Sleep quality among type 2 diabetes mellitus patients in a private hospital setting in Yangon, Myanmar Open Access

Sleep is very important and can impact every facet of our daily life. Poor sleep quality also affects our health and well-being [1]. People who sleep well have a better cognitive function, improved memory, healthier immune systems, increased attentiveness and show better performance throughout the day [2]. However, 29% of the general adult population is affected by sleep disorders [3]. Epidemiologic studies revealed that shorter sleep duration and poor sleep quality can increase the risk of obesity [4] and lifestyle-related diseases such as cardiovascular diseases [5], type 2 diabetes mellitus (T2DM) [1, 6], poor glycemic control [7–11] and poor quality of life (QoL) [12–14].

Sleep disorders are commonly seen in T2DM patients because of the disease itself, its complications and other comorbidities [15–17]. The prevalence of poor sleep quality among T2DM patients ranges from 45% to 70% [9, 15, 18–22]. It is estimated that 53.4% of diabetic patients have suffered from impaired sleep compared to 29% in the general population [3]. Patients who suffer from poor sleep quality also have a dramatically increased burden when trying to maintain a healthy diet and take regular exercise as part of self-care among diabetes patients [21].

Diabetes is a global health problem and is prevalent in Myanmar [23]. In 2014, the national prevalence of diabetes mellitus among Myanmar adults aged 25–64 years was 10.5%, higher than the global diabetes prevalence of 8.8% [24]. Poor glycemic control may lead to long-term complications such as neuropathy [25], nephropathy [26] and retinopathy [27], whereas good control significantly reduces the risk of cardiovascular diseases [28] and the rate of amputation in T2DM patients by 40% and 60% [29].

Sleep disorders are common alongside depressive symptoms among T2DM patients [13, 15, 30]. The combined effect of sleep disorders and depressive symptoms has been shown to increase the risk of a patient having a poor QoL [13]. Other studies have suggested that good family relationships are associated with good sleep quality [31–33] and a better QoL [12–14]. Sleep quality plays a significant role in physical and mental health outcomes [1] as well as better diabetes self-care behaviors in T2DM patients [12]. Numerous studies related to sleep quality among T2DM patients have been conducted and have been well covered in Western countries. However, this information is rare in and pertaining to Myanmar. To the best of our knowledge, there has been no published literature regarding sleep quality among T2DM patients using the Pittsburgh Sleep Quality Index (PSQI) in Myanmar. The factors determining sleep quality among T2DM patients in Myanmar are still unclear. Therefore, this study aimed to assess the prevalence of poor sleep quality as well as its associated factors among T2DM patients in a private hospital in Yangon, Myanmar, using PSQI, a validated tool for the subjective measure of sleep quality [34].

This hospital-based cross-sectional study was conducted at an outpatient diabetic clinic of a private hospital with 180 beds, in Lanmadaw Township in Yangon, Myanmar. The study population was T2DM patients who were attending the diabetic clinic as outpatients during the study period.

In 2016, the total number of T2DM patients in the study hospital was 1300 [35]. The study sample size was estimated using the single-proportion formula with finite population correction [36] and a 95% confidence interval (CI). Due to there being no published literature for sleep quality in Myanmar, the prevalence of poor sleep quality was assumed to be 50%. Precision was set at 0.0525% and the sample size was calculated to be 275. To represent the population of 1300 T2DM patients, a total of 289 T2DM patients were recruited to allow for a nonresponse rate of 5%. Inclusion criteria were: (1) patients who were diagnosed with T2DM by a physician and were treated with hypoglycemic medication for at least six months; and (2) a willingness to participate in the study. Exclusion criteria were: (1) patients aged under 30 years because it is difficult to differentiate between T2DM and other types; (2) patients aged above 80 years; (3) patients who had other endocrine disorders, such as thyroid disease or chronic use of glucocorticoids; (4) patients who were pregnant or lactating; (5) patients who were seriously ill or had cognitive impairment as determined by interaction with the researcher when asking for informed consent.

