“How Do Greeks Sleep?” A Cross-Sectional Study among Indigenous and Minority Populations

Abstract

Objective: To investigate sleep habits and their potential relationship with several sociodemographic, lifestyle and health related characteristics among indigenous and minority populations in Northeastern Greece.

Materials and methods: Nine hundred fifty seven adults aged 19 to 86 years were enrolled in this cross-sectional study. Self-reported structured questionnaires were utilized.

Results:The reported mean sleep duration on a weekly basis was 6:26±1:10 hours (range, 04:00 to 10:00 hours); sleep duration was 26 min longer on weekends (p < 0.001). In multivariate linear regression analysis, older age (β=-26.7 min, p=0.010), being divorced or widowed (β=-29.0 min, p < 0.001), high alcohol (β=-39.7 min, p < 0.001) or coffee (β=-36.9 min, p=0.006) consumption, screen exposure before bedtime for 1-2 hours (β=-18.9 min, p=0.004) or > one hour (β=-34.4 min, p < 0.001), having a child aged under six years (β=-62.3 min, p < 0.001), napping for > 30 min during the day (β=-35.2 min, p < 0.001) and morbidity (β=-21.5 min, p < 0.001) were independently associated with short sleep duration and lower sleep efficiency. Moreover, a tendency towards short sleep duration was associated with anxiety (β=-8.8 min, p=0.078) and depression (β=-12.8 min, p=0.029). Obesity (β=10.7 min, p=0.047), being a university student (β=41.0 min, p=0.002), high financial status (β=16.6 min, p=0.037) and high adherence to Mediterranean diet (β=15.4 min, p=0.002) were associated with long sleep duration.

Conclusion:This study illustrates the association of sleep disturbances with several sociodemographic and health-related factors and dictates conduction of larger scale prospective studies to evaluate causality on the relationship between sleep patterns and lifestyle factors.

Keywords:sleep duration, sleep efficiency, Greeks.

INTRODUCTION

Sleep deprivation constitutes a major public health concern affecting modern societies, as nowadays both sleep quantity and quality are substantially disturbed, resulting in a decline of the quality of life (1, 2). Sleep inefficiency may confine daytime concentration, vigilance and coordination and provoke irritability, fatigue and malaise (3). Moreover, poor sleep may be associated with adverse health outcomes such as anxiety (4), depression (5) hyperlipidemia (6), diabetes mellitus (7), hypertension (8) and cardiovascular disease (9).

The prevalence of sleep problems among adults varies across the world. Kocevska et al have studied sleep characteristics in 1.1 million people from the United States, the Netherlands and the UK, and concluded that insomnia was more prevalent among Americans compared to Hollanders and Britons (10). Furthermore, Kerkhof revealed a high rate of both sleep disturbances and sleep inefficiency among Hollanders aged 18 to 70 (11). Moreover, Blume et al demonstrated that only 31% of Austrians assigned themselves as "good sleepers" and 86% reported sleep problems lasting more than six months (12). Additionally, according to Lee and Sibley, more than a third of New Zealanders are experiencing sleep problems such as short sleep duration (13). Finally, Magee et al underlined that both short and long sleep duration were strongly related to poor general health and lower quality of life in Australian adults between 45 and 95 years old (14).

Data from a number of cohort studies indicate large differences in sleep duration and time in bed regarding sex, age, body mass index (BMI), marital status and other socioeconomic parameters in different countries. For example, Leng et al having studied a population of Britons aged 45-90 years highlighted that sleep duration was lower among non-workers, women and the elderly (15). In addition, Ursin et al concluded that 10% and 12.2% of Norwegian men and women respectively suffered from insomnia (16). Finally, Wilsmore et al pointed out that lifestyle parameters such us alcohol consumption, depression or shift-working among New Zealanders aged 16-84 were in correlation with insomnia and day time sleepiness (17).

