et al., 2000, Eaton et al., 1992, Haapea et al.,
2008, Suvisaari et al., 2009). Therefore, the fact that the participation rate in
the mental health interview was only 57%, poses a challenge to the analysis
and interpretation of the results.
The method chosen to account for non-participation was multiple imputation

(MI), which is considered superior to other methods of handling missing
94
data (Li et al., 2015, Mackinnon, 2010). Register data, which was available
for all participants, was used in the imputation. The variables included in the
imputation model were chosen based on their correlation with both
missingness and with the outcomes, depressive disorders. Sensitivity
analyses were carried out with varying numbers of imputed data sets. The
results based on MI are thus considered more reliable and are reported as
primary results. In the case of Sub-studies II-III, there was little difference
between the weighted-only and imputed results. In Sub-study I, however, the
prevalence was very different, which is understandable given the increased
non-participation among people with hospitalisations for depression or other
mental disorders. For this reason, the results cannot be considered as reliable
as results based on a sample with a higher participation rate. Also, there was
considerable imprecision in the point estimates based on results where
imputation had a larger impact, such as prevalence among the older age
groups.
Finally, the register data used in the imputation only included people with
hospitalisations for mental disorders, and was not able to capture people in
outpatient treatment, or those without treatment. Also, lifetime instead of
past year treatments were utilised in the imputation. It is difficult to assess
the impact of these limitations of the register data on the imputed results.
However, the fact that any register data was available for imputation can be
considered an advantage over other, less complete sources of information for
imputation.
95
Conclusions
7 CONCLUSIONS
This study contributed to the existing body of literature documenting the

large burden of depressive disorders. It showed that depressive disorders are
common in the Finnish population and that their prevalence has increased
over the past decade; that certain negative experiences and individual
characteristics place people at a higher risk of developing these disorders;
that depressive disorders often have a chronic course and negative
consequences on the individual’s life and health; and that people with
depressive and alcohol use disorders have a higher mortality risk than the
general population.
The findings of this study are in line with earlier literature and bring clarity
to some issues where controversial findings existed. The impact of non-
participation on prevalence estimates should be taken into account in health
surveys, and also in estimates of burden of mental disorders. Unlike previous
studies from earlier decades, we found an increasing prevalence of depressive
disorders. Further studies should examine whether this is a global
phenomenon or specific to Finland.
In terms of reducing the burden of depressive disorders, it is crucial to focus

efforts on prevention. Several types of psychological interventions, both
targeted and universal, are effective in preventing incidence of depression
(van Zoonen et al., 2014). Based on this study, people with subclinical
depressive symptoms, anxiety disorders, low trust and multiple childhood
adversities are at a particularly high risk of developing depression and could
be suitable target groups for preventive interventions. In addition, welfare
policies should target families with children to minimise the impact of
parental illness, unemployment, financial difficulties and other adversities on
children (Wahlbeck and McDaid, 2012). Specific policies to reduce bullying
could be effective in reducing incidence of depression (Scott et al., 2014).
We found no evidence of low socioeconomic position being a risk factor for

depression, but dysthymia was more prevalent among those with low or
intermediate education. There was no evidence to support the causation
theory, where low socioeconomic position causes depression. Since
dysthymia was more prevalent among those with low education, and
previously low income has been associated with current depressive disorders
in the same survey, selection seems more likely than causation to explain the
link between depressive disorders and low SEP in the Finnish context.
Therefore, policy efforts should aim at protecting persons with persistent
depression from downward social mobility or failure to achieve desired
educational and income level. These efforts could include reduction of stigma
96
and discrimination in educational institutions and work places, and
improving early access to treatment (Muntaner et al., 2004). Stigma related
to depression is still very common, with 41% of people with MDD in high-
income countries reporting being “shunned” or avoided by others, and 23%
reporting discrimination in regard to their physical health problems
(Lasalvia et al., 2015).
As opposed to new-onset depression, we found few predictors of persistence

of depression. However, in clinical treatment settings, it is important to
recognise that severity of the disorder is an important predictor of the
longitudinal course, perhaps more so than comorbidity or other individual
characteristics. The fact that unmarried people are more likely to have a
persistent course of illness means that they warrant special attention in
treatment settings. Research efforts could be directed toward finding better
ways to distinguish those at a higher risk of adverse outcomes. Recently, a
more densely connected network of symptoms was found to predict
longitudinal course of depression in a network analysis using Bayesian
methods (van Borkulo et al., 2015). Different prediction algorithms have
been developed to predict risk of recurrence in depressed people (van Loo et
al., 2015a, Wang et al., 2014), where a wide variety of risk factors, including
demographic characteristics, level of depressive and anxiety symptoms,
psychiatric and family history, physical health and life events, contribute to
the risk of recurrence.
Beyond the diagnostic status and remission, we found broad and long-term
adverse outcomes in subclinical symptoms, health-related quality of life and
self-rated health. All of these were more pronounced in individuals with
dysthymia, even after recovery from the disorder was achieved. Since the
negative health impacts of depressive disorders persist more than a decade
after the initial diagnosis, a history of depressive disorder should be regarded
as a sign of increased health needs in the health care system, and a system of
periodical general health check-ups could be considered even after formal
recovery. Also self-care and monitoring in the recovery period could be
encouraged, and different online and other tools developed for this purpose.
As did numerous other studies, we found increased mortality among people

with depressive and alcohol use disorders, and the risk was increased despite
long follow-up and extensive controlling for confounders. Little doubt
remains over the excess mortality in mental disorders. In fact, as Graham
Thornicroft writes: “It has been clear for more than 50 years that people with
the more disabling forms of mental illness do not live as long as those
without mental illness. -- What are the implications of these sobering
findings? Three things are clear. Firstly, the time for descriptive research
alone is over. We now need evidence-based interventions that can reduce
excess mortality. -- Thirdly, this huge loss of life among people with mental
97
Conclusions
illness needs to be recognised as a human rights disgrace.” (Thornicroft,

2013) In the same spirit, Suetani comments: “Scarce research money would
be better invested in assessing how to implement proven interventions and in
developing better treatments and prevention strategies.” (Suetani et al.,
2015). These interventions to reduce excess mortality include adequate
treatment of depression, both antidepressants and psychological treatment
(Cuijpers and Schoevers, 2004). The physical health needs of persons with
mental disorders and the inequities in access to health care also need to be
adequately addressed (Lawrence and Kisely, 2010).
Based on this study, recommendations for further research include:
1. Assess recent trends of prevalence of depressive disorders in other

settings besides Finland;
2. Identify reasons for the increasing prevalence of depressive
disorders, particularly in women;
3. Examine ways to improve participation of people with mental
disorders in population studies and the statistical methods to
account for their non-participation;
4. Identify better predictors of different negative outcomes of
depressive disorders that can be used in the clinical practice;
5. Assess implementation and effectiveness of interventions to reduce
excess mortality in mental disorders.
Some policy recommendations based on this study are:

1. Target preventive programmes to people with subclinical
depressive symptoms, anxiety disorders, low trust, and childhood
adversities;
2. Sustain implementation of welfare policies that target families with
children to minimise the impact of parental illness, unemployment,
financial difficulties and other adversities, in order to prevent onset
of depressive disorders in these children. Implement policies to
reduce bullying and assess their impact on reducing incidence of
depression;
3. Protect individuals with depression from downward social
movement, or failure to achieve desired educational level, via
reduction of stigma and discrimination in educational institutions
and work places, facilitating return to work after sick leave, and
improving early access to treatment;
4. Offer special support to unmarried or lonely depressed individuals,
either at the treatment facilities or the non-governmental sector;
5. Implement policies to tackle the vast inequity in mortality between
people with mental disorders and those without. Address the
physical health needs of people with mental disorders, both during
the symptomatic phase and after formal recovery.
98
REFERENCES
Aalto, A. M., Elovainio, M., Kivimäki, M., Uutela, A. & Pirkola, S.

(2012). The Beck Depression Inventory and General Health Questionnaire as
measures of depression in the general population: a validation study using
the Composite International Diagnostic Interview as the gold standard.
Psychiatry Res 197, 163-71.
Agosti, V. (2014). Predictors of remission from chronic depression: a
prospective study in a nationally representative sample. Compr Psychiatry
55, 463-7.
Akbaraly, T. N., Brunner, E. J., Ferrie, J. E., Marmot, M. G.,
Kivimaki, M. & Singh-Manoux, A. (2009). Dietary pattern and
depressive symptoms in middle age. The British Journal of Psychiatry 195,
408-413.
Ali, F. A., Israr, S. M., Ali, B. S. & Janjua, N. Z. (2009). Association of
various reproductive rights, domestic violence and marital rape with
depression among Pakistani women. BMC Psychiatry 9, 77.
Anacker, C., Zunszain, P. A., Carvalho, L. A. & Pariante, C. M.
(2011). The glucocorticoid receptor: pivot of depression and of
antidepressant treatment? Psychoneuroendocrinology 36, 415-25.
Andersen, I., Thielen, K., Nygaard, E. & Diderichsen, F. (2009).
Social inequality in the prevalence of depressive disorders. Journal of
Epidemiology and Community Health 63, 575-581.
Andrade, L., Caraveo-Anduaga, J. J., Berglund, P., Bijl, R. V., De
Graaf, R., Vollebergh, W., Dragomirecka, E., Kohn, R., Keller, M.,
Kessler, R. C., Kawakami, N., Kilic, C., Offord, D., Ustun, T. B. &
Wittchen, H. U. (2003). The epidemiology of major depressive episodes:
results from the International Consortium of Psychiatric Epidemiology
(ICPE) Surveys. Int J Methods Psychiatr Res 12, 3-21.
Andrews, G., Henderson, S. & Hall, W. D. (2001). Prevalence,
comorbidity, disability and service utilisation. Overview of the Australian
National Mental Health Survey 178, 145-153.
Andrews, G. & Peters, L. (1998). The psychometric properties of the
Composite International Diagnostic Interview. Soc Psychiatry Psychiatr
Epidemiol 33, 80-8.
Anthony, J. C. & Petronis, K. R. (1991). Suspected risk factors for
depression among adults 18-44 years old. Epidemiology 2, 123-32.
Bakhshaie, J., Zvolensky, M. J. & Goodwin, R. D. (2015). Cigarette
smoking and the onset and persistence of depression among adults in the
United States: 1994-2005. Compr Psychiatry 60, 142-8.
Bank, T. W. (1993). World Development Report 1993: Investing in Health.
Baxter, A. J., Page, A. & Whiteford, H. A. (2011). Factors Influencing
Risk of Premature Mortality in Community Cases of Depression: A Meta-
Analytic Review.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. & Erbaugh, J.
(1961). An inventory for measuring depression. Arch Gen Psychiatry 4, 561-
71.
99
Conclusions
Berk, M., Williams, L. J., Jacka, F. N., O'Neil, A., Pasco, J. A.,
Moylan, S., Allen, N. B., Stuart, A. L., Hayley, A. C., Byrne, M. L. &
Maes, M. (2013). So depression is an inflammatory disease, but where does
the inflammation come from? BMC Med 11, 200.
Bijl, R. V., Ravelli, A. & van Zessen, G. (1998). Prevalence of psychiatric
disorder in the general population: results of The Netherlands Mental Health
Survey and Incidence Study (NEMESIS). Soc Psychiatry Psychiatr
Epidemiol 33, 587-95.
Björkenstam, E., Björkenstam, C., Holm, H., Gerdin, B. &
Ekselius, L. (2015). Excess cause-specific mortality in in-patient-treated
individuals with personality disorder: 25-year nationwide population-based
study. Br J Psychiatry 207, 339-45.
Blanco, C., Okuda, M., Markowitz, J. C., Liu, S. M., Grant, B. F. &
Hasin, D. S. (2010). The epidemiology of chronic major depressive disorder
and dysthymic disorder: results from the National Epidemiologic Survey on
Alcohol and Related Conditions. J Clin Psychiatry 71, 1645-56.
Boden, J. M. & Fergusson, D. M. (2011). Alcohol and depression.
Addiction 106, 906-14.
Boden, J. M., Fergusson, D. M. & Horwood, L. J. (2010). Cigarette
smoking and depression: tests of causal linkages using a longitudinal birth
cohort. Br J Psychiatry 196, 440-6.
Borrow, A. P. & Cameron, N. M. (2014). Estrogenic mediation of
serotonergic and neurotrophic systems: implications for female mood
disorders. Prog Neuropsychopharmacol Biol Psychiatry 54, 13-25.
Bowes, L., Joinson, C., Wolke, D. & Lewis, G. (2015). Peer
victimisation during adolescence and its impact on depression in early
adulthood: prospective cohort study in the United Kingdom. BMJ 350.
Breslau, J., Miller, E., Jin, R., Sampson, N. A., Alonso, J.,
Andrade, L. H., Bromet, E. J., de Girolamo, G., Demyttenaere, K.,
Fayyad, J., Fukao, A., Galaon, M., Gureje, O., He, Y., Hinkov, H.
R., Hu, C., Kovess-Masfety, V., Matschinger, H., Medina-Mora, M.
E., Ormel, J., Posada-Villa, J., Sagar, R., Scott, K. M. & Kessler, R.
C. (2011). A multinational study of mental disorders, marriage, and divorce.
Acta Psychiatr Scand 124, 474-86.
Bromet, E., Andrade, L. H., Hwang, I., Sampson, N. A., Alonso, J.,
de Girolamo, G., de Graaf, R., Demyttenaere, K., Hu, C., Iwata, N.,
Karam, A. N., Kaur, J., Kostyuchenko, S., Lepine, J. P., Levinson,
D., Matschinger, H., Mora, M. E., Browne, M. O., Posada-Villa, J.,
Viana, M. C., Williams, D. R. & Kessler, R. C. (2011). Cross-national
epidemiology of DSM-IV major depressive episode. BMC Med 9, 90.
Brugha, T., Bebbington, P., Singleton, N., Melzer, D., Jenkins, R.,
Lewis, G., Farrell, M., Bhugra, D., Lee, A. & Meltzer, H. (2004).
Trends in service use and treatment for mental disorders in adults
throughout Great Britain. Br J Psychiatry 185, 378-384.
Brugha, T. S., Bebbington, P. E. & Jenkins, R. (1999). A difference
that matters: comparisons of structured and semi-structured psychiatric
diagnostic interviews in the general population. Psychol Med 29, 1013-20.
Bulloch, A., Williams, J., Lavorato, D. & Patten, S. (2014).
Recurrence of major depressive episodes is strongly dependent on the
number of previous episodes. Depress Anxiety 31, 72-6.
100
Butterworth, P., Rodgers, B. & Windsor, T. D. (2009). Financial
hardship, socio-economic position and depression: Results from the PATH
Through Life Survey. Social Science & Medicine 69, 229-237.
Buttner, M. M., Mott, S. L., Pearlstein, T., Stuart, S., Zlotnick, C. &
O'Hara, M. W. (2013). Examination of premenstrual symptoms as a risk
factor for depression in postpartum women. Arch Womens Ment Health 16,
219-25.
Castaneda, A. E., Suvisaari, J., Marttunen, M., Perälä, J., Saarni,
S. I., Aalto-Setälä, T., Aro, H., Koskinen, S., Lönnqvist, J. & Tuulio-
Henriksson, A. (2008). Cognitive functioning in a population-based
sample of young adults with a history of non-psychotic unipolar depressive
disorders without psychiatric comorbidity. Journal of Affective Disorders
110, 36-45.
Chaiton, M. O., Cohen, J. E., O'Loughlin, J. & Rehm, J. (2009). A
systematic review of longitudinal studies on the association between
depression and smoking in adolescents. BMC Public Health 9, 356.
Chang, C. K., Hayes, R. D., Perera, G., Broadbent, M. T.,
Fernandes, A. C., Lee, W. E., Hotopf, M. & Stewart, R. (2011). Life
expectancy at birth for people with serious mental illness and other major
disorders from a secondary mental health care case register in London. PLoS
One 6, e19590.
Charlson, F. J., Ferrari, A. J., Flaxman, A. D. & Whiteford, H. A.
(2013). The epidemiological modelling of dysthymia: Application for the
Global Burden of Disease Study 2010. J Affect Disord 151, 111-20.
Chen, J., Cai, Y., Cong, E., Liu, Y., Gao, J., Li, Y., Tao, M., Zhang,
K., Wang, X., Gao, C., Yang, L., Li, K., Shi, J., Wang, G., Liu, L.,
Zhang, J., Du, B., Jiang, G., Shen, J., Zhang, Z., Liang, W., Sun, J.,
Hu, J., Liu, T., Wang, X., Miao, G., Meng, H., Li, Y., Hu, C., Li, Y.,
Huang, G., Li, G., Ha, B., Deng, H., Mei, Q., Zhong, H., Gao, S.,
Sang, H., Zhang, Y., Fang, X., Yu, F., Yang, D., Liu, T., Chen, Y.,
Hong, X., Wu, W., Chen, G., Cai, M., Song, Y., Pan, J., Dong, J.,
Pan, R., Zhang, W., Shen, Z., Liu, Z., Gu, D., Wang, X., Liu, X.,
Zhang, Q., Li, Y., Chen, Y., Kendler, K. S., Shi, S. & Flint, J. (2014).
Childhood sexual abuse and the development of recurrent major depression
in Chinese women. PLoS One 9, e87569.
Cole, D. A., Dukewich, T. L., Roeder, K., Sinclair, K. R., McMillan,
J., Will, E., Bilsky, S. A., Martin, N. C. & Felton, J. W. (2014). Linking
peer victimization to the development of depressive self-schemas in children
and adolescents. J Abnorm Child Psychol 42, 149-60.
Colman, I. & Ataullahjan, A. (2010). Life course perspectives on the
epidemiology of depression. Can J Psychiatry 55, 622-32.
Colman, I., Kingsbury, M., Garad, Y., Zeng, Y., Naicker, K., Patten,
S., Jones, P. B., Wild, T. C. & Thompson, A. H. (2015). Consistency in
adult reporting of adverse childhood experiences. Psychological Medicine
FirstView, 1-7.
Coupland, C. A., Dhiman, P., Barton, G., Morriss, R., Arthur, A.,
Sach, T. & Hippisley-Cox, J. (2011). A study of the safety and harms of
antidepressant drugs for older people: a cohort study using a large primary
care database. Health Technol Assess 15, 1-202, iii-iv.
101
Conclusions
Cradock-O'Leary, J., Young, A. S., Yano, E. M., Wang, M. & Lee, M.

