Open Access

Volume: 40, Article ID: e2018018, 9 pages

https://doi.org/10.4178/epih.e2018018

ORIGINAL ARTICLE

Sleep quality and associated factors among the elderly

living in rural Chiang Rai, northern Thailand
Weerakorn Thichumpa1, Nopporn Howteerakul1, Nawarat Suwannapong2, Visasiri Tantrakul3
1
Department of Epidemiology, Faculty of Public Health, Mahidol University, Bangkok, Thailand; 2Department of Public Health Administration,
Faculty of Public Health, Mahidol University, Bangkok, Thailand; 3Department of Medicine, Faculty of Medicine Ramathibodi Hospital, Mahidol
University, Bangkok, Thailand

OBJECTIVES: This study aimed to characterize the prevalence of poor sleep quality and to identify associated factors among
community-dwelling elderly individuals in northern Thailand.
METHODS: A cross-sectional study was conducted among 266 randomly selected elderly people in a sub-district in rural Chi-
ang Rai Province, northern Thailand. The participants were interviewed using the Thai version of the Pittsburgh Sleep Quality
Index (PSQI).
RESULTS: Roughly 44.0% of the participants had poor sleep quality (PSQI score, >5), 9.4% used sleep medication, 27.1% had
poor family relationships, and 12.0% had mild depression. Multiple logistic regression analysis indicated that being female (odds
ratio [OR], 1.74; 95% confidence interval [CI], 1.10 to 3.02), a higher education level (OR, 3.03; 95% CI, 1.34 to 6.86 for primary
school; OR, 2.48; 95% CI, 1.31 to 5.44 for higher than primary school), mild depression (OR, 2.65; 95% CI, 1.11 to 6.36), and poor
family relationships (OR, 3.65; 95% CI, 1.98 to 6.75) were significantly associated with poor sleep quality.
CONCLUSIONS: The prevalence of poor sleep quality among the elderly was moderately high. Healthcare providers should
regularly conduct screenings for sleep quality and depression; provide sleep health education; and conduct interventions to en-
courage participating in family activities, resolving conflicts, sharing ideas, and making compromises within the family.

KEY WORDS: Sleep hygiene, Elderly, Prevalence, Depression, Family relations

INTRODUCTION often more pronounced in individuals who are living with multi-
ple chronic diseases such as cancer, stroke, and the decreasing per-
Poor sleep quality is a common problem in the general popula- formance associated with aging [6,10,11]. Poor sleep quality can
tion, often due to work-related demands and lifestyle changes [1- have profound physical effects on the elderly, including fatigability
3]. Problems with sleep increase sharply with age [4-9]. Poor sleep and an increased risk of falls. These effects threaten both mobility
quality is associated with serious negative consequences on physi- and independence [3,4,7]. Poor sleep quality and sleep depriva-
cal, mental, and social aspects of well-being [10]. These effects are tion are also associated with effects on activities of daily living and
cognitive impairment in the elderly [12,13]. Moreover, both are
Correspondence: Nopporn Howteerakul
significantly related with psychological problems, mood changes,
Department of Epidemiology, Faculty of Public Health, Mahidol and depression [1,7,14,15]. Previous studies have indicated that
University, 420/1 Rajvithi Road, Ratchathewi, Bangkok 10400, Thailand supportive ties and close relationships among family members are
E-mail: [email protected] positively related to good sleep quality, while aversive ties predict
Received: Feb 7, 2018 / Accepted: May 14, 2018 / Published: May 14, 2018 poor sleep quality [16,17].
This article is available from: http://e-epih.org/ The prevalence of poor sleep quality varies according to the de-
This is an open-access article distributed under the terms of the Creative mographic characteristics of the population. In Australia, epide-
Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
which permits unrestricted use, distribution, and reproduction in any medium, miological surveys have found that roughly 13.0-33.0% of elderly
provided the original work is properly cited. people reported frequent sleep difficulties or problems maintain-
2018, Korean Society of Epidemiology ing sleep [18]. In China, Li et al. [8] reported that the prevalence