Data regarding sleep habits among Greeks remain scarce (18) and are mainly derived from studies on children (19) and adolescents (20). In the present study, self-reported questionnaires were utilized in order to include a large number of participants and to collect data for different sleep characteristics. The aim of this large cross-sectional study was to describe several sleep characteristics such as bed time, rise time, time in bed, sleep duration and sleep efficiency in relation to several sociodemographic characteristics, lifestyle habits and health related characteristics of the adult population of northern-eastern Greece.

METHOD

Study sample and research design

The study population in this cross-sectional study consisted of 957 participants [439 (45.9%) males and 518 (54.1%) females] with a mean age of 49.62±14.79 years (range 19-86 years; median age 50 years). The sample selection was based on a two-stage stratified sampling scheme on all adult people (aged ≥18 years) living in the region of Thrace and was conducted between September 2016 and May 2019. Thrace, the Northeastern prefecture of Greece, is characterized by cultural diversity with various national, ethnolinguistic and religious groups. Its population consists of the indigenous Christian Orthodox population (65% of the region population); the Muslim minority, which is the dominant minority group (30% of the population in Thrace), including the Pomaks and the Roma-Gypsies; and the descendants of Armenian refugees and a lot of expatriated Greeks from countries of the former Soviet Republics who settled in Thrace (estimated 5% of the population of Thrace). In the first stage of the sampling procedure, the area of Thrace was divided into two strata by the degree of urbanization. The urbanization levels were urban (≥10,000 inhabitants) and semi-urban or rural (<10,000 inhabitants) areas. According to the 2011 census, which constituted the sampling frame in our study, the urban population of Thrace accounted for approximately 40% of the total population of this area. In the second stage, subjects were recruited proportionally to each stratum size through a method of random generation of telephone numbers on the basis of the area code. After the aim of the study was explained to all participants, they agreed to have field researchers visiting their home and to complete the study questionnaires in an hour-long interview. The overall response rate was 71%. The following persons were excluded from this study: people aged <18 years; pregnant women, due to the special characteristics of their sleeping habits; night shift workers, owing to the inconsistency in their sleep-wake pattern; persons living in institutions for chronically ill people; inmates of nursing homes and penitentiary establishments.

Ethics

All procedures performed in the present study were in accordance with the ethical standards of the Democritus University Ethics Committee, which approved its conduct (No 42570/294), and with the standards of the Helsinki declaration (1964) and its later amendments. Informed consent was obtained from all study participants.

Covariates

A structured questionnaire was used to collect: a) sociodemographic characteristics (gender, age, place of residence, education level, marital, cultural, employment and financial status); b) health related characteristics [subjective general health status, chronic disease morbidity, anxiety (21), depression (22, 23) and body mass index (24)]; and c) lifestyle and dietary habits [smoking status, alcohol consumption, coffee consumption, caffeine consumption in the evening, adherence to the Mediterranean diet (25), time from dinner to bedtime, time watching TV or using a computer before bedtime, physical activity, the presence of a child aged under six years and nap during the day]. (Supplementary material) Description of the variables has been also presented in Zissimopoulou et al (26).

Measures of sleep

Participants provided information on their nighttime sleep by answering the following sleep questions of the questionnaire: "At what time do you normally go to bed?", "At what time do you normally get up?" and "On the average, how many hours do you sleep per day?". (Supplementary material) Responses were obtained for an average weekday and weekend day over the previous month. Time in bed was calculated as the difference between rise time and bedtime. As an indicator of the overall time in bed and sleep duration on a weekly basis, weighted mean measures were calculated using the following formulas: weighted time in bed = 5/7*(time in bed on a weekday) + 2/7*(time in bed on a weekend day) and weighted sleep duration = 5/7* (sleep duration on a weekday) + 2/7*(sleep duration on a weekend day). Sleep efficiency was assessed as the ratio of nighttime sleep duration and time in bed.