L. (2002). Use of general medical services by VA patients with psychiatric
disorders. Psychiatr Serv 53, 874-8.
Crump, C., Sundquist, K., Winkleby, M. A. & Sundquist, J. (2013a).
Mental disorders and risk of accidental death. Br J Psychiatry 203, 297-302.
Crump, C., Sundquist, K., Winkleby, M. A. & Sundquist, J. (2013b).
Mental disorders and vulnerability to homicidal death: Swedish nationwide
cohort study. Bmj 346, f557.
Cuijpers, P. & Schoevers, R. A. (2004). Increased mortality in depressive
disorders: a review. Curr Psychiatry Rep 6, 430-7.
Cuijpers, P. & Smit, F. (2002). Excess mortality in depression: a meta-
analysis of community studies. J Affect Disord 72, 227-36.
Cuijpers, P., Vogelzangs, N., Twisk, J., Kleiboer, A., Li, J. &
Penninx, B. W. (2014a). Comprehensive meta-analysis of excess mortality
in depression in the general community versus patients with specific
illnesses. Am J Psychiatry 171, 453-62.
Cuijpers, P., Vogelzangs, N., Twisk, J., Kleiboer, A., Li, J. &
Penninx, B. W. (2014b). Is excess mortality higher in depressed men than
in depressed women? A meta-analytic comparison. J Affect Disord 161, 47-
54.
De Graaf, R., Bijl, R. V., Ravelli, A., Smit, F. & Vollebergh, W. A.
(2002). Predictors of first incidence of DSM-III-R psychiatric disorders in
the general population: findings from the Netherlands Mental Health Survey
and Incidence Study. Acta Psychiatr Scand 106, 303-13.
de Graaf, R., Bijl, R. V., Smit, F., Ravelli, A. & Vollebergh, W. A.
(2000). Psychiatric and sociodemographic predictors of attrition in a
longitudinal study: The Netherlands Mental Health Survey and Incidence
Study (NEMESIS). Am J Epidemiol 152, 1039-47.
de Graaf, R., ten Have, M., van Gool, C. & van Dorsselaer, S. (2012).
Prevalence of mental disorders and trends from 1996 to 2009. Results from
the Netherlands Mental Health Survey and Incidence Study-2. Soc
Psychiatry Psychiatr Epidemiol 47, 203-13.
de Graaf, R., van Dorsselaer, S., Tuithof, M. & ten Have, M. (2013).
Sociodemographic and psychiatric predictors of attrition in a prospective
psychiatric epidemiological study among the general population. Result of
the Netherlands Mental Health Survey and Incidence Study-2. Compr
Psychiatry 54, 1131-9.
De Silva, M. J., McKenzie, K., Harpham, T. & Huttly, S. R. (2005).
Social capital and mental illness: a systematic review. J Epidemiol
Community Health 59, 619-27.
Denollet, J., Maas, K., Knottnerus, A., Keyzer, J. J. & Pop, V. J.
(2009). Anxiety predicted premature all-cause and cardiovascular death in a
10-year follow-up of middle-aged women. J Clin Epidemiol 62, 452-6.
Dierker, L. C., Avenevoli, S., Stolar, M. & Merikangas, K. R. (2002).
Smoking and depression: an examination of mechanisms of comorbidity. Am
J Psychiatry 159, 947-53.
Dijkstra-Kersten, S. M. A., Biesheuvel-Leliefeld, K. E. M., van der
Wouden, J. C., Penninx, B. W. J. H. & van Marwijk, H. W. J. (2015).
Associations of financial strain and income with depressive and anxiety
disorders. Journal of Epidemiology and Community Health 69, 660-665.
102
DiMatteo, M. R., Lepper, H. S. & Croghan, T. W. (2000). Depression
is a risk factor for noncompliance with medical treatment: meta-analysis of
the effects of anxiety and depression on patient adherence. Arch Intern Med
160, 2101-7.
Dohrenwend, B., Levav, I., Shrout, P., Schwartz, S., Naveh, G.,
Link, B., Skodol, A. & Stueve, A. (1992). Socioeconomic status and
psychiatric disorders: the causation-selection issue. Science 255, 946-952.
Dowrick, C., Shiels, C., Page, H., Ayuso-Mateos, J. L., Casey, P.,
Dalgard, O. S., Dunn, G., Lehtinen, V., Salmon, P. & Whitehead,
M. (2011). Predicting long-term recovery from depression in community
settings in Western Europe: evidence from ODIN. Soc Psychiatry Psychiatr
Eaton, W. W., Anthony, J. C., Tepper, S. & Dryman, A. (1992).
Psychopathology and attrition in the epidemiologic catchment area surveys.
Am J Epidemiol 135, 1051-9.
Eaton, W. W., Martins, S. S., Nestadt, G., Bienvenu, O. J., Clarke,
D. & Alexandre, P. (2008a). The Burden of Mental Disorders. Epidemiol
Rev 30, 1-14.
Eaton, W. W., Muntaner, C., Bovasso, G. & Smith, C. (2001).
Socioeconomic status and depressive syndrome: the role of inter- and intra-
generational mobility, government assistance, and work environment. J
Health Soc Behav 42, 277-94.
Eaton, W. W., Shao, F., Nestadt , G., Lee, H. B., Bienvenu, O. J. &
Zandi, P. (2008b). Population-based study of first onset and chronicity in
major depressive disorder. Arch Gen Psychiatry 65, 513-20.
Eaton, W. W., Shao, H., Nestadt, G., Lee, B., Bienvenu, O. & Zandi,
P. (2008c). POpulation-based study of first onset and chronicity in major
depressive disorder. Archives of General Psychiatry 65, 513-520.
Edwards, A. C. & Kendler, K. S. (2012). A twin study of depression and
nicotine dependence: shared liability or causal relationship? J Affect Disord
142, 90-7.
Elovainio, M., Pulkki-Råback, L., Hakulinen, C., Ferrie, J. E.,
Jokela, M., Hintsanen, M., Raitakari, O. T. & Keltikangas-
Järvinen, L. (2015). Childhood and adolescence risk factors and
development of depressive symptoms: the 32-year prospective Young Finns
follow-up study. Journal of Epidemiology and Community Health 69, 1109-
1117.
Ernst, C., Schmid, G. & Angst, J. (1992). The Zurich Study. XVI. Early
antecedents of depression. A longitudinal prospective study on incidence in
young adults. Eur Arch Psychiatry Clin Neurosci 242, 142-51.
Everson-Rose, S. A., House, J. S. & Mero, R. P. (2004). Depressive
symptoms and mortality risk in a national sample: confounding effects of
health status. Psychosom Med 66, 823-30.
Fagiolini, A. & Goracci, A. (2009). The effects of undertreated chronic
medical illnesses in patients with severe mental disorders. J Clin Psychiatry
70 Suppl 3, 22-9.
Fekadu, A., Medhin, G., Kebede, D., Alem, A., Cleare, A. J., Prince,
M., Hanlon, C. & Shibre, T. (2015). Excess mortality in severe mental
illness: 10-year population-based cohort study in rural Ethiopia. Br J
103
Conclusions
Ferrari, A. J., Charlson, F. J., Norman, R. E., Flaxman, A. D.,