www.e-epih.org | 1
Epidemiol Health 2018;40:e2018018

of poor sleep quality among the elderly in rural Anhui Province sub-district, in rural Chiang Rai Province in northern Thailand.
was 49.7%. In that study, the associated factors included age, physi- Wiang Ka Long is 785 km away from Bangkok, the capital of Thai-
cal and mental health, and nutritional intake. In Korea, Park et al. land (Figure 1). This sub-district was selected because of its high
[15] reported that approximately 60.0% of the elderly aged 65 years proportion of elderly inhabitants. In 2015, the elderly accounted
and above, sampled from 5 community health centers in Gyeong- for 23.9% (1,618 of 6,751) of the population [24]. The sample size
gi Province and Seoul, reported having poor sleep quality. In Thai- was estimated using the single-proportion formula with finite
land, epidemiologic data from 1997 to 2015 indicated that the prev- population correction [33] and a 95% confidence interval (CI).
alence of poor sleep quality ranged from 42.7 to 57.1%, and that The sample size calculation was based on 46.3% of elderly Thais
the factors leading to sleep disturbances were pain, anxiety, depres- reporting insomnia in a previous epidemiological survey [20].
sion, and poor perceived health [19-21]. A more recent study by Precision was set at 5.5%, and the sample size was calculated to be
Aunjitsakul et al. [22] revealed that the prevalence of poor sleep 266 to represent the population of 1,618 elderly inhabitants [24].
quality among the elderly seeking care at the outpatient department A total of 279 elderly individuals were recruited to allow for a non-
of a teaching hospital in a southern province of Thailand was 70.8%. response rate of approximately 5.0%. The sub-district has 15 vil-
The main complaints were short sleep duration (99.7%), poor sleep lages, and 17-20 households per village were randomly selected.
efficiency (61.6%), and daytime dysfunction (51.6%). However, The participants were selected using probability proportional to
that study did not identify any factors associated with poor sleep size sampling, to obtain study samples that represented each vil-
quality in its multivariable analysis. The cited studies used differ- lage and age group (60-69, 70-79, and ≥ 80 years). If no one was
ent types of sleep quality measures, different definitions of older home, the nearest household was selected. When a household had
adults (some used 60 years of age as a cutoff for inclusion, while more than 1 eligible elderly member, an elderly household mem-
others used 65 years), and varied in terms of study location, in- ber was chosen at random. The inclusion criteria were being a reg-
cluding urban, rural, and hospital settings. This should be kept in istered resident of the study area, being aged ≥ 60 years, having no
mind when comparing the prevalence rates of poor sleep quality problems hearing and speaking, and being willing to participate
in older adults. in the study. The elderly who were unavailable for an interview
The proportion of the elderly in Thailand has been steadily in- during the study or those who the research assistant determined,
creasing. The number of people over 60 is projected to increase through interacting with them, to be too ill or cognitively impaired
from 10.73 million in 2015 to 18.36 million in 2030, when this age to participate were excluded. Of the 279 participants, 13 were ex-
group will account for 26.9% of the total population [23]. cluded because their questionnaires were incomplete. Therefore,
Wiang Ka Long is a sub-district in rural Chiang Rai Province. responses from 266 participants were analyzed.
Demographic data from the Chiang Rai Provincial Health Office
show that the proportion of the elderly (age 60 and above) in this Instrument and measurements
sub-district has increased from 18.9% in 2005 to 24.9% in 2015, The research instruments of this study consisted of 5 parts.
and will continue to increase quickly in future decades [24]. In General characteristics included age, sex, education, occupa-
Chiang Rai Province, the 5 leading diseases of the elderly are hy- tion, religious activities, chronic diseases, daytime napping, use of
pertension, diabetes mellitus, ischemic heart disease, stroke, and sleep medication, alcohol consumption, smoking, and perceived
chronic obstructive pulmonary disease [25]. All these chronic dis- environmental factors disturbing sleep quality. These variables have
eases are associated with poor sleep quality [26,27]. Previous stud- been reported to be associated with poor sleep quality in previous
ies have also found sleep problems to be common among the el- studies [4,7,15,19-21,26,27,34,35].
derly, to the point that many elderly individuals perceive sleep prob- Depression was measured using the Thai version of the Patient
lems as part of the normal aging process [28,29]. Therefore, sleep Health Questionnaire [36,37], which contains 9 questions used to
problems in the elderly are frequently underexplored and untreat- screen for depression. Each question is rated on scale of 0 to 3,
ed [15,30]. In Thailand, information about sleep quality among and the total score is 27 points. A score of < 7 is defined as no de-
the community-dwelling elderly is still scarce, particularly for those pressive symptoms, 7-12 as mild depression, 13-18 as moderate
living in rural areas. We conducted a study examining the preva- depression, and 19-27 as severe depression. The Cronbach alpha
lence of poor sleep quality and its associated factors using a local- of this instrument was found to be 0.92.
ized version of the Pittsburgh Sleep Quality Index (PSQI) [31,32] Activities of daily living were assessed with the modified Bar-
to help fill this information gap. The results of our study can be thel Index [38,39], which comprises 10 questions used to assess
used to identify appropriate interventions to promote the physical hygiene, bowel control, bladder control, toileting, feeding, mov-
and mental health of the elderly with poor sleep quality. ing, ambulation, chair transfer, dressing, stair climbing, and bath-
ing oneself. The total score is 20. A score of 0-4 is defined as de-
MATERIALS AND METHODS pendence, 5-11 as some dependence, and 12-20 as independence.
The Cronbach alpha was 0.95.
Study participants Family relationships were measured by the Family Relation-
This cross-sectional study was conducted in the Wiang Ka Long ships Scale developed by the Technical Promotion and Support

2 | www.e-epih.org
 Thichumpa W et al. : Assessing sleep quality among elderly Thais

Figure 1. Map showing Wiang Ka Long sub-district.