Statistical analysis

Statistical analysis of data was performed using IBM Statistical Package for the Social Sciences (SPSS), version 19.0 (IBM Corp., Armonk, NY, USA). The normality of quantitative variables was tested with Kolmogorov-Smirnov test. Quantitative variables were expressed as mean ± stan- dard deviation (SD) and qualitative variables were expressed as absolute and relative (%) frequencies. In particular, the mean estimated time of sleep characteristics (i.e., bedtime, rise time, time in bed and sleep duration) were expressed as HH: MM. In the univariate analysis, the association of subjects’ characteristics with time in bed, sleep duration and sleep efficiency was assessed using Student’s t test and analysis of variance (ANOVA). Three separated multivariate linear regression models were constructed to explore the independent association of subjects’ sociodemographic characteristics, lifestyle habits and health related characteristics on time in bed sleep duration and sleep efficiency. Unstandardized b regression coefficients and their standard error (SE) were estimated as the measure of the above associations. All tests were two tailed and statistical significance was considered for p values < 0.05.

RESULTS

Subjects’ characteristics

Participants’ sociodemographic, health related and lifestyle characteristics are shown in Tables 1 and 2. One third of the entire cohort (32.7%) was of low educational level and half of the subjects (49.7%) belonged to the low financial class. The prevalence of smoking was 34.8%. Half of all subjects (49.7%) used to consume at least one glass of alcohol per week, while the vast majority (91.2%) used to consume at least one cup of coffee per day. Only 22.4% achieved high adherence to the Mediterranean diet, 15.9% were physically active. The majority (58.2%) used to take a nap during the day. Regarding subjects’ health related characteristics, the prevalence of obesity was 37.3%. Morbidity rate was 56.5%; however, the majority (77.0%) of subjects reported a good subjective health status. The internal consistency of Zung self-rating Anxiety Scale and Beck Depression Inventory (BDI), which were used to assess the level of anxiety and depression disorders, respectively, were very high (Cronbach’s á coefficient was 0.81 and 0.89, respectively). The prevalence of anxiety and depression disorders was 33.6% and 28.4%, respectively.

Bedtime and rise time

On weekdays, bedtime ranged from 17:00 pm to 03:00 am with a mean value of 23:29 pm and rise time ranged from 04:30 am to 11:00 am with a mean value of 06:53 am. Older (aged > 70 years) subjects (22:35 pm), those with low education level (22:56 pm) and Greek Muslims (22:59 pm) used to go to bed earlier; on the contrary, younger (aged < 30 years) subjects (00:14 am), residents of urban areas (23:52 pm) and those with high education (00:02 pm) and financial (23:52 pm) level used to go to bed later. Moreover, older (aged > 70 years) subjects (06:28 am), those having a child aged under six years (06:40 am), individuals with chronic disease (06:43 am), anxiety (06:43 am) and employees (06:44 am) used to wake up earlier, while subjects following a Mediterranean diet (07:13 am), those who consumed > four cups of coffee per day (07:32 am), younger (<30-year-old) subjects (07:39 am) and unemployed (07:07 am) or university students (09:23 am) used to wake up later.

On weekends, the mean bedtime was 23:55 pm (range, 17:00 pm to 04:00 am; 26 min later than the weekdays, p < 0.001) and the mean rise time 07:45 am (range, 4:30 pm to 12:30 am; 52 min later than the weekdays, p < 0.001).

Time in bed, sleep duration and sleep efficiency

Time in bed was calculated as the difference between the reported bedtime and rise time. Mean time in bed was 07:24±01:05 hours on weekdays (range, 05:00 to 11:00 hours) and 26 min longer (p < 0.001) on weekends (07:50±01:00 hours). The mean self-reported sleep duration was 6:19±1:10 hours on the weekdays (range, 04:00 to 10:00 hours) and 6:45±1:15 hours on the weekends; sleep duration was statistically significantly longer on weekends (by 26 min, p < 0.001). On a weekly basis, the mean time in bed was 07:32±01:00 hours and the mean sleep duration was 6:26±1:10 hours. It was observed that the time in bed was more than one hour longer than sleep duration (p < 0.001).