Patten, S. B., Vos, T. & Whiteford, H. A. (2013a). The epidemiological
modelling of major depressive disorder: application for the Global Burden of
Disease Study 2010. PLoS One 8.
Ferrari, A. J., Charlson, F. J., Norman, R. E., Patten, S. B.,
Freedman, G., Murray, C. J., Vos, T. & Whiteford, H. A. (2013b).
Burden of depressive disorders by country, sex, age, and year: findings from
the global burden of disease study 2010. PLoS Med 10, e1001547.
Ferrari, A. J., Somerville, A., Baxter, A. J., Norman, R. E., Patten,
S. B., Vos, T. & Whiteford, H. A. (2013c). Global variation in the
prevalence and incidence of major depressive disorder: a systematic review
of the epidemiological literature. Psychol Med 43, 471-81.
Fichter, M. M., Quadflieg, N., Fischer, U. C. & Kohlboeck, G. (2010).
Twenty-five-year course and outcome in anxiety and depression in the Upper
Bavarian Longitudinal Community Study. Acta Psychiatr Scand 122, 75-85.
Statistics Finland. (2014). Causes of death [e-publication]. (ed. S.
Finland). Statistics Finland.
Statistics Finland (2015). Koulutustasomittain (Indicator of level of
education).
First, M., Anthony, JC, Tepper, S, Dryman, A. (1997). Structured
Clinical Interview for DSM–IV Axis I Disorders, Research Version, Non-
patient Edition (SCID–I/NP). New York State Psychiatric Institute.
Flint, J. & Kendler, K. (2014). The Genetics of Major Depression. Neuron
81, 484-503.
Forsman, A. K., Nyqvist, F., Schierenbeck, I., Gustafson, Y. &
Wahlbeck, K. (2012). Structural and cognitive social capital and depression
among older adults in two Nordic regions. Aging Ment Health 16, 771-9.
Frank, E., Prien, R. F., Jarrett, R. B., Keller, M. B., Kupfer, D. J.,
Lavori, P. W., Rush, A. J. & Weissman, M. M. (1991).
Conceptualization and Rationale for Consensus Definitions of Terms in
Major Depressive Disorder: Remission, Recovery, Relapse and Recurrence.
Arch Gen Psychiatry 48, 851-855.
Frayne, S. M., Halanych, J. H., Miller, D. R., Wang, F., Lin, H.,
Pogach, L., Sharkansky, E. J., Keane, T. M., Skinner, K. M., Rosen,
C. S. & Berlowitz, D. R. (2005). Disparities in diabetes care: impact of
mental illness. Arch Intern Med 165, 2631-8.
Fried, E. I., Nesse, R. M., Zivin, K., Guille, C. & Sen, S. (2014).
Depression is more than the sum score of its parts: individual DSM
symptoms have different risk factors. Psychol Med 44, 2067-76.
Fryers, T., Melzer, D. & Jenkins, R. (2003). Social inequalities and the
common mental disorders: a systematic review of the evidence. Soc
Fujiwara, T. & Kawachi, I. (2008). A prospective study of individual-level
social capital and major depression in the United States. J Epidemiol
Community Health 62, 627-33.
Fuller-Thomson, E., Battiston, M., Gadalla, T. M. & Brennenstuhl,
S. (2014). Bouncing back: remission from depression in a 12-year panel
study of a representative Canadian community sample. Soc Psychiatry
Psychiatr Epidemiol 49, 903-10.
104
Furukawa, T. A., Fujita, A., Harai, H., Yoshimura, R., Kitamura, T.
& Takahashi, K. (2008). Definitions of recovery and outcomes of major
depression: results from a 10-year follow-up. Acta Psychiatr Scand 117, 35-
40.
Gan, Y., Gong, Y., Tong, X., Sun, H., Cong, Y., Dong, X., Wang, Y.,
Xu, X., Yin, X., Deng, J., Li, L., Cao, S. & Lu, Z. (2014). Depression
and the risk of coronary heart disease: a meta-analysis of prospective cohort
studies. BMC Psychiatry 14, 371.
Georgakis, M. K., Thomopoulos, T. P., Diamantaras, A. A.,
Kalogirou, E. I., Skalkidou, A., Daskalopoulou, S. S. & Petridou, E.
T. (2016). Association of Age at Menopause and Duration of Reproductive
Period With Depression After Menopause: A Systematic Review and Meta-
analysis. JAMA Psychiatry 73, 139-49.
Gilman, S. E., Trinh, N. H., Smoller, J. W., Fava, M., Murphy, J. M.
& Breslau, J. (2013). Psychosocial stressors and the prognosis of major
depression: a test of Axis IV. Psychol Med 43, 303-16.
Gissler, M., Laursen, T. M., Osby, U., Nordentoft, M. & Wahlbeck,
K. (2013). Patterns in mortality among people with severe mental disorders
across birth cohorts: a register-based study of Denmark and Finland in 1982-
2006. BMC Public Health 13, 834.
Global Burden of Disease Study 2013 Collaborators. (2015). Global,
regional, and national incidence, prevalence, and years lived with disability
for 301 acute and chronic diseases and injuries in 188 countries, 1990-2013:
a systematic analysis for the Global Burden of Disease Study 2013. Lancet
386, 743-800.
Gold, P. W. (2015). The organization of the stress system and its
dysregulation in depressive illness. Mol Psychiatry 20, 32-47.
Goldberg, D. P., Gater, R., Sartorius, N., Ustun, T. B., Piccinelli,
M., Gureje, O. & Rutter, C. (1997). The validity of two versions of the
GHQ in the WHO study of mental illness in general health care. Psychol Med
27, 191-7.
Gu, L., Xie, J., Long, J., Chen, Q., Chen, Q., Pan, R., Yan, Y., Wu,
G., Liang, B., Tan, J., Xie, X., Wei, B. & Su, L. (2013). Epidemiology of
major depressive disorder in mainland china: a systematic review. PLoS One
8, e65356.
Gureje, O. (2011). Dysthymia in a cross-cultural perspective. Curr Opin
Haapea, M., Miettunen, J., Läärä, E., Joukamaa, M., Järvelin, M.,
Isohanni , M. & Veijola, J. M. (2008). Non-participation in a field survey
with respect to psychiatric disorders. Scand J Public Health 36, 728-36.
Haapea, M., Miettunen, J., Veijola, J., Lauronen, E., Tanskanen,
P. & Isohanni, M. (2007). Non-participation may bias the results of a
psychiatric survey: an analysis from the survey including magnetic resonance
imaging within the Northern Finland 1966 Birth Cohort. Soc Psychiatry
Hajek, A., Brettschneider, C., Ernst, A., Lange, C., Wiese, B.,
Prokein, J., Weyerer, S., Werle, J., Pentzek, M., Fuchs, A., Stein,
J., Bickel, H., Mosch, E., Heser, K., Jessen, F., Maier, W., Scherer,
M., Riedel-Heller, S. G. & Konig, H. H. (2015). Complex coevolution of
105
Conclusions
depression and health-related quality of life in old age. Qual Life Res 24,
2713-22.
Hardeveld, F., Spijker, J., de Graaf, R., Nolen, W. A. & Beekman,
A. T. (2010). Prevalence and predictors of recurrence of major depressive
disorder in the adult population. Acta Psychiatr Scand 122, 184-91.
Hardeveld, F., Spijker, J., de Graaf, R., Nolen, W. A. & Beekman,
A. T. (2013). Recurrence of major depressive disorder and its predictors in
the general population: results from the Netherlands Mental Health Survey
and Incidence Study (NEMESIS). Psychol Med 43, 39-48.
Härkänen, T., Karvanen, J., Tolonen, H., Lehtonen, R., Djerf, K.,
Juntunen, T. & Koskinen, S. (2016). Systematic handling of missing data
in complex study designs - experiences from the Health 2000 and 2011
Surveys. Journal of Applied Statistics (accepted for publication).
Harkness, K. L., Alavi, N., Monroe, S. M., Slavich, G. M., Gotlib, I.
H. & Bagby, R. M. (2010). Gender differences in life events prior to onset
of major depressive disorder: the moderating effect of age. J Abnorm Psychol
119, 791-803.
Haro, J. M., Arbabzadeh-Bouchez, S., Brugha, T., de Girolamo, G.,
Guyer, M. E., Jin, R., Lepine, J. P., Mazzi, F., Reneses, B., Vilagut,
G., Sampson, N. & Kessler, R. C. (2006). Concordance of the Composite
International Diagnostic Interview Version 3.0 (CIDI 3.0) with standardized
clinical assessments in the WHO World Mental Health surveys. Int J
Methods Psychiatr Res 15, 167-80.
Harris, E. C. & Barraclough, B. (1998). Excess mortality of mental
disorder. Br J Psychiatry 173, 11-53.
Hasin, D. S., Goodwin, R. D., Stinson, F. S. & Grant, B. F. (2005).
Epidemiology of major depressive disorder: results from the National
Epidemiologic Survey on Alcoholism and Related Conditions. Arch Gen
Hawkley, L. C. & Cacioppo, J. T. (2010). Loneliness matters: a
theoretical and empirical review of consequences and mechanisms. Ann
Behav Med 40, 218-27.
Hawton, K., Casanas, I. C. C., Haw, C. & Saunders, K. (2013). Risk
factors for suicide in individuals with depression: a systematic review. J
Affect Disord 147, 17-28.
Hayes, J. F., Miles, J., Walters, K., King, M. & Osborn, D. P. (2015).
A systematic review and meta-analysis of premature mortality in bipolar
affective disorder. Acta Psychiatr Scand 131, 417-25.
Heiskanen, T. H., Koivumaa-Honkanen, H. T., Niskanen, L. K.,
Lehto, S. M., Honkalampi, K. M., Hintikka, J. J. & Viinamaki, H. T.
(2013). Depression and major weight gain: a 6-year prospective follow-up of
outpatients. Compr Psychiatry 54, 599-604.
Heistaro, S. e. (2008). Methodology Report. Health 2000 survey. In
Publications of the National Public Health Institute, KTL (ed. N. P. H.
Institute), p. 248. National Public Health Institute: Helsinki.
Hiroeh, U., Kapur, N., Webb, R., Dunn, G., Mortensen, P. B. &
Appleby, L. (2008). Deaths from natural causes in people with mental
illness: a cohort study. J Psychosom Res 64, 275-83.
106
Holma, K. M., Holma, I. A., Melartin, T. K., Rytsälä, H. J. &
Isometsä, E. T. (2008). Long-term outcome of major depressive disorder
in psychiatric patients is variable. J Clin Psychiatry 69, 196-205.
Holma, K. M., Melartin, T. K., Haukka, J., Holma, I. A., Sokero, T.
P. & Isometsä, E. T. (2010). Incidence and predictors of suicide attempts
in DSM-IV major depressive disorder: a five-year prospective study. Am J
Holmstrand, C., Bogren, M., Mattisson, C. & Bradvik, L. (2015).
Long-term suicide risk in no, one or more mental disorders: the Lundby
Study 1947-1997. Acta Psychiatr Scand.
Holwerda, T. J., Schoevers, R. A., Dekker, J., Deeg, D. J., Jonker,
C. & Beekman, A. T. (2007). The relationship between generalized anxiety
disorder, depression and mortality in old age. Int J Geriatr Psychiatry 22,
241-9.
Horwath, E., Johnson, J., Klerman, G. L. & Weissman, M. M.
(1992). DEpressive symptoms as relative and attributable risk factors for
first-onset major depression. Archives of General Psychiatry 49, 817-823.
Hovens, J. G. F. M., Giltay, E. J., van Hemert, A. M. & Penninx, B.
W. J. H. (2016). Childhood maltreatment and the course of depressive and
anxiety disorders: the contribution of personality characteristics. Depression
and Anxiety 33, 27-34.
Irwin, M. R. & Cole, S. W. (2011). Reciprocal regulation of the neural and
innate immune systems. Nat Rev Immunol 11, 625-32.
Isometsä, E. (2014). Suicidal behaviour in mood disorders--who, when,
and why? Can J Psychiatry 59, 120-30.
Jacobi, F., Hofler, M., Strehle, J., Mack, S., Gerschler, A., Scholl,
L., Busch, M. A., Hapke, U., Maske, U., Seiffert, I., Gaebel, W.,
Maier, W., Wagner, M., Zielasek, J. & Wittchen, H. U. (2015).
Twelve-months prevalence of mental disorders in the German Health
Interview and Examination Survey for Adults - Mental Health Module
(DEGS1-MH): a methodological addendum and correction. Int J Methods
Psychiatr Res.
Jenkins, J. M. & Curwen, T. (2008). Change in adolescents' internalizing
symptomatology as a function of sex and the timing of maternal depressive
symptomatology. J Am Acad Child Adolesc Psychiatry 47, 399-405.
Joffe, H., Chang, Y., Dhaliwal, S., Hess, R., Thurston, R., Gold, E.,
Matthews, K. A. & Bromberger, J. T. (2012). Lifetime history of
depression and anxiety disorders as a predictor of quality of life in midlife
women in the absence of current illness episodes. Arch Gen Psychiatry 69,
484-92.
John, U., Rumpf, H. J., Bischof, G., Hapke, U., Hanke, M. & Meyer,
C. (2013). Excess mortality of alcohol-dependent individuals after 14 years
and mortality predictors based on treatment participation and severity of
alcohol dependence. Alcohol Clin Exp Res 37, 156-63.
Joinson, C., Heron, J., Araya, R. & Lewis, G. (2013). Early menarche
and depressive symptoms from adolescence to young adulthood in a UK
cohort. J Am Acad Child Adolesc Psychiatry 52, 591-8.e2.
Jokela, M., Virtanen, M., Batty, G. D. & Kivimaki, M. (2015).
Inflammation and Specific Symptoms of Depression. JAMA Psychiatry, 1-2.
107
Conclusions
Jones, R., Heim, D., Hunter, S. & Ellaway, A. (2014). The relative
influence of neighbourhood incivilities, cognitive social capital, club
membership and individual characteristics on positive mental health. Health
Place 28, 187-93.
Joutsenniemi, K., Martelin, T., Martikainen, P., Pirkola, S. &
Koskinen, S. (2006). Living arrangements and mental health in Finland. J
Epidemiol Community Health 60, 468-75.
Judd, L. L., Akiskal, H. S., Zeller, P. J., Paulus, M., Leon, A. C.,
Maser, J. D., Endicott, J., Coryell, W., Kunovac, J. L., Mueller, T.
I., Rice, J. P. & Keller, M. B. (2000). Psychosocial Disability During the
Long-term Course of Unipolar Major Depressive Disorder. Arch Gen
Kananen, L., Surakka, I., Pirkola, S., Suvisaari, J., Lönnqvist, J.,
Peltonen, L., Ripatti, S. & Hovatta, I. (2010). Childhood adversities are
associated with shorter telomere length at adult age both in individuals with
an anxiety disorder and controls. PLoS One 25.
Kaplan, G. A., Roberts, R. E., Camacho, T. C. & Coyne, J. C. (1987).
Psychosocial predictors of depression. Prospective evidence from the human
population laboratory studies. Am J Epidemiol 125, 206-20.
Karasz, A. (2005). Cultural differences in conceptual models of depression.
Soc Sci Med 60, 1625-35.
Kendler, K. S., Aggen, S. H. & Neale, M. C. (2013). EVidence for
multiple genetic factors underlying dsm-iv criteria for major depression.
JAMA Psychiatry 70, 599-607.
Kendler, K. S. & Gardner, C. O. (2014). Sex differences in the pathways
to major depression: a study of opposite-sex twin pairs. Am J Psychiatry 171,
426-35.
Kendler, K. S., Gardner, C. O. & Prescott, C. A. (2002). Toward a
comprehensive developmental model for major depression in women. Am J
Kendler, K. S., Gardner, C. O. & Prescott, C. A. (2006). Toward a
comprehensive developmental model for major depression in men. Am J
Kendler, K. S., Myers, J. & Zisook, S. (2008). Does bereavement-
related major depression differ from major depression associated with other
stressful life events? Am J Psychiatry 165, 1449-55.
Kessler, R. C., Abelson, J., Demler, O., Escobar, J. I., Gibbon, M.,
Guyer, M. E., Howes, M. J., Jin, R., Vega, W. A., Walters, E. E.,
Wang, P., Zaslavsky, A. & Zheng, H. (2004). Clinical calibration of
DSM-IV diagnoses in the World Mental Health (WMH) version of the World
Health Organization (WHO) Composite International Diagnostic Interview
(WMHCIDI). Int J Methods Psychiatr Res 13, 122-39.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Koretz, D.,
Merikangas, K. R., Rush, A. J., Walters, E. E. & Wang, P. S. (2003).
The epidemiology of major depressive disorder: results from the National
Comorbidity Survey Replication (NCS-R). JAMA 289, 3095-105.
Kessler, R. C. & Bromet, E. J. (2013). The epidemiology of depression
across cultures. Annu Rev Public Health 34, 119-38.
Kessler, R. C., Chiu, W. T., Demler, O., Merikangas, K. R. &
Walters, E. E. (2005). Prevalence, severity, and comorbidity of 12-month
108
DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen
Kessler, R. C., Gruber, M., Hettema, J. M., Hwang, I., Sampson, N.
& Yonkers, K. A. (2008). Co-morbid major depression and generalized
anxiety disorders in the National Comorbidity Survey follow-up. Psychol
Med 38, 365-74.
Kessler, R. C. & Ustun, T. B. (2004). The World Mental Health (WMH)
Survey Initiative Version of the World Health Organization (WHO)
Composite International Diagnostic Interview (CIDI). Int J Methods
Psychiatr Res 13, 93-121.
Kiecolt-Glaser, J. K., Derry, H. M. & Fagundes, C. P. (2015).
Inflammation: Depression Fans the Flames and Feasts on the Heat. Am J
Kiecolt-Glaser, J. K. & Glaser, R. (2002). Depression and immune
function: central pathways to morbidity and mortality. J Psychosom Res 53,
873-6.
Kigar, S. L. & Auger, A. P. (2013). Epigenetic mechanisms may underlie
the aetiology of sex differences in mental health risk and resilience. J
Neuroendocrinol 25, 1141-50.
Kind, P., Hardman, G. & Macran, S. (1999). UK Population Norms for
EQ-5D. In Discussion Paper 172. Centre for Health Economics, University of
York.
Kisely, S., Smith, M., Lawrence, D., Cox, M., Campbell, L. A. &
Maaten, S. (2007). Inequitable access for mentally ill patients to some
medically necessary procedures. CMAJ 176, 779-84.
Kivelä, S. L., Köngäs-Saviaro, P., Laippala, P., Pahkala, K. & Kesti,
E. (1996). Social and psychosocial factors predicting depression in old age: a
longitudinal study. Int Psychogeriatr 8, 635-44.
Klein, D. N., Glenn, C. R., Kosty, D. B., Seeley, J. R., Rohde, P. &
Lewinsohn, P. M. (2013). Predictors of first lifetime onset of major
depressive disorder in young adulthood. J Abnorm Psychol 122, 1-6.
Klein, D. N., Shankman, S. A. & Rose, S. (2006). Ten-year prospective
follow-up study of the naturalistic course of dysthymic disorder and double
depression. Am J Psychiatry 163, 872-80.
Klein, D. N., Shankman, S. A. & Rose, S. (2008). Dysthymic disorder
and double depression: prediction of 10-year course trajectories and
outcomes. J Psychiatr Res 42, 408-15.
Klein, J. P., Roniger, A., Schweiger, U., Spath, C. & Brodbeck, J.
(2015). The association of childhood trauma and personality disorders with
chronic depression: A cross-sectional study in depressed outpatients. J Clin
Psychiatry 76, e794-801.
Koivumaa-Honkanen, H., Tuovinen, T., Honkalampi, K.,
Antikainen, R., Hintikka, J., Haatainen, K. & Viinamäki, H. (2008).
Mental health and well-being in a 6-year follow-up of patients with
depression. Social Psychiatry and Psychiatric Epidemiology 43, 688-696.
Kosidou, K., Dalman, C., Lundberg, M., Hallqvist, J., Isacsson, G.
& Magnusson, C. (2011). Socioeconomic status and risk of psychological
distress and depression in the Stockholm Public Health Cohort: a
population-based study. J Affect Disord 134, 160-7.
109
Conclusions
Koskinen, S., Lundqvist, A. and Ristiluoma, N. (2012). Terveys,