Offices of the Thai Department of Social Welfare [40]. This in- cial activity.’ Higher scores represent more daytime dysfunction.
strument comprises 10 questions with a 3-point response format: Participants with a T-PSQI global score of > 5 were defined as
0 (never), 1 (sometimes), and 2 (always). Scores of > 10 are de- having poor sleep quality, while T-PSQI global scores of ≤ 5 were
fined as good relationships, and scores ≤ 10 as poor relationships. defined as good sleep quality. The Cronbach alpha was 0.84 [32].
The Cronbach alpha was 0.95.
Sleep quality was examined using the Thai version of the Pitts- Data collection
burgh Sleep Quality Index (T-PSQI) [31,32], a self-rated question- Data were collected by the principal investigator and 2 trained
naire used to assess subjective sleep quality and sleep disturbance research assistants between June and September 2016. Written in-
over 1 month. It contains 19 items, which generate 7 component formed consent was obtained from all participants. The research
scores: subjective sleep quality (1 item), sleep latency (2 items), team explained the objectives of the study and interviewed the
sleep duration (1 item), habitual sleep efficiency (3 items), sleep participants using structured questionnaires.
disturbance (9 items), use of sleep medication (1 item) and day-
time dysfunction (2 items). Each component has a range of 0-3 Statistical analysis
points, ranging from 0 (no difficulty) to 3 (severe difficulty) [41]. Data were analyzed using SPSS version 18.0 (SPSS Inc., Chica-
The totals of the 7 component scores are added to generate the fi- go, IL, USA). Descriptive statistics such as percentage, mean, and
nal score (range, 0-21). Higher scores indicate poorer subjective standard deviation (SD) were used to describe all study factors.
sleep quality. Habitual sleep efficiency was calculated by: hours of Multiple logistic regression was used to obtain odds ratios (ORs)
sleep/(get up time – usual bed time)× 100. Sleep disturbance was and 95% CIs to examine the associations between factors ana-
calculated from the T-PSQI items ‘wake up in the middle of the lyzed in the study and poor sleep quality. In the analysis, sleep
night or early morning,’ ‘need to get up to use the bathroom,’ ‘can- quality was dichotomized into good sleep quality and poor sleep
not breathe comfortably,’ ‘cough or snore loudly,’ ‘feel too cold,’ ‘feel quality. All study factors with a p-value < 0.05 in the univariate
too hot,’ ‘have bad dreams,’ and ‘have pain.’ Higher scores repre- analysis and biological plausibility were tested for the absence of
sent more sleep disturbances. Daytime dysfunction was calculat- multicollinearity among study factors by inspecting the magnitude
ed from the T-PSQI items ‘during the past month, how often have of the standard error (SE) of each variable. The magnitude of the
you taken medicine to help you sleep (prescribed or over the coun- SEs should hover around 0.001-5.0 [42]. In model I, the predic-
ter)’ and ‘during the past month, how often you have you had trou- tors were adjusted for age, sex, and education level. In model II,
ble staying awake while driving, eating meals, or engaging in so- the predictors were adjusted for age, sex, education level, mild de-

www.e-epih.org | 3
Epidemiol Health 2018;40:e2018018

pression, and family relationships. The p-values < 0.05 were con- Table 1. General characteristics of the subjects (n=266)
sidered to indicate statistical significance. Variable n (%)
Sex
Ethical approval Male 108 (40.6)
The study protocol was approved by the ethics committee of the
Female 158 (59.4)
Faculty of Public Health, Mahidol University (COA. no. MUPH Age (yr)
2016-086; June 7, 2016). 60-69 150 (56.4)
70-79 83 (31.2)
RESULTS ≥ 80 33 (12.4)
Mean±SD (range) 67.8±7.1 (60-85)
General characteristics Education level
Of the 266 participants, 59.4% were female, and 56.4% were No school 59 (22.2)
aged 60-69 years. The mean± SD age was 67.8± 7.1 years. Moreo- Primary school (grade 4 intheformer Thai 68 (25.6)
ver, 52.2% had finished primary school or higher, 47.4% were not educational system)
working or a housewife, 34.2% were agricultural laborers, and >Primary school (higher than grade 4) 139 (52.2)
Occupation
9.8% were self-employed. Religious activities were performed by
None/housewife 126 (47.4)
95.9% of the participants. Of the respondents, 59.8% had at least 1
Agriculture 91 (34.2)
chronic disease; 38.2% had hypertension, and 14.7% had diabetes. Self-employed 26 (9.8)
Furthermore, 22.2% often napped in the daytime, 9.4% used sleep Employee 20 (7.5)
medication, 19.2% reported that they sometimes or often drank Others 3 (1.1)
alcohol, and 10.1% reported that they sometimes or often smoked. Religious activities (e.g., alms-giving, listening to sermons,
The environmental factors reported as disturbing sleep quality meditation)
were noise (12.4%), feeling too hot (6.0%), and light (3.8%), as Yes 255 (95.9)
shown in Table 1. No 11 (4.1)
Have 1 or more chronic diseases (hypertension, 159 (59.8)
diabetes, etc.)
Activities of daily living, depression, and family
Type of chronic disease1
relationships Hypertension 102 (38.2)
As shown in Table 2, only 1.1% of the participants reported
Diabetes 39 (14.7)
they had some dependence on others for daily living activities.
Arthritis 23 (8.6)
The mean Barthel Index mean ± SD score was 19.7 ± 1.4. Symp- Heart disease 10 (3.8)
toms of mild depression were shown in 12.0% of the participants. Others (cancer, kidney disease, hypercholesterolemia, 82 (30.8)
The mean± SD depression score was 2.6± 2.7. Of the participants, etc.)
27.1% had poor family relationships. The mean family relation- Daytime napping
ship mean ± SD score was 14.6± 2.2. No 85 (31.9)
Sometimes 122 (45.9)
Sleep quality Often 59 (22.2)
The mean ± SD T-PSQI score was 5.86± 2.47. The prevalence of Use of sleep medication 25 (9.4)
poor sleep quality (T-PSQI global score, > 5) was 44.0%. The Alcohol consumption
mean± SD sleep duration was 7.0± 1.3 hours. The mean± SD sleep Never or no longer drink 215 (80.8)
latency was 30.0 ± 27.5 minutes. The mean ± SD sleep efficiency Sometimes or often drink 51 (19.2)
among good sleepers was 92.6 ± 7.7%, while it was 83.7 ± 16.6% Smoking
among poor sleepers. The mean T-PSQI score of participants with Never or no longer smoke 239 (89.8)
at least 1 chronic disease or mild depression was slightly higher Sometimes or often smoke 27 (10.1)
than that of participants without these conditions (5.02 vs. 4.63 Reported environmental factors disturbing sleep quality1
for chronic disease and 6.25 vs. 4.68 for depression), as shown in None 186 (69.9)
Table 3. Noise 33 (12.4)
Table 4 shows the number and percentage of poor sleepers (PSQI Feeling too hot 16 (6.0)
> 5) by sleep quality components and items. Among the 117 poor Too much light 10 (3.8)
sleepers, 14.5% reported poor subjective sleep quality, 86.3% re- Feeling too cold 5 (1.9)
ported sleep latency of more than 15 minutes, 23.1% reported a Other (e.g. uncomfortable mattress, insects, 22 (8.3)
good or recommended sleep duration (> 7 hours), 59.8% reported bad smells)
a low but possibly appropriate sleep duration (6-7 hours), 41.9% SD, standard deviation.
had poor sleep efficiency (< 85.0%), 82.1% reported frequent sleep 1
Multiple responses.