Time in bed, sleep duration and sleep efficiency in relation to all subjects’ characteristics are shown in Tables 1 and 2. Shorter time in bed was found among subjects of high education and financial level and among employed individuals, while longer time in bed was found among older (>70-year-old) subjects, those with low education level, Greek Muslims and expatriated Greeks, household and university students. Regarding the association of gender and age with sleep duration, no gender-related difference was observed (p=0.191), while sleep duration was decreasing as age increased p < 0.001). However, it was observed that females used to sleep more than males in the first three age groups (< 30-year-old: p=0.071; 31-40 years: p < 0.001; 41-50-year-old: p=0.048), while males used to sleep more than females in the last three age groups (51-60-year-old: p=0.010; 61-70-year-old: p=0.074; >70 years: p=0.008); overall, sleep duration was longer in females aged ≤ 50 years (06:50 vs 06:26 hours, p < 0.001), while sleep duration was longer in males aged > 50 years (06:20 vs 06:00 hours, p=0.006). Furthermore, shorter sleep duration was found among older subjects (aged > 70 years), those being divorced or widowed, those who consumed more than six glasses of alcohol/week and those having a child aged under six years, while longer sleep duration was found among younger subjects (aged < 30 years), singles, household, university students and those who watched TV or used a computer for < one hour before bedtime.

Sleep efficiency on a weekly basis ranged from 50% to 100% with a mean value of 86±11%. Sleep efficiency less than 85%, 80% and 75% was observed in 34%, 23% and 17% of participants, respectively. Sleep efficiency was similar on weekdays (85%) and weekends (86%). Worse sleep efficiency was found among household, older subjects (aged > 70 years) subjects, divorced or widowed individuals, subjects with bad subjective health status and those who consumed more than four cups of coffee/day or more than six glasses of alcohol/week; on the contrary, better sleep efficiency was found in subjects of high education and financial level, those having higher physical activity and university students.

Independent determinants of time in bed

Multivariate linear regression analysis revealed the following significant independent determinants for time in bed (which explain the 34.5% of its variation): (i) shorter time in bed was associated with middle age (p=0.026 for 31-40-year-old persons; p < 0.001 for 41-50-year-old persons; p < 0.001 for 51-60-year-old persons; p=0.034 for 61-70-year-old persons compared to those aged < 30 years), being divorced or widowed (p=0.009), caffeine consumption in the evening (p < 0.001), having dinner 1-2 hours (p=0.006) and > two hours (p < 0.001) before bedtime, watching TV or using a computer for more than two hours before bedtime (p < 0.001), having a child aged under six years (p=0.002), napping for > 30 min (p < 0.001) and having good subjective health status (p=0.033); (ii) longer time in bed was associated with being Greek Muslim (p=0.001) or expatriated Greek (p=0.006) compared to Greek Christians, university student (p=0.010), household (p < 0.001), ex- (p=0.006) and current (p=0.051) smoking, high adherence to the Mediterranean diet (p=0.002) and obesity (p=0.001) (Tables 3 and 4).

Independent determinants of sleep duration

Multivariate linear regression analysis revealed the following significant independent determinants for sleep duration (which explain the 34.1% of its variation): (i) shorter sleep duration was associated with older age (p=0.010), being divorced or widowed (p < 0.001), high alcohol consumption (> six glasses/week) (p < 0.001), high coffee consumption (> four cups/day) (p=0.006), watching TV or using a computer before bedtime for 1-2 hours (p=0.004) or for > two hours (p < 0.001), child aged under six years (p < 0.001), napping for >30 min (p < 0.001), morbidity (p < 0.001), anxiety (p=0.078) and depression (p=0.029); (ii) longer sleep duration was associated with Greek Muslims (p=0.001), being unemployed (p=0.081), university student (p=0.002), retired (p=0.081), household (p < 0.001), having high financial status (p=0.037), ex- (p=0.064) and current (p=0.040) smoking, high adherence to the Mediterranean diet (p=0.002), having dinner 1-2 hours (p=0.011) and > two hours (p=0.001) before bedtime and obesity (p=0.047) (Tables 3 and 4).