toimintakyky ja hyvinvointi Suomessa 2011. Terveyden ja hyvinvoinnin
laitos: Tampere.
Kruijshaar, M. E., Barendregt, J., Vos, T., de Graaf, R., Spijker, J.
& Andrews, G. (2005). Lifetime prevalence estimates of major depression:
an indirect estimation method and a quantification of recall bias. Eur J
Kuk, A. Y., Li, J. & Rush, A. J. (2010). Recursive subsetting to identify
patients in the STAR*D: a method to enhance the accuracy of early
prediction of treatment outcome and to inform personalized care. J Clin
Lahelma, E., Laaksonen, M., Martikainen, P., Rahkonen, O. &
Sarlio-Lahteenkorva, S. (2006). Multiple measures of socioeconomic
circumstances and common mental disorders. Soc Sci Med 63, 1383-99.
Lamers, F., Beekman, A. T. F., van Hemert, A. M., Schoevers, R. A.
& Penninx, B. W. J. H. (2016). Six-year longitudinal course and outcomes
of subtypes of depression. The British Journal of Psychiatry 208, 62-68.
Lamers, F., Hoogendoorn, A. W., Smit, J. H., van Dyck, R.,
Zitman, F. G., Nolen, W. A. & Penninx, B. W. (2012).
Sociodemographic and psychiatric determinants of attrition in the
Netherlands Study of Depression and Anxiety (NESDA). Compr Psychiatry
53, 63-70.
Laramée, P., Leonard, S., Buchanan-Hughes, A., Warnakula, S.,
Daeppen, J.-B. & Rehm, J. (2015). Risk of All-Cause Mortality in Alcohol-
Dependent Individuals: A Systematic Literature Review and Meta-Analysis.
EBioMedicine 2, 1394-1404.
Lasalvia, A., Van Bortel, T., Bonetto, C., Jayaram, G., van
Weeghel, J., Zoppei, S., Knifton, L., Quinn, N., Wahlbeck, K.,
Cristofalo, D., Lanfredi, M., Sartorius, N. & Thornicroft, G. (2015).
Cross-national variations in reported discrimination among people treated
for major depression worldwide: the ASPEN/INDIGO international study.
The British Journal of Psychiatry 207, 507-514.
Laursen, T. M., Wahlbeck, K., Hallgren, J., Westman, J., Osby, U.,
Alinaghizadeh, H., Gissler, M. & Nordentoft, M. (2013). Life
expectancy and death by diseases of the circulatory system in patients with
bipolar disorder or schizophrenia in the Nordic countries. PLoS One 8,
e67133.
Lawrence, D. & Kisely, S. (2010). Inequalities in healthcare provision for
people with severe mental illness. J Psychopharmacol 24, 61-8.
Lee, J. (1981). Covariance adjustment of rates based on the multiple logistic
regression model. J Chronic Dis 34, 415-26.
Lee, S., Tsang, A., Huang, Y. Q., He, Y. L., Liu, Z. R., Zhang, M. Y.,
Shen, Y. C. & Kessler, R. C. (2009). The epidemiology of depression in
metropolitan China. Psychol Med 39, 735-47.
Lehtinen, V., Joukamaa, M., Lahtela, K., Raitasalo, R., Jyrkinen,
E., Maatela, J. & Aromaa, A. (1990). Prevalence of mental disorders
among adults in Finland: basic results from the Mini Finland Health Survey.
Acta Psychiatr Scand 81, 418-25.
Lehtinen, V., Sohlman, B., Nummelin, T., Salomaa, M., Ayuso-
Mateos, J. L. & Dowrick, C. (2005). The estimated incidence of
110
depressive disorder and its determinants in the Finnish ODIN sample. Soc
Lehto, S. M., Heiskanen, T., Hintikka, J., Niskanen, L., Koivumaa-
Honkanen, H., Tolmunen, T., Honkalampi, K. & Viinamäki, H.
(2008a). Metabolic syndrome--the impact of depression. Ann Epidemiol 18,
871.
Lehto, S. M., Hintikka, J., Niskanen, L., Tolmunen, T., Koivumaa-
Honkanen, H., Honkalampi, K. & Viinamäki, H. (2008b). Low HDL
cholesterol associates with major depression in a sample with a 7-year
history of depressive symptoms. Prog Neuropsychopharmacol Biol
Leinonen, T., Martikainen, P., Laaksonen, M. & Lahelma, E. (2014).
Excess mortality after disability retirement due to mental disorders:
variations by socio-demographic factors and causes of death. Soc Psychiatry
Leskelä, U., Rytsälä, H., Komulainen, E., Melartin, T. K., Sokero,
T. P., Lestelä-Mielonen, P. S. & Isometsä, E. (2006). The influence of
adversity and perceived social support on the outcome of major depressive
disorder in subjects with different levels of depressive symptoms. Psychol
Med 36, 779-88.
Lev-Ran, S., Roerecke, M., Le Foll, B., George, T. P., McKenzie, K.
& Rehm, J. (2014). The association between cannabis use and depression: a
systematic review and meta-analysis of longitudinal studies. Psychol Med
44, 797-810.
Li, P., Stuart, E. A. & Allison, D. B. (2015). Multiple imputation: A
flexible tool for handling missing data. JAMA 314, 1966-1967.
Lichtman, J. H., Froelicher, E. S., Blumenthal, J. A., Carney, R.
M., Doering, L. V., Frasure-Smith, N., Freedland, K. E., Jaffe, A.
S., Leifheit-Limson, E. C., Sheps, D. S., Vaccarino, V. & Wulsin, L.
(2014). Depression as a risk factor for poor prognosis among patients with
acute coronary syndrome: systematic review and recommendations: a
scientific statement from the American Heart Association. Circulation 129,
1350-69.
Lindeman, S., Hämäläinen, J., Isometsä, E., Kaprio, J.,
Poikolainen, K., Heikkinen, M. & Aro, H. (2000). The 12-month
prevalence and risk factors for major depressive episode in
Finland: representative sample of 5993 adults. Acta Psychiatr Scand 102,
178-84.
Lorant, V., Croux, C., Weich, S., Deliege, D., Mackenbach, J. &
Ansseau, M. (2007). Depression and socio-economic risk factors: 7-year
longitudinal population study. Br J Psychiatry 190, 293-8.
Lorant, V., Deliege, D., Eaton, W., Robert, A., Philippot, P. &
Ansseau, M. (2003). Socioeconomic inequalities in depression: a meta-
analysis. Am J Epidemiol 157, 98-112.
Loret de Mola, C., de Franca, G. V., Quevedo Lde, A. & Horta, B. L.
(2014). Low birth weight, preterm birth and small for gestational age
association with adult depression: systematic review and meta-analysis. Br J
Luppino, F. S., de Wit, L. M., Bouvy, P. F., Stijnen, T., Cuijpers, P.,
Penninx, B. W. & Zitman, F. G. (2010). Overweight, obesity, and
111
Conclusions
depression: a systematic review and meta-analysis of longitudinal studies.