4 | www.e-epih.org
 Thichumpa W et al. : Assessing sleep quality among elderly Thais

Table 2. ADLs, depression, and family relationships among the sub- disturbances (≥ 1/wk), 18.0% used sleep medication (< 1/ wk or
jects ≥ 1/wk), and 35.9% reported daytime dysfunction (< 1/wk or
Variable n (%) ≥ 1/wk).
Table 5 depicts the details of reported sleep disturbance prob-
ADLs (BI score)
Independence (12-20) 263 (98.9)
Some dependence (5-11) 3 (1.1)
Table 3. The T-PSQI global score among the subjects according to
the presence ofchronic diseases and depression
Dependence (0-4) 0 (0.0)
Mean±SD (range) 19.7±1.4 (6-20) Variables Mean± SD
Depression (score) No. of chronic diseases
Mild (7-12) 32 (12.0) None 4.63±2.26
None (<7) 234 (88.0) One 5.02±2.45
Mean±SD (range) 2.6±2.7 (0-12) Two or more 5.01±2.45
Family relationships (score) Depression (score)
Good (≥10) 194 (72.9) None (<7) 4.68±2.24
Poor (<10) 72 (27.1) Mild depression (7-12) 6.25±2.82
Mean±SD (range) 14.6±2.2 (0-20)
T-PSQI, Thai version of the Pittsburgh Sleep Quality Index; SD, standard
ADL, activities of daily living; BI, Barthel Index; SD, standard deviation. deviation.

Table 4. Good sleepers and poor sleepers among the subjects by sleep quality componentsand items1
Good sleeper Poor sleeper Total
Sleep quality componentsand items
PSQI≤5 (n=149) PSQI>5 (n=117) (n=266)

Subject sleep quality

Good (0, 1) 149 (100.0) 100 (85.5) 249 (93.6)
Poor (2, 3) 0 (0.0) 17 (14.5) 17 (6.4)
Sleep latency (min)2
Good (≤15) (0) 77 (51.7) 16 (13.7) 93 (35.0)
Poor (>15) (1-3) 72 (48.3) 101 (86.3) 173 (65.0)
Sleep duration (hr)
Good or recommended (>7) (0)3 119 (79.9) 27 (23.1) 146 (54.9)
Low, may be appropriate (6-7) (1)3 0 (0.0) 70 (59.8) 70 (26.3)
Poor (≤6) (2, 3) 30 (20.1) 20 (17.1) 50 (18.8)
Habitual sleep efficiency (%)4
Good (≥85) (0) 127 (85.2) 68 (58.1) 195 (73.3)
Poor (<85) (1-3) 22 (14.8) 49 (41.9) 71 (26.7)
Sleep disturbance (times/wk)5
Low (<1 or not during past month) (0, 1) 115 (77.2) 21 (17.9) 136 (51.1)
High (≥1) (2, 3) 34 (22.8) 96 (82.1) 130 (48.9)
Use of sleep medication
Never in 1 month (0) 145 (97.3) 96 (83.1) 241 (90.6)
<1/wk or ≥1/wk (1-3) 4 (3.4) 21 (18.0) 25 (9.4)
Daytime dysfunction6
Never in 1 month (0) 135 (90.6) 75 (64.1) 210 (78.9)
<1/wk or ≥1/wk (1-3) 14 (9.4) 42 (35.9) 56 (21.1)
Values are presented as number (%).
PSQI, Pittsburgh Sleep Quality Index.
1
Scores range from 0 to 3; 0 indicates no difficulty and 3 indicates severe difficulty.
2
The proportion of the first item of the sleep latency component was presented to make it easier to understand.
3
Defined according to the National Sleep Foundation [43].
4
Habitual sleep efficiency was calculated as: hours of sleep/(get up time – usual bed time)×100.
Derived from PSQI items 5b-5j; higher scores indicate more sleep disturbance; item 5j is an open-ended question for describing trouble sleeping.
5

It was not included in the analysis, as there were no responses.