Independent determinants of sleep efficiency

Multivariate linear regression analysis revealed the following significant independent determinants for sleep efficiency (which explain the 41.9% of its variation): (i) worse sleep efficiency was associated with older age (p=0.049), being divorced or widowed (p=0.001), expatriated Greeks (p=0.050), high alcohol consumption (> six glasses/week) (p < 0.001), high coffee consumption (> four cups/day) (p < 0.001), watching TV or using a computer before bedtime for 1-2 hours (p=0.007) or for > two hours (p=0.001), having a child aged under six years (p < 0.001), napping for >30 min during the day (p-0.047) and morbidity (p < 0.001); (ii) better sleep efficiency was associated with age categories 31-40 years (p < 0.001) and 41-50 years (p=0.001), singles (p=0.003), high education level (p < 0.001), being unemployed (p=0.056), retired (p=0.081) and household (p=0.005) compared to being employed, having high financial status p=0.037), caffeine consumption in the evening (> 6 pm) (p < 0.001), high adherence to the Mediterranean diet (p=0.009), having dinner 1-2 hours (p<0.001) and > two hours (p < 0.001) before bedtime and having good subjective health status (p=0.029) (Tables 3 and 4).

DISCUSSION

This cross-sectional study was conducted to assess the potential association of adults’ sleep patterns, which include time in bed, sleep duration and sleep efficiency with sociodemographic characteristics, utilizing a populationbased sample from the rural region of Thrace, in Northeastern Greece.

Concerning time in bed, shorter time was more prevalent among middle aged (41-60), divorced or widowed subjects and parents of minors that consumed caffeine after 6 pm, had a nap lasting more than 30 minutes during the day and were exposed to TV or computer screens for > two hours before bedtime. On the other hand, longer time in bed was associated with high adherence to Mediterranean diet, tobacco use, obesity and being Greek Muslim or expatriated Greek or a university student. In keeping with our results, Watson et al exhibited that lower caffeine consumption was related to longer time in bed (27). Similarly, Farazdaq et al also concluded that divorced or widowed individuals spend shorter time in bed (28). In contrast, results from existing literature regarding the relationship between time in bed, obesity and having young children differ from ours as in the studies of Gubelmann et al (29) and Hart et al (30), obesity was associated with shorter time in bed and according to Sekine et al (31) having young kids presented a u-shaped correlation with time in bed, respectively. Additionally, in the studies of Thomas et al (32) and Leng et al (15), older participants spent longer time in bed than the younger ones, whereas in our study being middle- aged (41-60-year-old) was associated with a shorter time in bed.

With regards to sleep duration, the present study demonstrated that shorter sleep duration was significantly related to depression, whereas a meta-analysis from Zhai et al concluded that both short and long sleep duration were associated with high risk of adults’ depression (33). In keeping with our results, Liu et al also reported that long sleep duration predisposed to obesity in adults (34). These findings contradicted the results of two recent meta-analyses that presented a significant relation between short sleep duration and risk of obesity in adults (35, 36). Concerning caffeine consumption, we pointed out a correlation between high caffeine (more than four cups/day) consumption and short sleep duration. Despite the fact that acute effects of caffeine on sleep were widely accepted, caffeine consumption in most of the literature did not relate to sleep duration (37, 38). Apart from that, we also pointed out a correlation between high alcohol consumption (more than six glasses per week) and short sleep duration. These results are consistent with the study of Chaput et al which highlighted that short-time sleepers were consuming more alcohol (39). Futhermore, Castro- Diehl et al have proven that a high adherence to Mediterranean diet was correlated with an adequate sleep duration, which was in accordance with our results that demonstrated a relationship between adherence to Mediterranean diet and longer sleep duration (40).