Arch Gen Psychiatry 67, 220-9.
Luutonen, S., Sohlman, B., Salokangas, R. K., Lehtinen, V. &
Dowrick, C. (2011). Weak sense of coherence predicts depression: 1-year
and 9-year follow-ups of the Finnish Outcomes of Depression International
Network (ODIN) sample. J Ment Health 20, 43-51.
Mackinnon, A. (2010). The use and reporting of multiple imputation in
medical research - a review. J Intern Med 268, 586-93.
Martens, E. J., de Jonge, P., Na, B., Cohen, B. E., Lett, H. &
Whooley, M. A. (2010). Scared to death? Generalized anxiety disorder and
cardiovascular events in patients with stable coronary heart disease:The
Heart and Soul Study. Arch Gen Psychiatry 67, 750-8.
Mattisson, C., Bogren, M., Horstmann, V., Munk-Jorgensen, P. &
Nettelbladt, P. (2007). The long-term course of depressive disorders in the
Lundby Study. Psychol Med 37, 883-91.
McCabe, P. J. & Christopher, P. P. (2015). Symptom and functional
traits of brief major depressive episodes and discrimination of bereavement.
Depress Anxiety.
McIntyre, R. S., Cha, D. S., Soczynska, J. K., Woldeyohannes, H.
O., Gallaugher, L. A., Kudlow, P., Alsuwaidan, M. & Baskaran, A.
(2013). Cognitive deficits and functional outcomes in major depressive
disorder: determinants, substrates, and treatment interventions. Depress
Anxiety 30, 515-27.
Menear, M., Dore, I., Cloutier, A. M., Perrier, L., Roberge, P.,
Duhoux, A., Houle, J. & Fournier, L. (2015). The influence of comorbid
chronic physical conditions on depression recognition in primary care: a
systematic review. J Psychosom Res 78, 304-13.
Meng, X. & D'Arcy, C. (2014). The projected effect of risk factor reduction
on major depression incidence: a 16-year longitudinal Canadian cohort of the
National Population Health Survey. J Affect Disord 158, 56-61.
Mezuk, B. & Kendler, K. S. (2012). Examining variation in depressive
symptoms over the life course: a latent class analysis. Psychol Med 42, 2037-
46.
Miettunen, J., Isohanni, M., Paunio, T., Freimer, N., Taanila, A.,
Ekelund, J., Järvelin, M. R., Joukamaa, M., Lichtermann, D.,
Koivumaa-Honkanen, H. & Veijola, J. (2012). Predicting depression
with psychopathology and temperament traits: the northern Finland 1966
birth cohort. Depress Res Treat 2012, 160905.
Miettunen, J. & Raevuori, A. (2012). A meta-analysis of temperament in
axis I psychiatric disorders. Compr Psychiatry 53, 152-66.
Mitchell, A. J. & Lawrence, D. (2011). Revascularisation and mortality
rates following acute coronary syndromes in people with severe mental
illness: comparative meta-analysis. Br J Psychiatry 198, 434-41.
Moffitt, T. E., Caspi, A., Taylor, A., Kokaua, J., Milne, B. J.,
Polanczyk, G. & Poulton, R. (2010). How common are common mental
disorders? Evidence that lifetime prevalence rates are doubled by prospective
versus retrospective ascertainment. Psychol Med 40, 899-909.
Moreno-Peral, P., Conejo-Ceron, S., Motrico, E., Rodriguez-
Morejon, A., Fernandez, A., Garcia-Campayo, J., Roca, M.,
Serrano-Blanco, A., Rubio-Valera, M. & Bellon, J. A. (2014). Risk
112
factors for the onset of panic and generalised anxiety disorders in the general
adult population: a systematic review of cohort studies. J Affect Disord 168,
337-48.
Moscati, A., Flint, J. & Kendler, K. S. (2015). Classification of anxiety
disorders comorbid with major depression: common or distinct influences on
risk? Depress Anxiety.
Moustgaard, H., Joutsenniemi, K., Sihvo, S. & Martikainen, P.
(2013). Alcohol-related deaths and social factors in depression mortality: a
register-based follow-up of depressed in-patients and antidepressant users in
Finland. J Affect Disord 148, 278-85.
Mueller, T. I., Leon, A. C., Keller, M. B., Solomon, D. A., Endicott,
J., Coryell, W., Warshaw, M. & Maser, J. D. (1999). Recurrence after
recovery from major depressive disorder during 15 years of observational
follow-up. Am J Psychiatry 156, 1000-6.
Muntaner, C., Eaton, W. W., Miech, R. & O'Campo, P. (2004).
Socioeconomic position and major mental disorders. Epidemiol Rev 26, 53-
62.
Murphy, J. A. & Byrne, G. J. (2012). Prevalence and correlates of the
proposed DSM-5 diagnosis of Chronic Depressive Disorder. J Affect Disord
139, 172-80.
Murphy, J. M., Laird, N. M., Monson, R. R., Sobol, A. M. &
Leighton, A. H. (2000). A 40-Year Perspective on the Prevalence of
Depression. Arch Gen Psychiatry 57, 209-215.
Murphy, J. M., Olivier, D. C., Sobol, A. M., Monson, R. R. &
Leighton, A. H. (1986). Diagnosis and outcome: depression and anxiety in
a general population. Psychol Med 16, 117-26.
Murray, C. J. L., Barber, R. M., Foreman, K. J., Ozgoren, A. A.,
Abd-Allah, F., Abera, S. F., Aboyans, V., Abraham, J. P., Abubakar,
I., Abu-Raddad, L. J., Abu-Rmeileh, N. M., Achoki, T., Ackerman,
I. N., Ademi, Z., Adou, A. K., Adsuar, J. C., Afshin, A., Agardh, E.
E., Alam, S. S., Alasfoor, D., Albittar, M. I., Alegretti, M. A., Alemu,
Z. A., Alfonso-Cristancho, R., Alhabib, S., Ali, R., Alla, F., Allebeck,
P., Almazroa, M. A., Alsharif, U., Alvarez, E., Alvis-Guzman, N.,
Amare, A. T., Ameh, E. A., Amini, H., Ammar, W., Anderson, H.
R., Anderson, B. O., Antonio, C. A. T., Anwari, P., Arnlöv, J.,
Arsenijevic, V. S. A., Artaman, A., Asghar, R. J., Assadi, R., Atkins,
L. S., Avila, M. A., Awuah, B., Bachman, V. F., Badawi, A., Bahit,
M. C., Balakrishnan, K., Banerjee, A., Barker-Collo, S. L.,
Barquera, S., Barregard, L., Barrero, L. H., Basu, A., Basu, S.,
Basulaiman, M. O., Beardsley, J., Bedi, N., Beghi, E., Bekele, T.,
Bell, M. L., Benjet, C., Bennett, D. A., Bensenor, I. M., Benzian, H.,
Bernabé, E., Bertozzi-Villa, A., Beyene, T. J., Bhala, N., Bhalla, A.,
Bhutta, Z. A., Bienhoff, K., Bikbov, B., Biryukov, S., Blore, J. D.,
Blosser, C. D., Blyth, F. M., Bohensky, M. A., Bolliger, I. W.,
Başara, B. B., Bornstein, N. M., Bose, D., Boufous, S., Bourne, R.
R. A., Boyers, L. N., Brainin, M., Brayne, C. E., Brazinova, A.,
Breitborde, N. J. K., Brenner, H., Briggs, A. D., Brooks, P. M.,
Brown, J. C., Brugha, T. S., Buchbinder, R., Buckle, G. C., Budke,
C. M., Bulchis, A., Bulloch, A. G., Campos-Nonato, I. R., Carabin,
H., Carapetis, J. R., Cárdenas, R., Carpenter, D. O., Caso, V.,
113
Conclusions
Castañeda-Orjuela, C. A., Castro, R. E., Catalá-López, F., Cavalleri,

F., Çavlin, A., Chadha, V. K., Chang, J.-C., Charlson, F. J., Chen,
H., Chen, W., Chiang, P. P., Chimed-Ochir, O., Chowdhury, R.,
Christensen, H., Christophi, C. A., Cirillo, M., Coates, M. M.,
Coffeng, L. E., Coggeshall, M. S., Colistro, V., Colquhoun, S. M.,
Cooke, G. S., Cooper, C., Cooper, L. T., Coppola, L. M., Cortinovis,
M., Criqui, M. H., Crump, J. A., Cuevas-Nasu, L., Danawi, H.,
Dandona, L., Dandona, R., Dansereau, E., Dargan, P. I., Davey, G.,
Davis, A., Davitoiu, D. V., Dayama, A., De Leo, D., Degenhardt, L.,
Del Pozo-Cruz, B., Dellavalle, R. P., Deribe, K., Derrett, S., Jarlais,
D. C. D., Dessalegn, M., Dharmaratne, S. D., Dherani, M. K., Diaz-
Torné, C., Dicker, D., Ding, E. L., Dokova, K., Dorsey, E. R.,
Driscoll, T. R., Duan, L., Duber, H. C., Ebel, B. E., Edmond, K. M.,
Elshrek, Y. M., Endres, M., Ermakov, S. P., Erskine, H. E., Eshrati,
B., Esteghamati, A., Estep, K., Faraon, E. J. A., Farzadfar, F., Fay,
D. F., Feigin, V. L., Felson, D. T., Fereshtehnejad, S.-M.,
Fernandes, J. G., Ferrari, A. J., Fitzmaurice, C., Flaxman, A. D.,
Fleming, T. D., Foigt, N., Forouzanfar, M. H., Fowkes, F. G. R.,
Paleo, U. F., Franklin, R. C., Fürst, T., Gabbe, B., Gaffikin, L.,
Gankpé, F. G., Geleijnse, J. M., Gessner, B. D., Gething, P., Gibney,
K. B., Giroud, M., Giussani, G., Dantes, H. G., Gona, P., González-
Medina, D., Gosselin, R. A., Gotay, C. C., Goto, A., Gouda, H. N.,
Graetz, N., Gugnani, H. C., Gupta, R., Gupta, R., Gutiérrez, R. A.,
Haagsma, J., Hafezi-Nejad, N., Hagan, H., Halasa, Y. A.,
Hamadeh, R. R., Hamavid, H., Hammami, M., Hancock, J.,
Hankey, G. J., Hansen, G. M., Hao, Y., Harb, H. L., Haro, J. M.,
Havmoeller, R., Hay, S. I., Hay, R. J., Heredia-Pi, I. B., Heuton, K.
R., Heydarpour, P., Higashi, H., Hijar, M., Hoek, H. W., Hoffman,
H. J., Hosgood, H. D., Hossain, M., Hotez, P. J., Hoy, D. G., Hsairi,
M., Hu, G., Huang, C., Huang, J. J., Husseini, A., Huynh, C.,
Iannarone, M. L., Iburg, K. M., Innos, K., Inoue, M., Islami, F.,
Jacobsen, K. H., Jarvis, D. L., Jassal, S. K., Jee, S. H., Jeemon, P.,
Jensen, P. N., Jha, V., Jiang, G., Jiang, Y., Jonas, J. B., Juel, K.,
Kan, H., Karch, A., Karema, C. K., Karimkhani, C., Karthikeyan,
G., Kassebaum, N. J., Kaul, A., Kawakami, N., Kazanjan, K.,
Kemp, A. H., Kengne, A. P., Keren, A., Khader, Y. S., Khalifa, S. E.
A., Khan, E. A., Khan, G., Khang, Y.-H., Kieling, C., Kim, D., Kim,
S., Kim, Y., Kinfu, Y., Kinge, J. M., Kivipelto, M., Knibbs, L. D.,
Knudsen, A. K., Kokubo, Y., Kosen, S., Krishnaswami, S., Defo, B.
K., Bicer, B. K., Kuipers, E. J., Kulkarni, C., Kulkarni, V. S.,
Kumar, G. A., Kyu, H. H., Lai, T., Lalloo, R., Lallukka, T., Lam, H.,
Lan, Q., Lansingh, V. C., Larsson, A., Lawrynowicz, A. E. B.,
Leasher, J. L., Leigh, J., Leung, R., Levitz, C. E., Li, B., Li, Y., Li, Y.,
Lim, S. S., Lind, M., Lipshultz, S. E., Liu, S., Liu, Y., Lloyd, B. K.,
Lofgren, K. T., Logroscino, G., Looker, K. J., Lortet-Tieulent, J.,
Lotufo, P. A., Lozano, R., Lucas, R. M., Lunevicius, R., Lyons, R.
A., Ma, S., Macintyre, M. F., Mackay, M. T., Majdan, M.,
Malekzadeh, R., Marcenes, W., Margolis, D. J., Margono, C.,
Marzan, M. B., Masci, J. R., Mashal, M. T., Matzopoulos, R.,
Mayosi, B. M., Mazorodze, T. T., McGill, N. W., McGrath, J. J.,
114
McKee, M., McLain, A., Meaney, P. A., Medina, C., Mehndiratta,
M. M., Mekonnen, W., Melaku, Y. A., Meltzer, M., Memish, Z. A.,
Mensah, G. A., Meretoja, A., Mhimbira, F. A., Micha, R., Miller, T.
R., Mills, E. J., Mitchell, P. B., Mock, C. N., Ibrahim, N. M.,
Mohammad, K. A., Mokdad, A. H., Mola, G. L. D., Monasta, L.,
Hernandez, J. C. M., Montico, M., Montine, T. J., Mooney, M. D.,
Moore, A. R., Moradi-Lakeh, M., Moran, A. E., Mori, R.,
Moschandreas, J., Moturi, W. N., Moyer, M. L., Mozaffarian, D.,
Msemburi, W. T., Mueller, U. O., Mukaigawara, M., Mullany, E. C.,
Murdoch, M. E., Murray, J., Murthy, K. S., Naghavi, M., Naheed,
A., Naidoo, K. S., Naldi, L., Nand, D., Nangia, V., Narayan, K. M.
V., Nejjari, C., Neupane, S. P., Newton, C. R., Ng, M., Ngalesoni, F.
N., Nguyen, G., Nisar, M. I., Nolte, S., Norheim, O. F., Norman, R.
E., Norrving, B., Nyakarahuka, L., Oh, I.-H., Ohkubo, T., Ohno, S.
L., Olusanya, B. O., Opio, J. N., Ortblad, K., Ortiz, A., Pain, A. W.,
Pandian, J. D., Panelo, C. I. A., Papachristou, C., Park, E.-K., Park,
J.-H., Patten, S. B., Patton, G. C., Paul, V. K., Pavlin, B. I., Pearce,
N., Pereira, D. M., Perez-Padilla, R., Perez-Ruiz, F., Perico, N.,
Pervaiz, A., Pesudovs, K., Peterson, C. B., Petzold, M., Phillips, M.
R., Phillips, B. K., Phillips, D. E., Piel, F. B., Plass, D., Poenaru, D.,
Polinder, S., Pope, D., Popova, S., Poulton, R. G., Pourmalek, F.,
Prabhakaran, D., Prasad, N. M., Pullan, R. L., Qato, D. M.,
Quistberg, D. A., Rafay, A., Rahimi, K., Rahman, S. U., Raju, M.,
Rana, S. M., Razavi, H., Reddy, K. S., Refaat, A., Remuzzi, G.,
Resnikoff, S., Ribeiro, A. L., Richardson, L., Richardus, J. H.,
Roberts, D. A., Rojas-Rueda, D., Ronfani, L., Roth, G. A.,
Rothenbacher, D., Rothstein, D. H., Rowley, J. T., Roy, N.,
Ruhago, G. M., Saeedi, M. Y., Saha, S., Sahraian, M. A., Sampson,
U. K. A., Sanabria, J. R., Sandar, L., Santos, I. S., Satpathy, M.,
Sawhney, M., Scarborough, P., Schneider, I. J., Schöttker, B.,
Schumacher, A. E., Schwebel, D. C., Scott, J. G., Seedat, S.,
Sepanlou, S. G., Serina, P. T., Servan-Mori, E. E., Shackelford, K.
A., Shaheen, A., Shahraz, S., Levy, T. S., Shangguan, S., She, J.,
Sheikhbahaei, S., Shi, P., Shibuya, K., Shinohara, Y., Shiri, R.,
Shishani, K., Shiue, I., Shrime, M. G., Sigfusdottir, I. D.,
Silberberg, D. H., Simard, E. P., Sindi, S., Singh, A., Singh, J. A.,
Singh, L., Skirbekk, V., Slepak, E. L., Sliwa, K., Soneji, S., Søreide,
K., Soshnikov, S., Sposato, L. A., Sreeramareddy, C. T., Stanaway,
J. D., Stathopoulou, V., Stein, D. J., Stein, M. B., Steiner, C.,
Steiner, T. J., Stevens, A., Stewart, A., Stovner, L. J.,
Stroumpoulis, K., Sunguya, B. F., Swaminathan, S., Swaroop, M.,
Sykes, B. L., Tabb, K. M., Takahashi, K., Tandon, N., Tanne, D.,
Tanner, M., Tavakkoli, M., Taylor, H. R., Ao, B. J. T., Tediosi, F.,
Temesgen, A. M., Templin, T., Ten Have, M., Tenkorang, E. Y.,
Terkawi, A. S., Thomson, B., Thorne-Lyman, A. L., Thrift, A. G.,
Thurston, G. D., Tillmann, T., Tonelli, M., Topouzis, F.,
Toyoshima, H., Traebert, J., Tran, B. X., Trillini, M., Truelsen, T.,
Tsilimbaris, M., Tuzcu, E. M., Uchendu, U. S., Ukwaja, K. N.,
Undurraga, E. A., Uzun, S. B., Van Brakel, W. H., Van De Vijver,
S., van Gool, C. H., Van Os, J., Vasankari, T. J.,
115
Conclusions
Venketasubramanian, N., Violante, F. S., Vlassov, V. V., Vollset, S.