6
Derived from PSQI items 7-8; higher scores indicate more daytime dysfunction.

www.e-epih.org | 5
Epidemiol Health 2018;40:e2018018

lems (> 1/wk) among poor sleepers. The most frequently reported In the multivariate analysis, age, sex, education level, and all
sleep problem was needing to get up to use the bathroom (71.8%), other variables found to be significant in the univariate analysis
followed by waking up in the middle of the night or early morn- were simultaneously included in the final analysis. In model I, the
ing (64.9%) and difficulty falling asleep within 30 minutes significant predictors were being female (OR, 1.90; 95% CI, 1.12
(57.3%). to 3.21), higher education level (OR, 3.13; 95% CI, 1.45 to 6.74 for
primary school; OR, 2.22; 95% CI, 1.06 to 4.63 for higher than
Factors associated with poor sleep quality primary school). The Hosmer and Lemeshow goodness-of-fit test
In the univariate analysis, 3 factors were significantly associated suggested the model was a good fit with the data, with p= 0.164.
with poor sleep quality among the elderly: a higher education lev- In model II, the significant predictors were being female (OR, 1.74;
el (OR, 2.72; 95% CI, 1.31 to 5.64 for primary school; OR, 1.83; 95% CI, 1.10 to 3.02), higher education level (OR, 3.03; 95% CI,
95% CI, 1.02 to 3.50 for higher than primary school), mild depres- 1.34 to 6.86 for primary school; OR, 2.48; 95% CI, 1.31 to 5.44 for
sion (OR, 3.22; 95% CI, 1.46 to 7.11), and poor family relation- higher than primary school), mild depression (OR, 2.65; 95% CI,
ships (OR, 3.95; 95% CI, 2.22 to 7.03). 1.11 to 6.36), and poor family relationships (OR, 3.65; 95% CI,
1.98 to 6.75). The Hosmer and Lemeshow goodness-of-fit test also
suggested a good fit with the data (p= 0.218), as shown in Table 6.
Table 5. Reported sleep disturbances among 117 poor sleepers
(PSQI>5)
Sleep disturbance (>1/wk) n (%)
DISCUSSION
Cannot fall asleep within 30 minutes 67 (57.3) The results of this study revealed that the prevalence of poor
Wake up in the middle of the night or early morning 76 (64.9) sleep quality among the elderly in the study area was 44.0%. The
Need to get up to use the bathroom 84 (71.8) mean± SD T-PSQI global score was 5.86± 2.47. This rate is com-
Cannot breathe comfortably 5 (4.3) parable with the prevalence of 42.7% found in a 1997 study of Thai
Cough or snore loudly 30 (25.6) community-dwelling elderly in the Nong Chok district of Bang-
Feel too cold 5 (4.3) kok reported by Sukying & Ninchaikovit [19]. Although Nong
Feel too hot 16 (13.6) Chok is now an urban area, in 1997 it was rural and agriculture
Have bad dreams 9 (7.7) was the main economic activity, making it similar to the present-
Have pain 20 (17.1) day Wiang Ka Long sub-district. An epidemiological study by
PSQI, Pittsburgh Sleep Quality Index. Sukying et al. [20] on insomnia among 40,111 individuals aged 60

Table 6. Factors associated with poor sleep quality among the subjects
Crude Model I1 Model II2
Variable p-value
OR (95%CI) aOR (95%CI) aOR (95% CI)
Age (yr)
60-69 1.00 (reference) 1.00 (reference) 1.00 (reference)
70-79 1.13 (0.66, 1.93) 1.31 (0.71, 2.40) 1.15 (0.62, 2.12) 0.38
≥80 0.73 (0.33, 1.59) 1.03 (0.43, 2.46) 0.90 (0.37, 2.20) 0.94
Sex
Male 1.00 (reference) 1.00 (reference) 1.00 (reference)
Female 1.62 (0.98, 2.67) 1.90 (1.12, 3.21) 1.74 (1.10, 3.02) 0.02
Education level
No school 1.00 (reference) 1.00 (reference) 1.00 (reference)
Primary school 2.72 (1.31, 5.64) 3.13 (1.45, 6.74) 3.03 (1.34, 6.86) 0.008
>Primary school 1.83 (1.02, 3.50) 2.22 (1.06, 4.63) 2.48 (1.31, 5.44) 0.02
Mild depression
No 1.00 (reference) - 1.00 (reference)
Yes 3.22 (1.46, 7.11) - 2.65 (1.11, 6.36) 0.03
Family relationships
Good 1.00 (reference) - 1.00 (reference)
Poor 3.95 (2.22, 7.03) - 3.65 (1.98, 6.75) <0.001
OR, odds ratio; CI, confidence interval; aOR, adjusted OR.
1
Model I: adjusted for age, sex and education level.
2
Model II: adjusted for age, sex, education level, mild depression, and family relationships.

6 | www.e-epih.org
 Thichumpa W et al. : Assessing sleep quality among elderly Thais

