Physical Neighborhood and Social Environment, Beliefs About Sleep, Sleep Hygiene behaviors, and Sleep Quality Among African Americans

Abstract

Objectives

African Americans (AAs) have a higher prevalence of sleep disorders than other racial/ethnic groups. However, little is known about the relationships among individual and neighborhood factors related to sleep quality in AAs. The purposes of this study were to (1) describe beliefs about sleep, sleep hygiene behaviors, and sleep quality among AAs; and (2) examine the relationships among sociodemographic characteristics, neighborhood environment, beliefs about sleep, sleep hygiene behaviors, and sleep quality.

Methods

We conducted a cross-sectional study of 252 AA men and women in the Greater New Haven, CT, USA community. We assessed their sociodemographic characteristics, neighborhood environment, beliefs about sleep, sleep hygiene, and sleep quality with the following measures, respectively: the Neighborhood Environment Scale, the brief version of Dysfunctional Beliefs and Attitudes about Sleep, the Sleep Hygiene Practice Scale, the Pittsburgh Sleep Quality Index. We performed descriptive statistics, correlations and multiple hierarchical regression.

Results

About 72% of the participants (mean age: 53.88±14.17 years, 77.8% women) reported experiencing sleep disturbance. People with poor sleep quality were more likely to report poorer neighborhood social environment (social cohesion), poorer overall neighborhood environment, more dysfunctional beliefs toward sleep, and poorer sleep hygiene than those who had good sleep quality. In the final multivariate model that controlled for a number of chronic comorbid conditions, neighborhood environment, beliefs about sleep, and sleep hygiene behaviors were significantly associated with sleep quality.

Conclusions

Future efforts are needed to improve sleep among AAs by considering both the individual’s belief about sleep, sleep hygiene behaviors and neighborhood factors.

INTRODUCTION

Poor sleep is linked to numerous physical and mental health outcomes including mood, cardiovascular and metabolic disorders.^1,2 African Americans have a disproportionately higher prevalence of sleep disorders than other racial/ethnic groups.² Poor sleep and sleep disparities may contribute to other health disparities in heart disease, obesity, and diabetes in African Americans.^1,2 Although the individual and social factors related to habitual sleep patterns have been increasingly studied in the general population, limited attention has been paid to habitual sleep patterns and the relationship between sleep-related cognition (beliefs about sleep), sleep behaviors (sleep hygiene practices), physical neighborhood and social environment and sleep quality among the high-risk group, African Americans.

Individuals’ beliefs about their conditions or illnesses influence their health behaviors; for example, individuals with positive attitudes and beliefs toward diabetes have better diabetes self-management and adherence to diabetes treatment.³ Similarly, rigidly held beliefs or unrealistic expectations about sleep are associated with clinically significant sleep disorders such as insomnia among both community participants and hospital patients.^4,5 Racial and ethnic groups tend to endorse different beliefs and attitudes about sleep;⁶ for example, in one study African American women were more likely to report misperceptions about sleep and beliefs that suggested less understanding of the importance of good sleep for health than their White counterparts.⁷ However, in a meta-analysis, the relationship between race and sleep was confounded with levels of education with much lower educational levels among the African American.^8,9

Individuals’ sleep-related beliefs and attitudes may contribute to their sleep behaviors. Sleep hygiene is defined as practicing behaviors that facilitate sleep and avoiding behaviors that interfere with sleep.¹⁰ Unhealthy sleep hygiene behaviors are associated with poor habitual sleep pattern and sleep quality.^11,12 Although limited data are available on African Americans’ sleep hygiene behaviors, one study showed that African American women are more likely to nap during the day to combat daytime sleepiness and are more likely to engage in activities other than sleep while in bed, including reading or watching TV, than white/European American women.⁷

Growing research also indicates that sleep is influenced not only by individual factors but also by one’s physical neighborhood and social environment. A neighborhood with excessive ambient noise, light, and a high crime rate can negatively influence sleep.^13,14 Characteristics of the neighborhood social environment, such as social cohesion and social support, could also affect sleep. In studies of the neighborhood social environment and sleep, lack of social cohesion was associated with short sleep duration, daytime sleepiness, and feeling unrested.^15,16 Disadvantaged physical neighborhoods and social environments are associated with short sleep duration, poor sleep quality, and sleep disorders.^17,18 In studies of multi-racial ethnic groups, the neighborhood disadvantage had a stronger influence on sleep among African Americans than among other racial/ethnic groups,¹³ even after controlling for individual income levels. However, most studies examining the relationships between sleep and neighborhood often omit other individual factors such as sleep cognitions (beliefs about sleep health) and sleep behaviors (sleep hygiene practices) that likely interact with each other. There is notably a lack of research describing individual sleep-related factors and neighborhood factors that influence sleep among African Americans, a population at high risk for sleep disorders and chronic conditions affected by poor sleep health.¹

Therefore, the purposes of this study were: (1) to describe beliefs about sleep, sleep hygiene behaviors and sleep quality among African Americans; and (2) to examine the relationships among individual factors (sociodemographic, beliefs about sleep, sleep hygiene behaviors), neighborhood environment and sleep quality.