• Part 1. Baseline characteristics of patients had 12 questions: age, sex, highest education, family income, marital status, working status, living with someone, alcohol drinking, smoking, body mass index (BMI), exercise and daytime napping. “Never smoked” referred to patients who smoked less than 100 cigarettes in their lifetime and did not currently smoke. “Ex-smoker” referred to patients who had smoked 100 or more cigarettes in their lifetime but did not smoke in the last 28 days. “Current smoker” referred to patients who had smoked 100 or more cigarettes in their lifetime and had smoked in the last 28 days [37]. “Never drank” referred to patients who had never tried any kind of alcoholic beverages or had had less than 12 alcoholic drinks in their lifetime. “Ex-drinker” referred to patients who had had at least 12 alcoholic drinks but none during the past 12 months. “Current drinker” referred to patients who had up to one drink per day for females and up to two drinks for males [38, 39].

• Part 2. Clinical factors had five questions: duration of diabetes, complication status, glycemic control, comorbidities and current medications.

• Part 3. Family relationships had ten questions that were translated into Myanmar by the researcher team from the Thai version of a family relationships questionnaire. Answers to the questions were rated on a scale of 0–2: always = 2, sometimes = 1 and never = 0. Total scores ranged from 0 to 20 points. According to the Thai family relationship scale, a total score >10 defined a good relationship [40]. Cronbach's alpha was 0.795.

• Part 4. Depression symptoms were assessed using the English version of the Patient Health Questionnaire-9 (PHQ-9) translated into Myanmar by the principal investigator. PHQ-9 comprises nine self-rated questions to screen for depression over the past two weeks, each question scores 0–3: “not at all” = 0, “several days” = 1, “more than half the days” = 2 and “nearly every day” = 3. Total scores ranged from 0 to 27 points. A score of <5 was defined as “normal”; a score of 5–9 as “mild depression”; a score of 10–14 as “moderate depression”; a score of 15–19 as “moderately severe depression”; and a score of ≥20 as “severe depression” [41]. Cronbach's alpha was 0.799.

• Part 5. Sleep quality was assessed using the English version of PSQI [34] translated into Myanmar by the researcher. PSQI comprises 19 self-rated questions that generate seven component scores: (1) subjective sleep quality; (2) sleep latency; (3) sleep duration; (4) habitual sleep efficiency; (5) sleep disturbances; (6) use of sleeping medication; and (7) daytime dysfunction. Each component scores 0–3 points. A score of “0” is “no difficulty,” while a score of “3” is “severe difficulty.” The total scores of seven components were added to yield one “global score,” which ranged from 0 to 21 points. Higher scores indicated poorer sleep quality. A cutoff point of >5 was defined as “poor sleep quality” and ≤5 as “good sleep quality” [42]. Cronbach's alpha was 0.741.

Data were collected by the principal investigator and two trained research assistants. Interviews were conducted at the outpatient department of a private hospital and the interviewer sat down with the participant in a comfortable place with appropriate privacy for the participant. Selected relevant laboratory results were collected from the patients' medical records. After the interview, the questionnaires were checked and verified for completeness and kept securely by the principal investigator.

Data analysis was performed using SPSS version 18.0 (SPSS Inc. Released 2008. SPSS Statistics for Windows, Version 18.0. Chicago: SPSS Inc.). Descriptive statistics such as mean, standard deviation, frequency and percentage were used to describe all study variables. All study factors with a p-value of <0.05 in crude analysis and biological plausibility were tested for multicollinearity [43] before conducting the multivariate analysis. Multiple logistic regression was used to obtain the odds ratio (OR) and 95% CI to determine the association between study factors and sleep quality of T2DM patients. The significance level was set at a p-value < 0.05.

The research proposal was approved by the Human Research Ethical Review Committee of the Faculty of Public Health, Mahidol University (COA. No. MUPH 2018-51; March 13, 2018) and the Ethics Review Committee on Medical Research Involving Human Subjects, Department of Medical Research, Ministry of Health and Sports, Myanmar (Approval No. Ethics/DMR/2018/045; March 21, 2018).

Of the 289 T2DM patients, 77.9% were aged ≥50 years. The mean ± SD age was 59.1 ± 11.4 years. Ages ranged from 30 to 80 years. 68.5% were female. 71.7% finished high school or lower. 57.1% had a monthly family income ≤300,000 kyat (223 USD). 65.1% were married. 42.2% were still working. 96.9% were living with a spouse, children or other relatives. 9.7% were current smokers and 9.7% were current drinkers. Regarding BMI, 64.4% were obese while 17.3% were in the normal range. 66.8% did not exercise and 64.4% were accustomed to daytime napping (Table 1).

