EBM, Design Study, Quantifying Risk

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 71

CAUSALITY, DESIGN STUDY,

QUANTIFYING OF RISK AND EBM

Dr Putu Moda Arsana SpPD – KEMD

Endocrinology & Metabolic section


Department of Internal Medicine
Brawijaya University
ASSESSMENT OF CAUSATION

• What is causation ?
• What study can show causation ?
• A direct test of causation
• Method of counting event
CAUSAL RELATIONSHIP IN
MEDICINE AND HEALTH CARE

• Physical sciences :
Metal :  temperature  length 
• Health : ( simple )
Infection with m. tuberculosis

Clinical tuberculosis ?
CAUSAL FACTOR

Infection M Tuberculosis

Poor nutrition

Age
Clinical tuberculosis

Genetic factors ?

Environmental condition
TYPES OF CAUSAL RELATIONSHIPS
• Necessary :
The outcome occurs only if the causal factor has operated
• Sufficient :
The operation of the causal factor always results in the
outcome
• Both :
The causal factor and the outcome have a fix relationship,
neither occurs without the other
• Neither :
The operation of the causal factor increases the frequency of
the outcome, but the outcome does not always result, and the
outcome can occur without the operation of the causal factor.
DEFINITION OF CAUSE

A factor is a cause of an event, if its


operation increases the frequency of
the event
NOTE

• Most situations in health and disease do


not fulfill the criteria of necessary or
sufficient
• Most of them were neither causal
factors
NOTE
If the time relationship is not clear,
and the concepts of necessary and
sufficient cause do not hold, we need
a quantitative assessment of the
relationship, base on observations,
not on one individual but on a number
of individual

( quantitative causation )
A DIRECT TEST OF CAUSATION
( RANDOMIZED CLINICAL TRIAL )

Take a group of
subject

Divide them into


two similar group

Apply the putative causal


factor to one group

Assess the frequency of the


outcome in both group

See if the frequency is higher in the group


exposed to the putative causal factor
METHODS OF COUNTING EVENTS
• Because there is limitation ( ethical and logistical ) of prospective
randomized study to prove “ causation “ in health care, we use
other study such as observational study to assess evidence to
support a causal relationship.
• So, causation needs to be defined in term of probabilities.
• Prevalence : the frequency of a characteristic, or the proportion
of a group which has the characteristic at one point in time.
• Incidence rate : the frequency of incidences; events such as
deaths or new diagnoses of disease, over a defined time period.
• The population at risk : the number of human that theoritically and
statistically more possible to get the disease ( e.g ; breast cancer
is more common in women than men ).
• Cumulative incidence : proportion of a group of subjects which
experiences an events from the start to the end of a specified
time period.
STUDY DESIGN WHICH CAN DEMONSTRATE
AND TEST CAUSATION
Implication of the definition of causal factor :
1. People who are affected by the causal agent will have a higher
frequency of the defined outcome
2. Individuals with the defined outcome will have a greater
frequency of past exposure to the causal agent
Comparative study

Study design for the 1 st statement : “ Cohort study “


Compare a group of people exposed to the putative causative
factor with a group not exposed
Study design for the 2 nd statement : “ a case control study “
Compare a group of people who have already experienced the
outcome with a group of people without the outcome
A CLASSIFICATION OF COMPARATIVE STUDY

Cohort means a group of people with some characteristic.


• Observational study : the researchers observe the natural events,
they do not involved or do intervention.
• Intervention study : the investigators control the assignment of
individuals to the intervention.
Case – control study
• Group of individuals are defined in term of whether they have or
have not already experienced the outcome under consideration,
and the exposure is then measured.
Surveys
• From the sampling perspective, surveys are carried out on
subjects chosen as all members or a representative sample of a
population but the sampling is not based on any defined exposure
or outcome
CLASSIFICATION BY TIME RELATIONSHIP

Present time only


( cross sectional study )

Past and present


( retrospective study )

Present and future


( prospective study )

Past Present Future


Information on past Time period during which Subjects followed in time
events found and subjects are entered in the and events recorded as
recorded study and data collection they occur, or as subjects
started are rechecked
Case – control studies :
• Cross-sectional or retrospective
Cohort studies :
• Can be prospective, cross-sectional or retrospective
Cross-sectional studies :
• All information is collected at one point in time
• Can be made if the putative causal factors and the
outcome state are stable. If the exposure or the
outcome change overtime, this design is weak.
MORE COMPLEX DESIGN