E., Wagner, G. R., Wagner, J., Waller, S. G., Wan, X., Wang, H.,
Wang, J., Wang, L., Warouw, T. S., Weichenthal, S., Weiderpass,
E., Weintraub, R. G., Wenzhi, W., Werdecker, A., Westerman, R.,
Whiteford, H. A., Wilkinson, J. D., Williams, T. N., Wolfe, C. D.,
Wolock, T. M., Woolf, A. D., Wulf, S., Wurtz, B., Xu, G., Yan, L. L.,
Yano, Y., Ye, P., Yentür, G. K., Yip, P., Yonemoto, N., Yoon, S.-J.,
Younis, M. Z., Yu, C., Zaki, M. E., Zhao, Y., Zheng, Y., Zonies, D.,
Zou, X., Salomon, J. A., Lopez, A. D. & Vos, T. (2015). Global,
regional, and national disability-adjusted life years (DALYs) for 306 diseases
and injuries and healthy life expectancy (HALE) for 188 countries, 1990–
2013: quantifying the epidemiological transition. The Lancet 386, 2145-2191.
Mykletun, A., Bjerkeset, O., Overland, S., Prince, M., Dewey, M. &
Stewart, R. (2009). Levels of anxiety and depression as predictors of
mortality: the HUNT study. Br J Psychiatry 195, 118-25.
Nabi, H., Shipley, M. J., Vahtera, J., Hall, M., Korkeila, J.,
Marmot, M. G., Kivimaki, M. & Singh-Manoux, A. (2010a). Effects of
depressive symptoms and coronary heart disease and their interactive
associations on mortality in middle-aged adults: the Whitehall II cohort
study. Heart 96, 1645-50.
Nabi, H., Singh-Manoux, A., Ferrie, J. E., Marmot, M. G.,
Melchior, M. & Kivimaki, M. (2010b). Hostility and depressive mood:
results from the Whitehall II prospective cohort study. Psychol Med 40, 405-
13.
Nanni, V., Uher, R. & Danese, A. (2012). Childhood maltreatment
predicts unfavorable course of illness and treatment outcome in depression:
a meta-analysis. Am J Psychiatry 169, 141-51.
Nieminen, T., Martelin, T., Koskinen, S., Aro, H., Alanen, E. &
Hyyppä, M. T. (2010). Social capital as a determinant of self-rated health
and psychological well-being. Int J Public Health 55, 531-42.
Nieminen, T., Martelin, T., Koskinen, S., Simpura, J., Alanen, E.,
Härkänen, T. & Aromaa, A. (2008). Measurement and socio-
demographic variation of social capital in a large population survey. Soc
Indic Res 85, 405-423.
Nivard, M. G., Dolan, C. V., Kendler, K. S., Kan, K. J., Willemsen,
G., van Beijsterveldt, C. E., Lindauer, R. J., van Beek, J. H., Geels,
L. M., Bartels, M., Middeldorp, C. M. & Boomsma, D. I. (2015).
Stability in symptoms of anxiety and depression as a function of genotype
and environment: a longitudinal twin study from ages 3 to 63 years. Psychol
Med 45, 1039-49.
Nock, M. K., Hwang, I., Sampson, N., Kessler, R. C., Angermeyer,
M., Beautrais, A., Borges, G., Bromet, E., Bruffaerts, R., de
Girolamo, G., de Graaf, R., Florescu, S., Gureje, O., Haro, J. M.,
Hu, C., Huang, Y., Karam, E. G., Kawakami, N., Kovess, V.,
Levinson, D., Posada-Villa, J., Sagar, R., Tomov, T., Viana, M. C. &
Williams, D. R. (2009). Cross-national analysis of the associations among
mental disorders and suicidal behavior: findings from the WHO World
Mental Health Surveys. PLoS Med 6, e1000123.
116
Nordentoft, M., Mortensen, P. B. & Pedersen, C. B. (2011). Absolute
risk of suicide after first hospital contact in mental disorder. Arch Gen
Norman, R. E., Byambaa, M., De, R., Butchart, A., Scott, J. & Vos,
T. (2012). The long-term health consequences of child physical abuse,
emotional abuse, and neglect: a systematic review and meta-analysis. PLoS
Med 9, e1001349.
Noteboom, A., Beekman, A. T., Vogelzangs, N. & Penninx, B. W.
(2015). Personality and social support as predictors of first and recurrent
episodes of depression. J Affect Disord 190, 156-161.
Nyqvist, F., Forsman, A. K., Giuntoli, G. & Cattan, M. (2013). Social
capital as a resource for mental well-being in older people: a systematic
review. Aging Ment Health 17, 394-410.
O'Connor, D. W. & Parslow, R. A. (2009). Different responses to K-10
and CIDI suggest that complex structured psychiatric interviews
underestimate rates of mental disorder in old people. Psychol Med 39 (9),
1527-31.
O'Connor, D. W. & Parslow, R. A. (2010). Differences in older people's
responses to CIDI's depression screening and diagnostic questions may point
to age-related bias. J Affect Disord 125 (1-3), 361-4.
Olfson, M., Gerhard, T., Huang, C., Crystal, S. & Stroup, T. (2015).
Premature mortality among adults with schizophrenia in the united states.
JAMA Psychiatry, 1-10.
Ormel, J., Oldehinkel, A. J., Nolen, W. A. & Vollebergh, W. (2004).
Psychosocial disability before, during, and after a major depressive episode: a
3-wave population-based study of state, scar, and trait effects. Arch Gen
Ostir, G. V. & Goodwin, J. S. (2006). High anxiety is associated with an
increased risk of death in an older tri-ethnic population. J Clin Epidemiol 59,
534-40.
Paananen, R., Ristikari, T., Merikukka, M. & Gissler, M. (2013a).
Social determinants of mental health: a Finnish nationwide follow-up study
on mental disorders. Journal of Epidemiology and Community Health.
Paananen, R., Santalahti, P., Merikukka, M., Ramo, A., Wahlbeck,
K. & Gissler, M. (2013b). Socioeconomic and regional aspects in the use of
specialized psychiatric care--a Finnish nationwide follow-up study. Eur J
Public Health 23, 372-7.
Palmaro, A., Dupouy, J. & Lapeyre-Mestre, M. (2015).
Benzodiazepines and risk of death: Results from two large cohort studies in
France and UK. Eur Neuropsychopharmacol 25, 1566-77.
Papakostas, G. I. (2014). Cognitive symptoms in patients with major
depressive disorder and their implications for clinical practice. J Clin
Papakostas, G. I. & Culpepper, L. (2015). Understanding and managing
Cognition in the Depressed Patient. J Clin Psychiatry 76, 418-25.
Park, A. L., Fuhrer, R. & Quesnel-Vallee, A. (2013). Parents' education
and the risk of major depression in early adulthood. Soc Psychiatry
Parker, G. (1987). Are the lifetime prevalence estimates in the ECA study
accurate? Psychol Med 17, 275-282.
117
Conclusions
Pasco, J. A., Williams, L. J., Jacka, F. N., Ng, F., Henry, M. J.,
Nicholson, G. C., Kotowicz, M. A. & Berk, M. (2008). Tobacco
smoking as a risk factor for major depressive disorder: population-based
study. Br J Psychiatry 193, 322-6.
Patten, S. B. (2001). Long-term medical conditions and major depression
in a Canadian population study at waves 1 and 2. J Affect Disord 63, 35-41.
Patten, S. B. (2003). Recall bias and major depression lifetime prevalence.
Soc Psychiatry Psychiatr Epidemiol 38, 290-6.
Patten, S. B. (2005). An analysis of data from two general health surveys
found that increased incidence and duration contributed to elevated
prevalence of major depression in persons with chronic medical conditions. J
Clin Epidemiol 58, 184-9.
Patten, S. B., Wang, J. L., Williams, J. V., Lavorato, D. H., Khaled,
S. M. & Bulloch, A. G. (2010). Predictors of the longitudinal course of
major depression in a Canadian population sample. Can J Psychiatry 55,
669-76.
Patten, S. B., Williams, J. V., Lavorato, D., Li Wang, J., Khaled, S.
& Bulloch, A. G. (2011). Mortality associated with major depression in a
Canadian community cohort. Can J Psychiatry 56, 658-66.
Patten, S. B., Williams, J. V., Lavorato, D. H., Bulloch, A. G.,
D'Arcy, C. & Streiner, D. L. (2012). Recall of recent and more remote
depressive episodes in a prospective cohort study. Soc Psychiatry Psychiatr
Patten, S. B., Williams, J. V., Lavorato, D. H., Fiest, K. M., Bulloch,
A. G. & Wang, J. (2015a). The prevalence of major depression is not
changing. Can J Psychiatry 60, 31-4.
Patten, S. B., Williams, J. V., Lavorato, D. H., Wang, J. L.,
McDonald, K. & Bulloch, A. G. (2015b). Descriptive epidemiology of
major depressive disorder in Canada in 2012. Can J Psychiatry 60, 23-30.
Patten, S. B., Williams, J. V. A., Lavorato, D. H., Wang, J. L.,
Bulloch, A. G. M. & Sajobi, T. (2016). The association between major
depression prevalence and sex becomes weaker with age. Social Psychiatry
and Psychiatric Epidemiology, 1-8.
Pedersen, C. B., Mors, O., Bertelsen, A., Waltoft, B. L., Agerbo, E.,
McGrath, J. J., Mortensen, P. B. & Eaton, W. W. (2014). A
comprehensive nationwide study of the incidence rate and lifetime risk for
treated mental disorders. JAMA Psychiatry 71, 573-81.
Penninx, B. W., Geerlings, S. W., Deeg, D. J., van Eijk, J. T., van
Tilburg, W. & Beekman, A. T. (1999). Minor and major depression and
the risk of death in older persons. Arch Gen Psychiatry 56, 889-95.
Penninx, B. W., Nolen, W. A., Lamers, F., Zitman, F., Smit, F.,
Spinhoven, P., Cuijpers, P., de Jong, P., van Marwijk, H., van der
Meer, K., Verhaak, P., Laurant, M., de Graaf, R., Hoogendijk, W.
J., van der Wee, N., Ormel, J., van Dyck, R. & Beekman, A. T.
(2011). Two-year course of depressive and anxiety disorders: results from the
Netherlands Study of Depression and Anxiety (NESDA). J Affect Disord 133,
76-85.
Perälä, J., Kuoppasalmi, K., Pirkola, S., Härkänen, T., Saarni, S.,
Tuulio-Henriksson, A., Viertiö, S., Latvala, A., Koskinen, S.,
118
Lönnqvist, J. & Suvisaari, J. (2010). Alcohol-induced psychotic disorder
and delirium in the general population. Br J Psychiatry 197, 200-6.
Perälä, J., Suvisaari, J., Saarni, S. I., Kuoppasalmi, K., Isometsä,
E., Pirkola, S., Partonen, T., Tuulio-Henriksson, A., Hintikka, J.,
Kieseppä, T., Härkänen, T., Koskinen, S. & Lönnqvist, J. (2007).
Lifetime prevalence of psychotic and bipolar I disorders in a general
population. Arch Gen Psychiatry 64, 19-28.
Perry, D. C., Sturm, V. E., Peterson, M. J., Pieper, C. F., Bullock,
T., Boeve, B. F., Miller, B. L., Guskiewicz, K. M., Berger, M. S.,
Kramer, J. H. & Welsh-Bohmer, K. A. (2016). Association of traumatic
brain injury with subsequent neurological and psychiatric disease: a meta-
analysis. J Neurosurg 124, 511-26.
Pettersson, A., Bostrom, K. B., Gustavsson, P. & Ekselius, L. (2015).
Which instruments to support diagnosis of depression have sufficient
accuracy? A systematic review. Nord J Psychiatry 69, 497-508.
Pirkola, S., Isometsä, E., Aro, H., Kestilä, L., J., H., Veijola, J.,
Kiviruusu, O. & Lönnqvist, J. (2005a). Childhood adversities as risk
factors for adult mental disorders: results from the Health 2000 study. Soc
Pirkola, S. P., Isometsä, E., Suvisaari, J., Aro, H., Joukamaa, M.,
Poikolainen, K., Koskinen, S., Aromaa, A. & Lönnqvist, J. K.
(2005b). DSM-IV mood-, anxiety- and alcohol use disorders and their
comorbidity in the Finnish general population--results from the Health 2000
Study. Soc Psychiatry Psychiatr Epidemiol 40, 1-10.
Plant, D. T., Pariante, C. M., Sharp, D. & Pawlby, S. (2015). Maternal
depression during pregnancy and offspring depression in adulthood: role of
child maltreatment. The British Journal of Psychiatry 207, 213-220.
Prince, M., Patel, V., Saxena, S., Maj, M., Maselko, J., Phillips, M.
R. & Rahman, A. (2007). No health without mental health. Lancet 370,
859-77.
Pulkki-Råbäck, L., Ahola, K., Elovainio, M., Kivimaki, M.,
Hintsanen, M., Isometsä, E., Lönnqvist, J. & Virtanen, M. (2012).
Socio-economic position and mental disorders in a working-age Finnish
population: the health 2000 study. Eur J Public Health 22, 327-32.
Rahe, C., Unrath, M. & Berger, K. (2014). Dietary patterns and the risk
of depression in adults: a systematic review of observational studies. Eur J
Nutr 53, 997-1013.
Rai, D., Zitko, P., Jones, K., Lynch, J. & Araya, R. (2013). Country-
and individual-level socioeconomic determinants of depression: multilevel
cross-national comparison. Br J Psychiatry.
Reed, V., Gander, F., Pfister, H., Steiger, A., Sonntag, H.,
Trenkwalder, C., Sonntag, A., Hundt, W. & Wittchen, H.-U. (1998).
To what degree does the Composite International Diagnostic Interview
(CIDI) correctly identify DSM-IV disorders? Testing validity issues in a
clinical sample. Int J Methods Psychiatr Res 7, 142-155.
Rhebergen, D., Batelaan, N. M., de Graaf, R., Nolen, W. A.,
Spijker, J., Beekman, A. T. & Penninx, B. W. (2011). The 7-year course
of depression and anxiety in the general population. Acta Psychiatr Scand
123, 297-306.
119
Conclusions
Rhebergen, D., Beekman, A. T., Graaf, R., Nolen, W. A., Spijker, J.,
Hoogendijk, W. J. & Penninx, B. W. (2009). The three-year naturalistic
course of major depressive disorder, dysthymic disorder and double
depression. J Affect Disord 115, 450-9.
Riemann, D. & Voderholzer, U. (2003). Primary insomnia: a risk factor
to develop depression? J Affect Disord 76, 255-9.
Riihimäki, K. A., Vuorilehto, M. S., Melartin, T. K. & Isometsä, E.
T. (2014). Five-year outcome of major depressive disorder in primary health
care. Psychol Med 44, 1369-79.
Ritsher, J. E., Warner, V., Johnson, J. G. & Dohrenwend, B. P.
(2001). Inter-generational longitudinal study of social class and depression: a
test of social causation and social selection models. Br J Psychiatry Suppl
40, s84-90.
Rock, P. L., Roiser, J. P., Riedel, W. J. & Blackwell, A. D. (2014).
Cognitive impairment in depression: a systematic review and meta-analysis.
Psychol Med 44, 2029-40.
Roerecke, M. & Rehm, J. (2013). Alcohol use disorders and mortality: A
systematic review and meta-analysis. Addiction 108, 1562-1578.
Rubin, D. B. (1978). Multiple imputations in sample surveys – a
phenomenological Bayesian approach to nonresponse. In Survey Research
Methods Section of the American Statistical Association, pp. 20-34.
Saarni, S. I., Härkänen, T., Sintonen, H., Suvisaari, J., Koskinen,
S., Aromaa, A. & Lönnqvist, J. (2006). The impact of 29 chronic
conditions on health-related quality of life: a general population survey in
Finland using 15D and EQ-5D. Qual Life Res 15, 1403-14.
Saarni, S. I., Suvisaari, J., Sintonen, H., Pirkola, S., Koskinen, S.,
Aromaa, A. & Lönnqvist, J. (2007). Impact of psychiatric disorders on
health-related quality of life: general population survey. Br J Psychiatry 190,
326-32.
Salokangas, R. K., Vaahtera, K., Pacriev, S., Sohlman, B. &
Lehtinen, V. (2002). Gender differences in depressive symptoms. An
artefact caused by measurement instruments? J Affect Disord 68, 215-20.
Salsberry, P. J., Chipps, E. & Kennedy, C. (2005). Use of general
medical services among Medicaid patients with severe and persistent mental
illness. Psychiatr Serv 56, 458-62.
Sanchez-Villegas, A., Delgado-Rodriguez, M., Alonso, A.,
Schlatter, J., Lahortiga, F., Serra Majem, L. & Martinez-Gonzalez,
M. A. (2009). Association of the Mediterranean dietary pattern with the
incidence of depression: the Seguimiento Universidad de Navarra/University
of Navarra follow-up (SUN) cohort. Arch Gen Psychiatry 66, 1090-8.
Saz, P. & Dewey, M. E. (2001). Depression, depressive symptoms and
mortality in persons aged 65 and over living in the community: a systematic
review of the literature. Int J Geriatr Psychiatry 16, 622-30.
Schiller, C. E., Meltzer-Brody, S. & Rubinow, D. R. (2015). The role of
reproductive hormones in postpartum depression. CNS Spectr 20, 48-59.
Scott, J. G., Moore, S. E., Sly, P. D. & Norman, R. E. (2014). Bullying
in children and adolescents: A modifiable risk factor for mental illness.
Australian and New Zealand Journal of Psychiatry 48, 209-212.
Scott, K. M. & Collings, S. C. (2010). Gender and the association between
mental disorders and disability. J Affect Disord 125, 207-12.
120
Scott, K. M., Lim, C., Al-Hamzawi, A. & et al. (2015). Association of
mental disorders with subsequent chronic physical conditions: World mental
health surveys from 17 countries. JAMA Psychiatry, 1-9.
Scott, K. M., Wells, J. E., Angermeyer, M., Brugha, T. S., Bromet,
E., Demyttenaere, K., de Girolamo, G., Gureje, O., Haro, J. M., Jin,
R., Karam, A. N., Kovess, V., Lara, C., Levinson, D., Ormel, J.,
Posada-Villa, J., Sampson, N., Takeshima, T., Zhang, M. & Kessler,
R. C. (2010). Gender and the relationship between marital status and first
onset of mood, anxiety and substance use disorders. Psychol Med 40, 1495-
505.
Seedat, S., Scott, K. M., Angermeyer, M. C., Berglund, P., Bromet,
E. J., Brugha, T. S., Demyttenaere, K., de Girolamo, G., Haro, J.
M., Jin, R., Karam, E. G., Kovess-Masfety, V., Levinson, D.,
Medina Mora, M. E., Ono, Y., Ormel, J., Pennell, B. E., Posada-
Villa, J., Sampson, N. A., Williams, D. & Kessler, R. C. (2009).
Cross-national associations between gender and mental disorders in the
World Health Organization World Mental Health Surveys. Arch Gen
Skapinakis, P., Weich, S., Lewis, G., Singleton, N. & Araya, R.
(2006). Socio-economic position and common mental disorders.
Slopen, N., Williams, D. R., Fitzmaurice, G. M. & Gilman, S. E.
(2011). Sex, stressful life events, and adult onset depression and alcohol
dependence: are men and women equally vulnerable? Soc Sci Med 73, 615-
22.
Smith, N. D. & Kawachi, I. (2014). State-level social capital and suicide
mortality in the 50 U.S. states. Soc Sci Med 120C, 269-77.
Somberg, T. C. & Arora, R. R. (2008). Depression and heart disease:
therapeutic implications. Cardiology 111, 75-81.
Sourander, A., Gyllenberg, D., Brunstein Klomek, A., Sillanmäki,
L., Ilola, A. & Kumpulainen, K. (2015). ASsociation of bullying behavior
at 8 years of age and use of specialized services for psychiatric disorders by
29 years of age. JAMA Psychiatry, 1-7.
Spijker, J., Bijl, R. V., de Graaf, R. & Nolen, W. A. (2001).
Determinants of poor 1-year outcome of DSM-III-R major depression in the
general population: results of the Netherlands Mental Health Survey and
Incidence Study (NEMESIS). Acta Psychiatr Scand 103, 122-30.
Spijker, J., de Graaf, R., Bijl, R. V., Beekman, A. T., Ormel, J. &
Nolen, W. A. (2002). Duration of major depressive episodes in the general
population: results from The Netherlands Mental Health Survey and
Incidence Study (NEMESIS). Br J Psychiatry 181, 208-13.
Spinhoven, P., Elzinga, B. M., van Hemert, A. M., de Rooij, M. &
Penninx, B. W. Childhood maltreatment, maladaptive personality types
and level and course of psychological distress: A six-year longitudinal study.
Journal of Affective Disorders.
Steel, Z., Marnane, C., Iranpour, C., Chey, T., Jackson, J. W.,
Patel, V. & Silove, D. (2014). The global prevalence of common mental
disorders: a systematic review and meta-analysis 1980–2013. International
Journal of Epidemiology 43, 476-493.
Stegenga, B. T., Geerlings, M. I., Torres-Gonzalez, F., Xavier, M.,
Svab, I., Penninx, B., Nazareth, I. & King, M. (2013). Risk factors for
121
Conclusions
onset of multiple or long major depressive episodes versus single and short
episodes. Soc Psychiatry Psychiatr Epidemiol 48, 1067-75.
Stegenga, B. T., Kamphuis, M. H., King, M., Nazareth, I. &
Geerlings, M. I. (2012). The natural course and outcome of major
depressive disorder in primary care: the PREDICT-NL study. Soc Psychiatry
Stuckler, D., Basu, S., Suhrcke, M., Coutts, A. & McKee, M. (2009).
The public health effect of economic crises and alternative policy responses
in Europe: an empirical analysis. Lancet 374, 315-23.
Suetani, S., Whiteford, H. A. & McGrath, J. J. (2015). An urgent call to
address the deadly consequences of serious mental disorders. JAMA
Psychiatry, 1-2.
Sullivan, P. F., Neale, M. C. & Kendler, K. S. (2000). Genetic
epidemiology of major depression: review and meta-analysis. Am J
Sundell, K. A., Gissler, M., Petzold, M. & Waern, M. (2011).
Antidepressant utilization patterns and mortality in Swedish men and
women aged 20-34 years. Eur J Clin Pharmacol 67, 169-78.
Susukida, R., Mojtabai, R. & Mendelson, T. (2015). Sex differences in
help seeking for mood and anxiety disorders in the National Comorbidity
Survey-Replication. Depress Anxiety.
Suvisaari, J., Aalto-Setälä, T., Tuulio-Henriksson, A., Härkänen,
T., Saarni, S. I., Perälä, J., Schreck, M., Castaneda, A. E., Hintikka,
J., Kestilä, L., Lähteenmaki, S., Latvala, A., Koskinen, S.,
Marttunen, M., Aro, H. & Lönnqvist, J. (2009). Mental disorders in
young adulthood. Psychol Med 39 (2), 287-99.
Takayanagi, Y., Spira, A. P., Roth, K. B., Gallo, J. J., Eaton, W. W.
& Mojtabai, R. (2014). Accuracy of reports of lifetime mental and physical
disorders: results from the Baltimore Epidemiological Catchment Area study.
JAMA Psychiatry 71, 273-80.
Talarowska, M., Berk, M., Maes, M. & Galecki, P. (2016).
Autobiographical memory dysfunctions in depressive disorders. Psychiatry
Clin Neurosci 70, 100-8.
Tanskanen, A., Hintikka, J., Honkalampi, K., Haatainen, K.,
Koivumaa-honkanen, H. & Viinamäki, H. (2004). Impact of multiple
traumatic experiences on the persistence of depressive symptoms – a
population-based study. Nordic Journal of Psychiatry 58, 459-464.
Therneau, T., Atkinson, B. & Ripley, B. (2015). rpart: Recursive
Partitioning and Regression Trees. In R package version 4.1-9.
Thornicroft, G. (2011). Physical health disparities and mental illness: the
scandal of premature mortality. Br J Psychiatry 199, 441-2.
Thornicroft, G. (2013). Premature death among people with mental
illness. Bmj 346, f2969.
Tiihonen, J., Lönnqvist, J., Wahlbeck, K., Klaukka, T., Tanskanen,
A. & Haukka, J. (2006). Antidepressants and the risk of suicide, attempted
suicide, and overall mortality in a nationwide cohort. Arch Gen Psychiatry
63, 1358-67.
Toffol, E., Heikinheimo, O. & Partonen, T. (2015). Hormone therapy
and mood in perimenopausal and postmenopausal women: a narrative
review. Menopause 22, 564-78.
122
Torikka, A., Kaltiala-Heino, R., Rimpela, A., Marttunen , M.,
Luukkaala, T. & Rimpela, M. (2014). Self-reported depression is
increasing among socio-economically disadvantaged adolescents - repeated
cross-sectional surveys from Finland from 2000 to 2011. BMC Public Health
14.
Trivedi, M. H. & Greer, T. L. (2014). Cognitive dysfunction in unipolar
depression: implications for treatment. J Affect Disord 152-154, 19-27.
Tsai, A. C., Lucas, M. & Kawachi, I. (2015). Association Between Social
Integration and Suicide Among Women in the United States. JAMA
Psychiatry.
Uher, R., Payne, J. L., Pavlova, B. & Perlis, R. H. (2014). Major
depressive disorder in DSM-5: implications for clinical practice and research
of changes from DSM-IV. Depress Anxiety 31, 459-71.
van Borkulo, C., Boschloo, L., Borsboom, D., Penninx, B. H.,
Waldorp, L. J. & Schoevers, R. A. (2015). ASsociation of symptom
network structure with the course of longitudinal depression. JAMA
Psychiatry, 1219-1226.
van Buuren, S. & Groothuis-Oudshoorn, K. (2011). mice: Multivariate
Imputation by Chained Equations in R. Journal of Statistical Software 45, 1-
67.
van Hout, H. P., Beekman, A. T., de Beurs, E., Comijs, H., van
Marwijk, H., de Haan, M., van Tilburg, W. & Deeg, D. J. (2004).
Anxiety and the risk of death in older men and women. Br J Psychiatry 185,
399-404.
van Loo, H. M., Aggen, S. H., Gardner, C. O. & Kendler, K. S.
(2015a). Multiple risk factors predict recurrence of major depressive disorder
in women. J Affect Disord 180, 52-61.
van Loo, H. M., Schoevers, R. A., Kendler, K. S., de Jonge, P. &
Romeijn, J. W. (2015b). Psychiatric comorbidity does not only depend on
diagnostic thresholds: An illustration with major depressive disorder and
generalized anxiety disorder. Depress Anxiety.
van Zoonen, K., Buntrock, C., Ebert, D. D., Smit, F., Reynolds, C.
F., 3rd, Beekman, A. T. & Cuijpers, P. (2014). Preventing the onset of
major depressive disorder: a meta-analytic review of psychological
interventions. Int J Epidemiol 43, 318-29.
Veijola, J., Puukka, P., Lehtinen, V., Moring, J., Lindholm, T. &
Väisänen, E. (1998). Sex differences in the association between childhood
experiences and adult depression. Psychol Med 28, 21-7.
Ventevogel, P., Jordans, M., Reis, R. & de Jong, J. (2013). Madness
or sadness? Local concepts of mental illness in four conflict-affected African
communities. Confl Health 7, 3.
Viinamäki, H., Haatainen, K., Honkalampi, K., Tanskanen, A.,
Koivumaa-Honkanen, H., Antikainen, R., Valkonen-Korhonen, M.
& Hintikka, J. (2006a). Which factors are important predictors of non-
recovery from major depression? A 2-year prospective observational study.
Nordic Journal of Psychiatry 60, 410-416.
Viinamäki, H., Tanskanen, A., Honkalampi, K., Koivumaa-
Honkanen, H., Antikainen, R., Haatainen, K. & Hintikka, J.
(2006b). Recovery from depression: a two-year follow-up study of general
population subjects. Int J Soc Psychiatry 52, 19-28.
123
Conclusions
Vogelzangs, N., Duivis, H. E., Beekman, A. T., Kluft, C.,