Table 2 shows that 42.2% of patients had been diagnosed with diabetes for six years or more. 64.0% reported no complications. Among patients with complications, peripheral neuropathy was found in the highest proportion of 22.2%, followed by nephropathy at 10.4%. 70.6% had poor glycemic control, which was HbA1c level >7%. Only 21.1% had no comorbidities. Among the comorbid diseases, hypertension stood as the highest at 61.6%, while hyperlipidemia was the second highest at 42.2%. Regarding medication, the largest group, 45.0%, were treated with a two-drug combination of oral hypoglycemic agents; this was followed by 32.9% with a three-drug combination of oral hypoglycemic agents.

Concerning depression status, 72.3% of 289 T2DM patients had “normal” scores while 19.4% had mild depressive symptoms, 5.2% had moderate depression and 3.1% had moderate–severe depression. Regarding family relationships, 91.7% had good family relationships (Table 3).

The T2DM patients usually went to bed on average at 10.00 pm and rose at 5.30 am. The mean ± SD of sleep latency was 32.17 ± 2.59 min. The mean ± SD of sleep duration per night was 6.48 ± 1.36 hours. Table 4 shows the number and proportion of seven components of sleep quality among 289 T2DM patients. Roughly, 14.2% had poor subjective sleep quality; 59.9% had prolonged sleep latency of >15 min; 22.2% had poor sleep duration of <6 h; 29.4% had low but appropriate sleep duration of 6–7 h; 36.7% had a poor sleep efficiency of <than 85.0%; 26.3% had high sleep disturbance of ≥1 time per week; 14.9% used sleep medication of <1 time a week or ≥1 time per week and 36.3% had daytime dysfunction of <1 time a week or ≥1 time per week. The total prevalence of poor sleep quality (PSQI global score >5) among T2DM patients was 48.4% (140/289), 42.9% (39/91) for men and 51.0% (101/198) for women.

The mean ± SD of the PSQI global score was 5.97 ± 3.45. The mean ± SD of the PSQI global score among T2DM patients with at least one complication, with depression or with poor family relationships, was higher than that of participants without complications (6.93 ± 3.46 vs 5.43 ± 3.33 for complication status, 8.53 ± 3.63 vs 5.00 ± 2.82 for depression and 8.87 ± 3.18 vs 5.71 ± 3.35 for family relationships) Table 5.

In the multiple logistic regression analysis, all significant factors from the crude analysis and biological plausibility were simultaneously entered in the final model. In model 1, the significant factors were older age (AOR = 2.06; 95% CI = 1.13–3.75), having completed ≤ high school as the highest education level (AOR = 2.59; 95% CI = 1.32–5.08) and glycemic control (AOR = 1.94; 95% CI = 1.13–3.33). In model 2, the significant factors were the presence of complications (AOR = 1.86, 95% CI = 1.04–3.35), poor family relationships (AOR = 5.09, 95% CI = 1.55–16.68) and depression (AOR = 7.52, 95% CI = 3.38–14.76), as depicted in Table 6.

This study revealed the prevalence of poor sleep quality among T2DM patients at 48.4%. This finding was similar to the studies conducted by Zhu et al. [18] – (47.1%) with T2DM patients who attended a university hospital in China – and Song et al. [19] – (49.0%) with T2DM patients using insulin therapy who attended a diabetes clinic or outpatient center at the Endocrinology Department, Zhejiang Provincial People's Hospital, Hangzhou, China. The prevalence of poor sleep quality in this current study falls in the range of previous epidemiologic studies [9, 15, 18–22], which ranges from 45 to 70%. The difference in the prevalence of poor sleep quality may be due to the sample size, whether the study site is in the community or hospital, socioeconomic status differences and lifestyle characteristics. However, this study's findings show that the prevalence of poor sleep quality among T2DM patients was much higher than among the general adult population where it is approximately 29% [3].

In the subgroup analysis, when the proportions of poor sleep quality components or items among the 140 poor sleep quality patients were assessed, the three most frequent complaints were prolonged sleep latency (80.7%), poor habitual sleep efficiency (68.6%) and daytime dysfunction (61.4%) (data not shown). This finding is in line with previous studies as discussed by Sakamoto et al. [10] and Skomro et al. [46].