Case-control studies within a cohort :


• “ nested “ case-control studies
• The case-control study was done within a large prospective
cohort study to get a deeper information about causation

Cohort and case-control analyses within a


cross-sectional survey
• A survey may be used to identify sub-group with different
exposures, who are then followed giving a prospective
cohort study; or sub-group can be defined in terms of an
outcome, and more detailed case-control comparisons can
then be made
Study design : sampling, time relationship and
analysis
Sampling system Time relationship Type of analysis

Cohort : Prospective, retrospective Cohort


Select on exposure or cross-sectional

Prospective
Intervention : Cohort
Assign exposure
Retrospective or cross-
Case-control : sectional Case-control
Select on outcome

Survey : Cross-sectional or Survey, plus cohort or


retrospective case-control analyses
Total or sample of a comparing subgroups
defined population
INTERVENTION TRIALS
Purpose :
• To measure the effect of the intervention on certain
predetermined outcomes.
Benefit :
• These trials are limited ( in humans ) to the evaluation of
interventions which are likely to be beneficial
• Usually can be used to asses the benefit of treatment
Optimal method :
• Randomized clinical trial ( RCT )
Sampling method :
• Randomization ( single or double blind )
Disadvantages :
• Organization, time, cost, resources and ethics.
Non randomized intervention trials :
COHORT AND INTERVENTION TRIALS

Relative risk or risk ratio :


• The ratio of the rate of disease among those exposed to the
rate in those not exposed

Odds ratio :
• The ratio of number of defined outcome to the undefined
outcome in the exposed group to the number of defined
outcome to the number of undefined outcome in the
unexposed group

Risk difference ( attributable risk = AR ) :


• The difference between the two rates
EXAMPLES

Result of the cohort study in pregnan women


Exposure Infant Infant not Total Prevalence of
malformed malformed malformation
(%)
Valproic acid 6 387 393 1.53

No valproic acid 7 1932 1939 0.36

Whole population 13 2319 2332 0.56

RR ( relative risk ) = 1.53/0.36 = 4.25


OR ( Odds ratio ) = 6/387 : 7/1932 = ( 6x1932 ) / ( 7 x 387 ) = 4.28
AR ( Attributable Risk or risk difference ) = 1.53 – 0.36 = 1.17 %
THE USES OF RISK DIFFERENCE AND
RELATIVE RISK

Risk difference :
• Describe the absolute quantity of the outcome which is
associated with the exposure
• Useful in considering the practical implications of studies
• Describe more causal association then relative risk

Relative risk :
• Useful in considering the estimation of incidence or mortality
associated with a particular exposure in a different population,
because empiric finding showed those incidence or mortality
were relative constant
• More valuable in evaluating whether a particular relationship
is or is not likely to be causal
• Easier to predict the effect of non causal factors on the
observed relative risk than on the observed attributable risk
PREVENTIVE FACTORS AND IMPLICATION TO
INTERVENTION STUDIES ; NUMBER NEEDED
TO TREAT ( NNT )
If the factors under consideration is preventive, the rate of outcome in
the exposed group will be less then in the unexposed group, and therefore
the relative risk will be less then one, and the risk difference will be
negative.

Treatment Death Survivors Total Death rate


per 1000
Captopril 2088 26940 29028 71.9

Placebo 2231 26791 29022 76.9

Relative Risk = 71.9 / 76.9 = 0.94


Risk Difference per 1000 = 71.9 – 76.9 = - 5
Death averted per 1000 treated = 76.9 – 71.9 = 5
Number treated to prevent one death ( NNT ) = 1000 / 5 = 200
THE RESULTS OBTAINED FROM
STUDIES OF CAUSATION
When you can measure what you are speaking about,
and express it in numbers, you know something
about it;
but when you cannot measure it, when you cannot
express it in numbers, your knowledge is of a
meagre and unsatisfactory kind :
it may be the beginning of knowledge, but you have
scarcely, in your thought, advanced to the stage
of science