years and above from 4 regions of Thailand found the prevalence greater earning capacity of the elderly who live in urban areas may
of poor sleep quality to be 46.3%. Choombuathong et al. [21] and lead to a reduction in some of the factors that cause low-quality
Aunjitsakul et al. [22] reported higher prevalence rates (57.1 and sleep. Additionally, the earning gap between the education levels of
70.8%, respectively) than that found in the current study. This is ‘some primary school’ and ‘completed primary school’ in rural ar-
most likely because these studies examined sleep quality among eas is unlikely to be dramatic, but the earning gap between the
the elderly in very different contexts (a nursing home and an out- levels of ‘completed primary school’ and ‘completed university’ in
patient department) from that of the current study. an urban setting is likely to be large.
When the proportions of poor sleep quality components/items Mild depression was an important predictor of sleep quality,
among the 117 poor sleepers were reviewed, the 2 most frequent consistent with previous studies [14,45]. The prevalence of depres-
complaints were poor sleep latency (86.3%) and frequent sleep dis- sion in Thailand has been reported to range from 5.9 to 38.9%
turbances (82.1%). This finding is congruent with previous stud- [46-48] depending on the study population and the instrument
ies [7,23] that found the most frequently occurring sleep problem used in the study. Females have been shown to be more likely to
in elderly was poor sleep latency, which occurs when people may have depression than males [48,49]. Depression can have several
spend some hours in bed before falling asleep. In our study, 57.3% impacts on daily activities, emotional well-being, and the risk of
(67 of 117) of the poor sleepers reported having difficulty falling suicide [50]. Therefore, healthcare providers should pay more at-
asleep within 30 minutes. Within the category of sleep disturbanc- tention to sleep habits in the elderly. Appropriate depression screen-
es, 71.8% (84 of 117) of the poor sleepers reported getting up to ing, access to mental health care, and interventions to manage de-
use the bathroom and 64.9% (76 of 117) reported waking up in pression might be able to prevent further potential negative effects
the middle of the night or early morning. This finding is consist- related to depression in the elderly.
ent with the studies by Knodel et al. [23] and Barthlen [44]. The A poor family relationship was the most significant predictor of
proportion of poor sleepers getting a good or recommended sleep sleep quality in this study. This finding is consistent with those of
duration (> 7 hours) was low (23.1%). Moreover, 64.9% (76 of 117) the study by Ailshire & Burgard [16], who found that having a
of the poor sleepers and 70.5% (105 of 149) of the good sleepers strained family relationship was associated with more troubled
napped in the daytime. Daytime napping might alleviate some of sleep, whereas a supportive family relationship was associated
the problems with poor sleep duration in both good and poor sleep- with less troubled sleep. Responses to 4 items in the family rela-
ers. It would be beneficial to investigate the influence of daytime tionship scale revealed that 28.6% of the elderly reported never
napping on nocturnal sleep in the community-dwelling elderly spending a significant amount of time participating in family ac-
aged 60 years and above. tivities such as reading, watching TV, and listening to the radio;
In the multivariate analysis, sex, education level, mild depres- 21.8% reported never resolving conflicts or consulting other fam-
sion, and family relationships were significant predictors of sleep ily members before making a decision; 21.1% reported never shar-
quality. In model I, sex changed from a nearly significant predic- ing ideas among family members; and 20.7% never experienced
tor (p = 0.059) in the univariate analysis to a significant predictor good compromises among family members (data not shown).
(p < 0.05) of sleep quality. This was because, as we found through Given the strong relationship between the quality of family rela-
subgroup and stratification-based analyses, more females than tionships and the quality of sleep, healthcare providers should
males had finished primary school or higher (56.8 vs. 43.2%), had provide interventions that promote good relationships among
mild depression (15.2 vs. 7.4%), and reported poor family rela- family members. Good family relationships can be a buffer against
tionships (27.8 vs. 25.9%). In addition, the mean T-PSQI score for the loneliness and depression that affect sleep quality in the elder-
females was higher than for males (6.15 ± 2.51 vs. 5.44 ± 2.09). ly [17].
These results suggest that sex alone did not account for sleep qual- This study had 4 main limitations that should be considered
ity. Instead, it was partially influenced by other variables in the when interpreting its findings. First, the results can be generalized
model. This finding is consistent with that of Sukying et al. [20]. only to the study area due to the power of the sample size calcula-
In this study, a higher level of education (completion of prima- tion. Second, sleep quality was measured by self-reporting only
ry school) was associated with poor sleep quality. This may have and no objective measures, such as a polysomnography test, were
been due to occupational stress, since some of the elderly with a included. Third, this study focused on family relationships, which
higher level of formal education were still working. This is in line were found to be a major factor affecting sleep quality in the el-
with the results of Yao et al. [45]. However, this result does not derly in previously published studies [16,17,44]. However, some
agree with the finding of Luo et al. [4] who found that Chinese el- elderly persons might have poor relationships with non-family
derly in Shanghai with less education had poorer sleep quality than members, such as friends or neighbors, that might affect their
those with more education. Possible reasons for this discrepancy sleep quality. This possibility was not investigated in the present
include differences between the studies regarding the mean years study. Fourth, the cross-sectional study design limits the degree to
of education and study location. In the Shanghai study, the partic- which conclusions can be drawn regarding the causal relationship
ipants’ education level was higher than that of the participants in between the factors analyzed in this study and sleep quality among
the present study (11.0 ± 4.5 vs. 3.0 ± 2.3 years). In addition, the the elderly in the study area. To increase the generalizability of the