In model 1 of the multivariate analysis, age, highest education level and glycemic control were significant predictors of sleep quality of T2DM patients although glycemic control attenuated the association between age, education and sleep quality. In model 2, only complication status, poor family relationships and depression were significant predictors of sleep quality when the duration of diabetes and those three factors were included in the final model. A possible explanation is that complication status, family relationship and depression are particularly important factors that affect sleep quality. These factors had a higher coefficient in the model for their potential to affect other variables (data not shown).

In this study, the presence of complications was a significant predictor of sleep quality. This finding is congruent with previous studies [22, 47, 48]. Of note, 36% of T2DM patients in this study reported the presence of complications including peripheral neuropathy (22.2%), nephropathy (10.4%), cardiovascular (8.0%) and retinopathy (4.8%). These complications caused nighttime awakenings and led to the development of sleep disorders, especially in older diabetes patients [47–49].

Poor family relationships were another important predictor of sleep quality in this study. Currently, there is little research into how family relationships affect the sleep quality of T2DM patients, but this finding is consistent with those of the studies by Ailshire and Burgard [31] and Thichumpa et al. [33], who found that a supportive family relationship was associated with less troubled sleep, whereas having strained family relationships was associated with more troubled sleep.

About 91.7% of T2DM patients had good family relationships, whereas 8.3% had poor family relationships. The most frequently reported problems of family relationships included never finding good compromises among family members (54.2%); never having the ability to share secrets within family (54.2%); never having the ability to share ideas among family members (41.7%); never having the ability to admonish one another (41.7%) (data not shown). Given the strong relationship between the quality of family relationships and the quality of sleep, healthcare providers should provide interventions that promote good relationships among family members. Good family relationships can be a buffer against loneliness and depression that negatively affect sleep quality [32].

Depression was the most important predictor of sleep quality. This might be due to the prevalence of depression being twice as high in diabetes patients than in the general population [50]. In addition, complications due to advanced T2DM may cause severe depression in response to the impacts of these disease complications [48, 50]. In this current study, 27.7% had depression. In a subgroup analysis, 24 T2DM patients with depression also had poor family relationships. The combined effects of depression and poor family relationships caused this group of T2DM patients to have poor sleep quality. Finding an association between depression and poor quality is consistent with the previous studies as discussed elsewhere [13, 15, 30].

In this study, age, sex and BMI were not significantly associated with sleep quality. This is in contrast to the findings of a study conducted by Darraj et al. [22] that showed that being elderly, female and having an abnormal BMI were significantly associated with poor sleep quality. A possible reason why this current study did not find these factors to be significant is that the majority of the participants were elderly women with abnormal BMI, meaning age, sex and BMI differences could not be fully assessed.

Some limitations should be considered when interpreting these findings. Firstly, sleep quality information relied on the patient's self-reported data. The results should be confirmed using polysomnography or actigraphy tests, which give an objective assessment of sleep quality. However, the use of these tests is limited by the need for well-trained staff and higher costs [18]. In addition, PHQ-9 is a screening test for depression, the screening result should be confirmed by clinical interview and observation by a psychiatrist. Secondly, data were collected from a single private hospital, the results could not be generalized to T2DM patients in public hospitals or other study areas. Thirdly, the cross-sectional study design limited the investigation of the causality relationship between the significant predictors and poor sleep quality. Finally, some patients might have poor relationships with nonfamily members, such as neighbors or friends that could also be a cause of their poor sleep quality. This reason was not assessed in the current study.

Based on the findings of this study, healthcare providers and hospital administrators should provide regular screening for sleep quality and depression and provide sleep health education at diabetic clinics. Healthcare personnel should focus on complication status, family relationships and depression status of T2DM patients and provide counseling programs at diabetic clinics. Nurses should be trained to provide consultations in order to improve family relationships and provide psychosocial support to relieve depression. Also, the prevalence of obstructive sleep apnea is quite high among diabetes patients and could contribute to poor sleep quality [12]. Further study is needed within this population to thoroughly describe the factors related to the reported sleep-related issues with coughing, intense snoring and breathing difficulties.

The prevalence of poor sleep quality among adults with T2DM was rather high. Further study should be conducted in Myanmar with a greater sample size and larger sample area covering a wider range of social and cultural contexts to improve the accuracy of estimated prevalence and predictors of poor sleep quality among T2DM patients in Myanmar.

The study was partially supported for publication by Faculty of Public Health, Mahidol University, Bangkok, Thailand.