Neuteboom, J., Hoogendijk, W., Smit, J. H., de Jonge, P. &
Penninx, B. W. (2012). Association of depressive disorders, depression
characteristics and antidepressant medication with inflammation. Transl
Psychiatry 2, e79.
Vollrath, M. & Angst, J. (1989). Outcome of panic and depression in a
seven-year follow-up: results of the Zurich study. Acta Psychiatr Scand 80,
591-596.
Wahlbeck, K. & McDaid, D. (2012). Actions to alleviate the mental health
impact of the economic crisis. World Psychiatry 11, 139-45.
Wahlbeck, K., Westman, J., Nordentoft, M., Gissler, M. &
Laursen, T. (2011). Outcomes of Nordic mental health systems: life
expectancy of patients with mental disorders. Br J Psychiatry 199, 453-458.
Walker, E., McGee, R. E. & Druss, B. G. (2015). Mortality in mental
disorders and global disease burden implications: A systematic review and
meta-analysis. JAMA Psychiatry 72, 334-341.
Wang, J., Williams, J., Lavorato, D., Schmitz, N., Dewa, C. &
Patten, S. B. (2010a). The incidence of major depression in Canada: the
National Population Health Survey. J Affect Disord 123, 158-63.
Wang, J. L., Patten, S., Sareen, J., Bolton, J., Schmitz, N. &
MacQueen, G. (2014). Development and validation of a prediction algorith
for use by health professionals in prediction of recurrence of major
depression. Depression and Anxiety 31, 451-457.
Wang, J. L., Schmitz, N. & Dewa, C. S. (2010b). Socioeconomic status
and the risk of major depression: the Canadian National Population Health
Survey. J Epidemiol Community Health 64, 447-52.
Waraich, P., Goldner, E. M., Somers, J. M. & Hsu, L. (2004).
Prevalence and incidence studies of mood disorders: a systematic review of
the literature. Can J Psychiatry 49, 124-38.
Waszczuk, M. A., Zavos, H. M., Gregory, A. M. & Eley, T. C. (2014).
The phenotypic and genetic structure of depression and anxiety disorder
symptoms in childhood, adolescence, and young adulthood. JAMA
Weich, S. & Lewis, G. (1998). Poverty, unemployment, and common
mental disorders: population based cohort study.
Wesselhoeft, R., Pedersen, C. B., Mortensen, P. B., Mors, O. &
Bilenberg, N. (2015). Gender-age interaction in incidence rates of
childhood emotional disorders. Psychol Med 45, 829-39.
White, J., Zaninotto, P., Walters, K., Kivimaki, M., Demakakos, P.,
Shankar, A., Kumari, M., Gallacher, J. & Batty, G. D. (2015). Severity
of depressive symptoms as a predictor of mortality: the English longitudinal
study of ageing. Psychol Med 45, 2771-9.
WHO (1993). The ICD-10 Classification of Mental and Behavioural
Disorders Diagnostic criteria for research. (ed. W. H. Organization): Geneva.
Whooley, M. A. & Wong, J. M. (2013). Depression and cardiovascular
disorders. Annu Rev Clin Psychol 9, 327-54.
Wiles, N. J., Haase, A. M., Gallacher, J., Lawlor, D. A. & Lewis, G.
(2007). Physical Activity and Common Mental Disorder: Results from the
Caerphilly Study. American Journal of Epidemiology 165, 946-954.
124
Wittchen, H.-U. & Pfister, H. (1997). DIA-X Interviews; Manual für
Screening-Verfahren und Interview; Interviewheft
Längsschnittuntersuchung (DIA-X-Lifetime); Ergänzungsheft (DIA-X-
Lifetime); Interviewheft Querschnittuntersuchung (DIA-X-12 Monate);
Ergänzungsheft (DIA-X-12 Monate); PC-Programm zur Durchführung des
Interviews (Längs- und Querschnittuntersuchung);
Auswertungsprogramm. . Swets & Zeitlinger: Frankfurt, Germany.
Wittchen, H. U. & Jacobi, F. (2005). Size and burden of mental disorders
in Europe--a critical review and appraisal of 27 studies. Eur
Neuropsychopharmacol 15, 357-76.
Wittchen, H. U., Jacobi, F., Rehm, J., Gustavsson, A., Svensson,
M., Jonsson, B., Olesen, J., Allgulander, C., Alonso, J., Faravelli,
C., Fratiglioni, L., Jennum, P., Lieb, R., Maercker, A., van Os, J.,
Preisig, M., Salvador-Carulla, L., Simon, R. & Steinhausen, H. C.
(2011). The size and burden of mental disorders and other disorders of the
brain in Europe 2010. Eur Neuropsychopharmacol 21, 655-79.
Wittchen, H. U., Lachner, G., Wunderlich, U. & Pfister, H. (1998).
Test-retest reliability of the computerized DSM-IV version of the Munich-
Composite International Diagnostic Interview (M-CIDI). Soc Psychiatry
Ylijoki-Sorensen, S., Boldsen, J. L., Lalu, K., Sajantila, A.,
Baandrup, U., Boel, L. W., Ehlers, L. H. & Boggild, H. (2014). Cost-
consequence analysis of cause of death investigation in Finland and in
Denmark. Forensic Sci Int 245c, 133-142.
Zahn-Waxler, C., Klimes-Dougan, B. & Slattery, M. J. (2000).
Internalizing problems of childhood and adolescence: prospects, pitfalls, and
progress in understanding the development of anxiety and depression. Dev
Psychopathol 12, 443-66.
Zeber, J. E., Copeland, L. A., McCarthy, J. F., Bauer, M. S. &
Kilbourne, A. M. (2009). Perceived Access to General Medical and
Psychiatric Care Among Veterans With Bipolar Disorder. American Journal
of Public Health 99, 720-727.
Ziegelstein, R. C. (2015). Depression and the risk of mortality: in search of
calipers. J Psychosom Res 78, 301-3.
Zivin, K., Yosef, M., Miller, E. M., Valenstein, M., Duffy, S., Kales,
H. C., Vijan, S. & Kim, H. M. (2015). Associations between depression
and all-cause and cause-specific risk of death: a retrospective cohort study in
the Veterans Health Administration. J Psychosom Res 78, 324-31.
Zukewich, N. & Norris, D. (2005). National Experiences and
International Harmonization in Social Capital Measurement: A Beginning.
Zvolensky, M. J., Bakhshaie, J., Sheffer, C., Perez, A. & Goodwin,
R. D. (2015). Major depressive disorder and smoking relapse among adults
in the United States: a 10-year, prospective investigation. Psychiatry Res
226, 73-7.
125
Recent Publications in this Series
NIINA MARKKULA
5/2016 Liisa Korkalo
Hidden Hunger in Adolescent Mozambican Girls: Dietary Assessment, Micronutrient Status,
and Associations between Dietary Diversity and Selected Biomarkers
6/2016 Teija Ojala
Lactobacillus crispatus and Propionibacterium freudenreichii: A Genomic and Transcriptomic
View dissertationes scholae doctoralis ad sanitatem investigandam
7/2016 César Araujo
universitatis helsinkiensis 24/2016
Prostatic Acid Phosphatase as a Regulator of Endo/Exocytosis and Lysosomal Degradation
8/2016 Jens Verbeeren
Prevalence, Predictors and Prognosis of Depressive Disorders in the General Population