www.e-epih.org | 7
Epidemiol Health 2018;40:e2018018

findings, this study should be replicated with a greater sample size urban community: results from the Shanghai aging study. PLoS
and larger study area. One 2013;8:e81261.
Our findings indicated that the prevalence of poor sleep quality 5. Amer MS, Hamza SA, El Akkad RM, Abdel Galeel YI. Does self-
among the elderly in the study area was high, and the major sig- reported sleep quality predict poor cognitive performance among
nificant predictors of poor sleep quality were mild depression and elderly living in elderly homes? Aging Ment Health 2013;17:788-
poor family relationships. Given the impact of poor sleep quality 792.
on health, healthcare providers should regularly screen for sleep 6. Ancoli-Israel S, Cooke JR. Prevalence and comorbidity of insom-
quality and depression and provide interventions targeting sleep nia and effect on functioning in elderly populations. J Am Geriatr
health education and behavior change. Educational messages should Soc 2005;53(7 Suppl):S264-S271.
include recommending that the elderly go to bed at the same time 7. Ancoli-Israel S, Ayalon L, Salzman C. Sleep in the elderly: normal
each night and wake up at the same time each morning, and that variations and common sleep disorders. Harv Rev Psychiatry 2008;
they should make their bedroom environment quiet, dark, and 16:279-286.
not too hot or cold. Further interventions could include providing 8. Li J, Yao YS, Dong Q, Dong YH, Liu JJ, Yang LS, et al. Characteri-
consultations to encourage family members to participate in ac- zation and factors associated with sleep quality among rural el-
tivities with their elderly family members, as well as providing in- derly in China. Arch Gerontol Geriatr 2013;56:237-243.
formal training on resolving conflicts, sharing ideas, and finding 9. Gu D, Sautter J, Pipkin R, Zeng Y. Sociodemographic and health
compromises. Psychosocial interventions with a non-pharmaco- correlates of sleep quality and duration among very old Chinese.
logical approach, such as community exercise programs, should Sleep 2010;33:601-610.
be also introduced to improve sleep quality and reduce depression 10. Subramanian S, Surani S. Sleep disorders in the elderly. Geriatrics
among the elderly in the community. 2007;62:10-32.
11. Beck SL, Schwartz AL, Towsley G, Dudley W, Barsevick A. Psy-
chometric evaluation of the Pittsburgh Sleep Quality Index in can-
ACKNOWLEDGEMENTS
cer patients. J Pain Symptom Manage 2004;27:140-148.
This study was supported for publication by the China Medical 12. Duclos C, Beauregard MP, Bottari C, Ouellet MC, Gosselin N.
Board, Faculty of Public Health, Mahidol University. This paper The impact of poor sleep on cognition and activities of daily liv-
was first presented as a poster at the 1st APACPH Bangkok Re- ing after traumatic brain injury: a review. Aust Occup Ther J 2015;
gional Conference and the 8th International Public Health Confer- 62:2-12.
ence Toward Achieving the 2030 Sustainable Development Goals, 13. Renu Sharma R, Das K, Randhawa H, Subodh BN. Sleep quality
May 25 to May 26, 2017. and its effect on activities of daily living among substance depend-
ent subjects. Delhi Psychiatry J 2015;18:32-39.
14. Tsuno N, Besset A, Ritchie K. Sleep and depression. J Clin Psy-
CONFLICT OF INTEREST
chiatry 2005;66:1254-1269.
The authors have no conflicts of interest to declare for this study. 15. Park JH, Yoo MS, Bae SH. Prevalence and predictors of poor sleep
quality in Korean older adults. Int J Nurs Pract 2013;19:116-123.
16. Ailshire JA, Burgard SA. Family relationships and troubled sleep
ORCID
among U.S. adults: examining the influences of contact frequency
Weerakorn Thichumpa: https://orcid.org/0000-0002-7947-4705; and relationship quality. J Health Soc Behav 2012;53:248-262.
Nopporn Howteerakul: https://orcid.org/0000-0003-3462-3778; 17. Kent RG, Uchino BN, Cribbet MR, Bowen K, Smith TW. Social
Nawarat Suwannapong: https://orcid.org/0000-0002-9904-3590; relationships and sleep quality. Ann Behav Med 2015;49:912-917.
Visasiri Tantrakul: https://orcid.org/0000-0001-7444-8674 18. Cunnington D, Junge MF, Fernando AT. Insomnia: prevalence,
consequences and effective treatment. Med J Aust 2013;199:S36-
S40.
REFERENCES
19. Sukying C, Ninchaikovit T. Sleep problems and their correlates in
1. Ohayon MM. Epidemiology of insomnia: what we know and a Thai community in Bangkok: a preliminary report. J Psychiatr
what we still need to learn. Sleep Med Rev 2002;6:97-111. Assoc Thailand 1997;42:66-76 (Thai).
2. Ohayon MM, Paiva T. Global sleep dissatisfaction for the assess- 20. Sukying C, Bhokakul V, Udomsubpayakul U. An epidemiological
ment of insomnia severity in the general population of Portugal. study on insomnia in an elderly Thai population. J Med Assoc
Sleep Med 2005;6:435-441. Thai 2003;86:316-324.
3. Sagayadevan V, Abdin E, Binte Shafie S, Jeyagurunathan A, Sam- 21. Choombuathong A, Chalopatham W, Dhongyooyen P, Chokchai­
basivam R, Zhang Y, et al. Prevalence and correlates of sleep prob- worrarat S, Thamwattana K, Changsap B, et al. Sleep quality and
lems among elderly Singaporeans. Psychogeriatrics 2017;17:43-51. factors related to sleep and self-esteem in the elderly. J Health Sci
4. Luo J, Zhu G, Zhao Q, Guo Q, Meng H, Hong Z, et al. Prevalence 2017;24:833-843 (Thai).
and risk factors of poor sleep quality among Chinese elderly in an 22. Aunjitsakul W, Pitanupong J, Werachattawan N, Anantapong K.