Our data concerning sleep efficiency were in accordance with the current literature as lower sleep efficiency was associated with advanced age (41), increased coffee (27) and alcohol consumption (42) and morbidity (43). Additionally, the present study has noted a statistical significance for the relation between poor sleep efficiency and being divorced or widowed. Arber et al also highlighted that divorced or widowed responders subjectively reported more sleep problems compared to married ones (44). The aforementioned study has also showed that the association between sleep disturbances and a low educational level was significant, a result that comes to an agreement with ours in view of the fact that the high educational level was related to better sleep efficiency. In the study of Barros et al, the prevalence of poor self-rated sleep was found higher in those aged 40 years or older and in unemployed individuals (45). Our study yielded contrary results, as sleep efficiency was better in those without occupation and in middle age individuals (31 to 50-year-old). Regarding the relationship between sleep efficiency and sex, some studies have found that women report worse sleep quality than men (41, 45). Although in our cohort females aged . 50 years had a longer sleep duration than males of corresponding age, gender-related differences on sleep efficiency were not observed.

Interestingly, and in contrast to the general view that university students had an inadequate sleep (46), our study revealed that being a university student was associated with a longer sleep duration, longer time in bed and better sleep efficiency. Cellini et al utilized actigraphy to reveal a normal time in bed and poor sleep efficiency among Italian university students (47), whereas Sivertsen et al used self-reported questionnaires that revealed a shorter time in bed and poor sleep quality among Norwegian university students (48). Such discrepancies could be attributed to the subjective nature of self-reported measures compared to actigraphy or the variation in sleep patterns among different geographical locations.

Overall, our study has several strengths, which include data from a large broad-based sample of a regional Thrace population. In addition, the present analysis provides a precise assessment of sleep quality and quantity concerning sociodemographic features. The sample is of high participation rate that ensures the randomness of the selected general population in this area. However, this study has some limitations. Above all, it was conducted in a rural population with religious, national and social perceptions of great variety. Despite the fact that these differences should not have impact on the population associations of sociodemographic factors, the findings might not be generalizable to other adult populations. Moreover, our study was susceptible to recall bias, as it was based on subjective measures that did not include polysomnography or actigraphy that make for objective measures. Furthermore, sleep duration, sleep quality and time in bed were all self-reported by study participants and thereby, misreporting could not be ruled out. This is more evident if we take into account previous research, demonstrating that self-reported sleep measures have a tendency to proffer an overestimation of sleep disturbances (49). It is important to mention that another limitation of this study concerns the seasonal variability, which may severely affect the wake time and sleep duration. Our data were not adjusted correspondently. Limitations also lie on the properties of cross-sectional design, since causal relationships between sleep patterns and sociodemographic and socioeconomic factors could not be established.

CONCLUSION

Our research revealed a relationship between several sociodemographic characteristics and sleep parameters such as time in bed, sleep duration and sleep efficiency among Northeastern Greeks. To summarize, short time in bed, short sleep duration and poor sleep efficiency were associated with high caffeine consumption, individuals with children under six years of age and high alcohol consumption, whereas long time in bed, long sleep duration and adequate sleep efficiency were correlated with high adherence to Mediterranean diet and being a university student. Sleep efficiency requires better definition and elucidation to assist in incorporation of its clinical implications in pertinent studies. Considering both the strengths and limitations of the present study, it highlights the necessity to conduct larger scale prospective studies in order to define causality on the relationships between sleep characteristics and sociodemographic factors.

Conflicts of interest: none declared.

Financial support: none declared.

TABLE 1.

Sleep characteristics in relation to subjects’ demographic and health related characteristics

TABLE 2.

Sleep characteristics in relation to subjects’ lifestyle habits

TABLE 3.

Results of multivariate linear regression analysis for the impact of subjects’ demographic and health related characteristics on their sleep characteristics. Data are expressed as unstandardized beta (β) coefficients with their standard error (SE)

TABLE 4.

Results of multivariate linear regression analysis for the impact of subjects’ lifestyle habits on their sleep characteristics. Data are expressed as unstandardized beta (β) coefficients with their standard error (SE)

SUPPLEMENTARY MATERIAL.

SUPPLEMENTARY MATERIAL

SUPPLEMENTARY MATERIAL.

SUPPLEMENTARY MATERIAL