Regulation of the Minor Spliceosome through Alternative Splicing and Nuclear Retention of the
U11/U12-65K mRNA
9/2016 Xiang Zhao
HMGB1 (Amphoterin) and AMIGO1 in Brain Development NIINA MARKKULA
10/2016 Tarja Pääkkönen (Jokinen)
Benign Familial Juvenile Epilepsy in Lagotto Romagnolo Dogs
11/2016 Nora Hiivala Prevalence, Predictors and Prognosis of Depressive
Patient Safety Incidents, Their Contributing and Mitigating Factors in Dentistry
12/2016 Juho Heinonen
Disorders in the General Population
Intravenous Lipid Emulsion for Treatment of Local Anaesthetic and Tricyclic Antidepressant
Toxicity
13/2016 Riikka Jokinen
Genetic Studies of Tissue-Specific Mitochondrial DNA Segregation in Mammals
14/2016 Sanna Mäkelä
Activation of Innate Immune Responses by Toll-like Receptors and Influenza Viruses
15/2016 Mari Hirvinen
Immunological Boosting and Personalization of Oncolytic Virotherapies for Cancer Treatment
16/2016 Sofia Montalvão
Screening of Marine Natural Products and Their Synthetic Derivatives for Antimicrobial and
Antiproliferative Properties
17/2016 Mpindi John Patrick
Bioinformatic Tools for Analysis, Mining and Modelling Large-Scale Gene Expression and Drug
Testing Datasets
18/2016 Hilla Sumanen
Work Disability among Young Employees Changes over Time and Socioeconomic Differences
19/2016 Oyediran Olulana Akinrinade
Bioinformatic and Genomic Approaches to Study Cardiovascular Diseases
20/2016 Prasanna Sakha
Development of Microfluidic Applications to Study the Role of Kainate Receptors in
Synaptogenesis
21/2016 Neha Shrestha
Mesoporous Silicon Systems for Oral Protein/Peptide-Based Diabetes MENTAL HEALTH UNIT
Mellitus Therapy
22/2016 Tanja Holopainen
NATIONAL INSTITUTE FOR HEALTH AND WELFARE AND
Targeting Endothelial Tyrosine Kinase Pathways in Tumor Growth and DEPARTMENT OF PSYCHIATRY
Metastasis FACULTY OF MEDICINE
23/2016 Jussi Leppilahti DOCTORAL PROGRAMME IN CLINICAL RESEARCH
Variability of Gingival Crevicular Fluid Matrix Metalloproteinase -8 Levels
in Respect to Point-of-Care Diagnostics in Periodontal Diseases UNIVERSITY OF HELSINKI
24/2016
Helsinki 2016 ISSN 2342-3161 ISBN 978-951-51-2070-0

Prevalence, Predictors and Prognosis of Depressive

Uploaded by

Copyright:

Available Formats

Prevalence, Predictors and Prognosis of Depressive

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Prevalence, Predictors and Prognosis of Depressive

Uploaded by

Copyright:

Available Formats

Recent Publications in this Series

Prevalence, Predictors and Prognosis of Depressive Disorders in the General Population

Helsinki 2016 ISSN 2342-3161 ISBN 978-951-51-2070-0

Mental Health Unit

PREVALENCE, PREDICTORS AND

To be presented, with the permission of the Faculty of Medicine of

Docent Samuli Saarni, MD, PhD

National Institute for Health and Welfare

Turku University Hospital and University of Turku,

Reviewed by Docent Soili Lehto, MD, PhD

Professor Jouko Miettunen, PhD

Opponent Professor emeritus Raimo K. Salokangas, MD, PhD

Cover image by Marcelino Martínez González

Dissertationes Scholae Doctoralis Ad Sanitatem Investigandam Universitatis

Depressive disorders are a major public health concern worldwide due to

This study confirms that depression is an increasing public health concern in

Masennushäiriöt ovat merkittävä kansanterveydellinen ongelma

Tulokset osoittavat, että yhdellä kymmenestä suomalaisesta aikuisesta oli

Tämä tutkimus vahvistaa, että masennushäiriöt ovat kasvava

In addition to the National Graduate School of Clinical Investigation, this

It has been a privilege to learn from my supervisors. I am grateful to Docent

My deepest gratitude and admiration goes to Research Professor Jaana

I thank my thesis committee members, Mika Gissler and Kristian Wahlbeck,

I thank my employers in Chile, Dr. Mauricio Gómez and Professor Báltica

I am so grateful to my friends, for their academic and, more importantly,

I am grateful to my parents, Ritva and Pekka, and my sister Emilia for

My husband Sebastián, I thank for his constructive, although sometimes

Contents ............................................................................................................... 10

List of original publications ..................................................................................15

Abbreviations ....................................................................................................... 16

1 Introduction .................................................................................................... 18

2 Review of the literature .................................................................................. 20

2.1 Definition of depressive disorders .......................................................... 20

2.1.1 ICD-10............................................................................................... 20

2.1.2 DSM-IV ............................................................................................ 21

2.1.3 DSM-5 .............................................................................................. 22

2.1.4 Main differences between the diagnostic systems .......................... 23

2.2 Prevalence of depressive disorders......................................................... 24

2.2.1 Measurement of depressive disorders in population studies ......... 24

2.2.1.1 Methods of assessing depressive disorders in

2.2.1.2 Methodological challenges in prevalence studies ............... 26

2.2.2 Global prevalence of major depressive disorder............................. 27

2.2.3 Prevalence of depressive disorders in Finland ............................... 31

2.2.4 Prevalence of dysthymia .................................................................. 31

2.2.5 The global burden of depressive disorders ..................................... 32

2.3 Prevalence trends of depressive disorders over time ............................. 33

2.4 Risk factors for depressive disorders...................................................... 33

2.4.1 Sociodemographic factors................................................................ 34

2.4.1.2 Age ........................................................................................ 35

2.4.1.3 Socioeconomic position ....................................................... 35

2.4.1.4 Marital status ....................................................................... 37

2.4.2 Family history and early life experiences ....................................... 37

2.4.2.1 Family history ...................................................................... 37

2.4.2.2 Early life experiences........................................................... 37

2.4.3 Social capital .................................................................................... 38

2.4.4 Chronic somatic conditions and health behaviours ....................... 39

2.4.5 Other psychiatric disorders and personality traits ......................... 40

2.5 Prognosis of depressive disorders .......................................................... 42

2.5.1 Key concepts of prognosis ............................................................... 42