8 | www.e-epih.org
 Thichumpa W et al. : Assessing sleep quality among elderly Thais

Sleep quality among elderly people within the outpatient depart- 36. Thai Ministry of Public Health. Guidebook of depressive disor-
ment of Songklanagarind Hospital, Thailand: a cross-sectional ders surveillance and care: provincial level; 2014 [cited 2016 Feb
study. Songkla Med J 2017;35:195-204. 10]. Available from: http://www.thaidepression.com/www/58/
23. Knodel J, Teerawichitchainan B, Prachuabmoh V, Pothisiri W. guidebookdepress.pdf (Thai).
The situation of Thailand’s older population an update based on 37. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a
the 2014 survey of older persons in Thailand; 2015 [cited 2017 brief depression severity measure. J Gen Intern Med 2001;16:606-
Dec 20]. Available from: https://www.psc.isr.umich.edu/pubs/ 613.
pdf/rr15-847.pdf. 38. Rachpukdee S, Howteerakul N, Suwannapong N, Tang-Aroonsin
24. Chiang Rai Province Health Office, Thailand. Standard popula- S. Quality of life of stroke survivors: a 3-month follow-up study. J
tion pyramid report. Chiang Rai: Chiang Rai Province Health Stroke Cerebrovasc Dis 2013;22:e70-e78.
Office; 2015, p. 5-10 (Thai). 39. Granger CV, Dewis LS, Peters NC, Sherwood CC, Barrett JE.
25. Chiang Rai Province Health Office, Thailand. Annual health re- Stroke rehabilitation: analysis of repeated Barthel index measures.
port 2015. Chiang Rai: Chiang Rai Province Health Office; 2015, Arch Phys Med Rehabil 1979;60:14-17.
p. 11 (Thai). 40. Technical Promotion and Support Offices. Sub-district social wel-
26. Foley D, Ancoli-Israel S, Britz P, Walsh J. Sleep disturbances and fare for the elderly with community participation. Bangkok: Tech-
chronic disease in older adults: results of the 2003 National Sleep nical Promotion and Support Offices; 2012, p. 108 (Thai).
Foundation Sleep in America Survey. J Psychosom Res 2004;56: 41. Hawkins J. Pittsburgh Sleep Quality Index (PSQI); 2016 [cited
497-502. 2017 Dec 10]. Available from: http://www.goodmedicine.org.uk/
27. Koyanagi A, Garin N, Olaya B, Ayuso-Mateos JL, Chatterji S, Leo­ files/assessment,%20pittsburgh%20psqi.pdf.
nardi M, et al. Chronic conditions and sleep problems among 42. Chan YH. Biostatistics 202: logistic regression analysis. Singapore
adults aged 50 years or over in nine countries: a multi-country Med J 2004;45:149-153.
study. PLoS One 2014;9:e114742. 43. Hirshkowitz M, Whiton K, Albert SM, Alessi C, Bruni O, Don-
28. Hoffman S. Sleep in the older adult: implications for nurses (CE). Carlos L, et al. National Sleep Foundation’s updated sleep dura-
Geriatr Nurs 2003;24:210-216. tion recommendations: final report. Sleep Health 2015;1:233-243.
29. Rodriguez JC, Dzierzewski JM, Alessi CA. Sleep problems in the 44. Barthlen GM. Sleep disorders. Obstructive sleep apnea syndrome,
elderly. Med Clin North Am 2015;99:431-439. restless legs syndrome, and insomnia in geriatric patients. Geriat-
30. Blackwell T, Yaffe K, Ancoli-Israel S, Schneider JL, Cauley JA, Hill- rics 2002;57:34-40.
ier TA, et al. Poor sleep is associated with impaired cognitive func- 45. Yao KW, Yu S, Cheng SP, Chen IJ. Relationships between person-
tion in older women: the study of osteoporotic fractures. J Ger- al, depression and social network factors and sleep quality in com-
ontol A Biol Sci Med Sci 2006;61:405-410. munity-dwelling older adults. J Nurs Res 2008;16:131-139.
31. Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. 46. Wongpoom T, Sukying C, Udomsubpayakul U. Prevalence of de-
The Pittsburgh Sleep Quality Index: a new instrument for psychi- pression among the elderly in Chiang Mai province. J Psychiatr
atric practice and research. Psychiatry Res 1989;28:193-213. Assoc Thailand 2011;56:103-116 (Thai).
32. Sitasuwan T, Bussaratid S, Ruttanaumpawan P, Chotinaiwattara- 47. Haseen F, Prasartkul P. Predictors of depression among older peo-
kul W. Reliability and validity of the Thai version of the Pittsburgh ple living in rural areas of Thailand. Bangladesh Med Res Counc
Sleep Quality Index. J Med Assoc Thai 2014;97 Suppl 3:S57-S67. Bull 2011;37:51-56.
33. Daniel WW, Cross CL. Biostatistics: a foundation for analysis in 48. Wongsuraprakit S, Santiprasitkul S. Situation of depression in
the health sciences. 10th ed. Hoboken: Wiley; 2013, p. 192. pre-screened risk groups in Muang district, Chiang Rai Province.
34. Halperin D. Environmental noise and sleep disturbances: a threat Thai J Nurs Counc 2012;27:91-105 (Thai).
to health? Sleep Sci 2014;7:209-212. 49. Wangtongkum S, Sucharitakul P, Wongjaroen S, Maneechompoo S.
35. Division of Sleep Medicine at Harvard Medical School. External Prevalence of depression among a population aged over 45 years
factors that influence sleep; 2007 [cited 2017 Dec 18]. Available in Chiang Mai, Thailand. J Med Assoc Thai 2008;91:1812-1816.
from: http://healthysleep.med.harvard.edu/healthy/science/how/ 50. Blazer DG. Depression in late life: review and commentary. J Ger-
external-factors. ontol A Biol Sci Med Sci 2003;58:249-265.

www.e-epih.org | 9