William Thomson, Lord Kelvin : Popular lecturer and


addresses ; 1891-94
SELECTION OF SUBJECTS FOR STUDY

• Target population
• Source population
• Eligible polpulation
• Study participants
TARGET POPULATION ( S )

The population ( s ) to which the result can be


applied
SOURCE POPULATION

• The population ( s ) defined in general terms and


enumerated if possible, from which eligible
subjects are drawn
ELIGIBLE POPULATION

• The population (s) of subjects eligible for


inclusion in the study ; should be defined
precisely and counted
STUDY PARTICIPANTS

• Those individuals who contribute data to the study


; the results apply directly only to this subjects
LEVELS OF SUBJECT SELECTION ( 1 )

Level of selection Main exclusions

Target population

Source population
Subjects not assessed
Subjects assessed and found not eligible
Subjects not classified because of inadequate data

Eligible population
Exclusions because of death, inability to cooperate,
administrative issues, confidentiality, voluntary non-
response... ( do not enter study )
Failure to complete study requirements, missing data...... ( do
not complete study )
Study participants
LEVELS OF SUBJECT SELECTION ( 2 )

Target population (s) : the population (s) to which the results can be applied

Direction of selection of Direction of application of


subjects results

Source population (s) : the population (s) defined in general terms and enumerated if possible,
from which eligible subjects are drawn

Selection Application

Eligible population (s) : the population (s) of subjects eligible for inclusion in the study , should
be defined precisely and counted

Selection Application

Study participants : those individuals who contribute data to the study , the results apply directly
only to these subjects
CLINICAL DAILY PRACTICE

Good evening Doctor, Would


You like to help me please ?
What should I do ?
I have been suffering from
chest pain since 3 hours ago
and……………
Diagnosis Treatment
Process process

Complaint Diagnose Treatment

Clinical Process
TREATMENT
PROCESS

CLINICAL
DECISIAN MAKING
Clinical
Expertise

Best Patient
Research CDM Values &
Evidence Preferences
OLD MODEL FOR CLINICAL DECISIONS

• Unsystematic observations/clinical
experience
• Pathophysiology plus pharmacology
• Extrapolation from intermediate
outcomes
• Authority of local experts
• Practitioners and patients not “equals”

EBM Working Group. JAMA 1992;268:2420-2425


NEW MODEL FOR CLINICAL
DECISIONS
• Systematic recording of observations -
reproducible and unbiased
• Mechanism of disease - necessary but not
sufficient
• Critical literature appraisal Vs authority
• Apply rules of evidence
• Full informed participation by patients

EBM Working Group. JAMA 1992;268:2420-2425


SO, HOW SHOULD WE CARE
OUR PATIENT ?

• Clinical Expertise ?  Yes


• Patient Values and
Preferences ?  Yes
• Best Research
 Yes
Evidence ?
EVIDENCE BASE
MEDICINE ?
EVIDENCE BASED MEDICINE

• • Definition: “The conscientious, explicit and


judicious use of current best evidence in
making decisions about the care of individual
patients” Sackett 1996
• • The practice of evidence based medicine
requires the integration of individual clinical
expertise, judgment and patient choice with
the best external evidence from systematic
research
BUT…IT’S A JUNGLE OUT THERE!

New information
available daily
The public has a high
interest and demand
And… the real world is…
• Haynes (An Intern Med 1986; 309 :105) ; 800
research articles in 4 famous journals 
valid only 19%
• Reid (JAMA 1995; 274: 651 ) ; 1300 research
articles on accuracy of diagnostic tools from urine
dipstick to MRI and CT scan 
valid only 6%

• Cohrane Collaboration (1996): 16,000 studies on


tx of mild hypertension  valid only 22 studies
What is in the real world?
Publication Bias
Research 1 Research 2 Research 3 Research 4 Research 5

Good results  published

Poor results not published

AHRQ  48% medical research were not published


SO, WHAT SHOULD WE DO IF WE
WANT TO APPLY BEST EVIDENCE
TO OUR PATIENTS ?
THE STEPS IN THE EBP/CDM PROCESS:

ASSESS 1. Start with the patient -- a clinical


the patient problem or question arises from the
care of the patient
ASK 2. Construct a well built clinical
the question question derived from the case
ACQUIRE 3. Select the appropriate resource(s)
the evidence and conduct a search

APPRAISE 4. Appraise that evidence for its validity


the evidence (closeness to the truth) and
applicability (usefulness in clinical
practice)
APPLY: 5. Return to the patient -- integrate
talk with the patient that evidence with clinical expertise,
patient preferences and apply it to
practice
Self-evaluation 6. Evaluate your performance with this
patient
ASSESS THE PATIENT
Diagnosis Treatment
Process process

Complaint Diagnose Treatment

Clinical Process
Case
The patient is a 15 year old female with a history of type 1
diabetes . Over the years he has been treating with insulin. She
understands that her diabetes puts her at a high risk for
neuropathy and She frustrated that She cannot hide that
complication. Lately She read from internet that intensive
insulin therapy could help her to delay that complication but
She still doubt about the efficacy of their information. She
wants to know if this insulin procedure really works.
PATIENT PROBLEMS

1.Type-1 Diabetes Mellitus


FORMULATING PICO

• P (Problem)
• I (Intervention)
• C (Comparison)
• O (Outcome)
PICO
• Problems : 1. DM
• Intervention : Intensive Insulin therapy
• Comparison : Conventional insulin therapy
• Outcome : Neuropathy
CLINICAL QUESTION

In patients with type-1 diabetes, is intensive


insulin therapy more effective than
conventional insulin therapy in decreasing
the probability of neuropathy?
SELECT THE APPROPRIATE
RESOURCE(S) AND CONDUCT A
SEARCH

Deseases Control and Complication


Trial in Type-1 DM
(DCCT)
APPRAISE / EVALUATING THE
EVIDENCE
VALIDITY, IMPORTANCE, APPLICABILITY
(VIA)
MAIN ISSUE OF THE CRITICAL
APPRAISAL

V ( valid ) : Are the result of the study valid ?


The answer : see on the method and sample characteristic
I ( Important ) : Are the valid result of this study important ?
The answer : see on the results
A ( Applicable ) : Can you apply this valid, important
evidence about this study for your patients ?
The answer : see on the discussion
ARE THE RESULTS OF THE
THERAPEUTIC TRIAL VALID ?
• Was the assignment of patients to treatments randomized?
• Was the randomization list concealed?
• Was follow-up of patients sufficiently long and complete?
• Were all patients analyzed in the groups to which they were
randomized?
• Were patients, clinicians, and study personnel kept “blind” to
treatment?
• Were the groups treated equally, apart from the experimental
treatment?
• Were the groups similar at the start of the trial apart from the
experimental therapy?
ARE THE VALID RESULTS OF THIS
RANDOMIZED TRIAL IMPORTANT ?

• What is the magnitude of the


treatment effect?
• How precise is the estimate of
the treatment effect?
Are the valid results of this
randomized trial important ?
Parameter of “ Important “ :
• ARR ( Absolute Risk reduction ) : CER – EER
• RRR ( Relative Risk Reduction ) : ( CER – EER ) / CER
• NNT ( Number Needed to treat ) : 1 / ARR

CER : Control Event Rate


EER : Experimental Event Rate
Note : Use 2 x 2 table to account CER and EER
SAMPLE CALCULATIONS
Occurrence of diabetic Relative risk Absolute risk Number
neuropathy at 5 years among reduction reduction needed to
insulin-dependent diabetics in (RRR) (ARR) treat
the DCCT trial (NNT)
Usual insulin Intensive CER – EER CER – EER 1/ARR
regimen insulin regimen CER
control event experimental
rate (CER) event rate
(EER)
9.6% 2.8% 9.6% – 2.8% 9.6% – 2.8% 1/6.8%
9.6% =15 patients
=71% =6.8%
95% CI 4.4% to 9.2% 11 to 23
95% confidence interval (CI) on an ARR and NNT

= ARR ± 1.96 CER x (1-CER) EER x (1-EER)


+
number of control patients number of experimental patients

= 6.8% ± 1.96 0.96 x 0.904 0.028 x 0.972


+
730 711

= 6.8% ± 2.4% = 4.4% - 9.2%

YOUR CALCULATIONS
Relative risk Absolute risk Number
reduction reduction needed to
(RRR) (ARR) treat (NNT)
CER EER CER – EER CER – EER 1/ARR
CER

95% CI
Can you apply this valid, important evidence about
therapy in caring for your patient?

Do these results apply to our patient?


Is our patient so different from those in the
study that its results cannot apply?
Is the treatment feasible in our setting?
What are our patient’s potential benefits and harms from the therapy?
Method I: f Risk of the outcome in our patient,
relative to patients in the trial.
Expressed as a decimal:______
NNT/f=______/______=______
(NNT for patients like ours)
Method II: 1/(PEERRRR) Our patient’s expected event rate if they
received the control treatment (PEER)
=______
1/(PEERRRR)=1/________=______
(NNT for patients like ours)
Are our patient’s values and preferences satisfied by the regimen and its consequences?

Do we and our patient have a clear assessment


of their values and preferences?
Are they met by this regimen and its
consequences?
WHAT SHOULD WE DO?

EVIDENCE EVIDENCE
TYPE OF STUDY

Meta Analysis

Systemic Review

Randomized Controlled Trial

Cohort studies

Case Control studies

Case Series/Case Report

Animal Research
 1++ High-quality meta-analyses, systematic reviews of randomized controlled trials
(RCTs), or RCTs with a very low risk of bias
 1+ Well-conducted meta-analyses, systematic reviews of RCTs or RCTs with a low
risk of bias)
 1– Meta-analyses, systematic reviews of RCTs or RCTs with a high risk of bias *
 2++ High-quality systematic reviews of non-RCT, case–control, cohort, controlled
before-and-after study (CBA) or interrupted time series (ITS) studies
 High quality non-RCT, case–control, cohort, CBA or ITS studies with a very low risk of
confounding, bias or chance and a high probability that the relation is causal
 2+ Well-conducted non-RCT, case–control, cohort, CBA or ITS studies with a very low
risk of confounding, bias or chance and a moderate probability that the relation is causal
 2– Non-RCT, case–control, cohort, CBA or ITS studies with a high risk of confounding,
bias or chance and a significant risk that the relationship is not causal *
 3 Non-analytic studies (for example, case reports, case series)
 4 Expert opinion, formal consensus

* Studies with a level of evidence '–' should not be used as a basis for making a
recommendation.
Classification of Recommendations and
Level of Evidence

Class I Class IIa Class IIb Class III

Benefit >>> Risk Benefit >> Risk Benefit ≥ Risk Risk ≥ Benefit
Additional studies with Additional studies with No additional studies
focused objectives broad objectives needed; needed
needed Additional registry data
would be helpful Procedure/Treatment
Procedure/ Treatment IT IS REASONABLE to should NOT be
SHOULD be perform Procedure/Treatment performed/administered
performed/ procedure/administer MAY BE CONSIDERED SINCE IT IS NOT
administered treatment HELPFUL AND MAY
BE HARMFUL

Level A Multiple (3-5) population risk strata evaluated


General consistency of direction and magnitude of effect

Level B Limited (2-3) population risk strata evaluated

Level C Very limited (1-2) population risk strata evaluated


Applying Classification of
Recommendations and Level of Evidence

Class I Class IIa Class IIb Class III

Benefit >>> Risk Benefit >> Risk Benefit ≥ Risk Risk ≥ Benefit
Additional studies with Additional studies with No additional studies
focused objectives broad objectives needed
needed needed; Additional
registry data would be Procedure/Treatment
Procedure/ Treatment IT IS REASONABLE helpful should NOT be
SHOULD be to perform performed/administered
performed/ procedure/administer Procedure/Treatment SINCE IT IS NOT
administered treatment MAY BE CONSIDERED HELPFUL AND MAY BE
HARMFUL

should is reasonable may/might be considered is not recommended


is recommended can be useful/effective/ may/might be reasonable is not indicated
is indicated beneficial usefulness/effectiveness is should not
is useful/effective/ is probably recommended unknown /unclear/uncertain is not
beneficial or indicated or not well established useful/effective/beneficial
may be harmful
CHALLENGES TO USING EBM IN
DAILY PRACTICE
• Lack of time
• Lack of access to information
• Lack of evidence
• Integrating patients values in decisions making
• Organizational barriers
• Financial barriers
EBM IN DAILY PRACTICE?

ACT LOCALLY
THINK GLOBALLY

You might also like