METHODS

Design, participants, and procedures

This current study is a sub-study of a cross-sectional study of African American men and women designed to understand obesity, obesity-risk behaviors (sleep, diet and physical activity) and the relationships of these factors with social networks.

Participants were recruited from the Greater New Haven area in Connecticut through flyers posted at African Methodist Episcopal (AME) churches and the Yale Center for Clinical Investigation (Yale’s CTSA) website. Since the primary purpose of the parent study was to conduct sociometric social network analysis, we enrolled participants from three AME churches in the Greater New Haven area to examine church-based social networks. Eligibility criteria for participants included the following: (1) men or women over 21 years of age, (2) self-reported Black or African American, and (3) able to speak and read in English. We excluded individuals who reported disabilities or acute/terminal conditions that affect daily physical activity (e.g., terminal cancer, dialysis), active psychiatric illnesses such as thought disorders, or self-reported illegal drug use in the past 6 months. After signing written informed consent forms, all participants completed the surveys and anthropometric measurements for the primary study and received a $30 gift card. All study protocols were reviewed by the Yale University Institutional Review Board prior to study implementation.

Variables and Measures

Sociodemographic Characteristics and Chronic Comorbid Conditions

Sociodemographic data were collected by self-report on: age, gender, race/ethnicity, marital status, educational level, employment status, and annual household income. Chronic comorbid conditions were asked by using the following questions: “Have you ever had any of the following heart conditions (heart failure, heart attack, angina, stroke)?” “Have your health care professionals told you that you have kidney failure?” “Do you have diabetes?” “Do you have high blood pressure (hypertension)?” “Do you have arthritis?” “Do you have the following lung problems (asthma, chronic obstructive pulmonary disease)?” “Do you have depression?”

Anthropometric Characteristics

Body weight and height were measured three times using a portable electronic scale and stadiometer. Body mass index (BMI) was calculated as weight (kg)/height squared (m²). Percent body fat was estimated using the same digital scale that measures foot-to-foot bioelectric impedance. Waist circumference was taken at the narrowest part of the torso, at the end of a normal expiration. Hip circumference was taken around the buttocks in a horizontal plane at the level of maximal extension of the buttocks. The average of three measurements was calculated. Inter- and intra-observer reliability were checked.¹⁹

Neighborhood Environment

Perceptions about the neighborhood environment were measured by the Neighborhood Environment Scale,²⁰ with six neighborhood dimensions: aesthetic quality (5 items), walking environment (7 items), availability of healthy foods (3 items), safety (3 items), violence (4 items), and social cohesion (4 items). For most subscales, responses for each item ranged from 1 (strongly agree) to 5 (strongly disagree). Responses for the scales on violence ranged from 1 (often) to 4 (never). Subscale scores were estimated by taking the average across all items within each dimension; a higher score indicated a worse neighborhood environment. The Neighborhood Environment Scale has good internal reliability with the Cronbach’s alpha range of .73 to .83.²⁰ The Cronbach’s alpha for the scale in our study was .93.

Sleep Hygiene

The Sleep Hygiene Practice Scale (SHPS) was used to assess sleep hygiene awareness and practices.¹¹ The SHPS has 19 items which include sleep-related behaviors (napping, caffeine/alcohol intake), night time activities (exercise, phone conversation), and bedroom environment (noise, light, temperature and bed partner). Respondents report on the average number of days per week in which they engaged in these activities during the previous month. Frequency scores (number of days per week) were calculated for each item, and higher frequency scores indicated worse sleep hygiene practices. The total hygiene practice scores ranged from 0 to 133. The Cronbach’s alpha for the SHPS was reported as .72 in other study²¹ and our study, the Cronbach’s alpha was .62.

Dysfunctional Beliefs and Attitudes about Sleep

The Dysfunctional Beliefs and Attitudes about Sleep (DBAS) is a 16-item measure used to evaluate the following sleep-related cognitions:²² consequences of insomnia (5 items), worry about sleep (6 items), sleep expectations (2 items), and medication (3 items). Each item is rated on a 10-point scale ranging from 1 (strongly disagree) to 10 (strongly agree). For each statement, participants rated their level of agreement/disagreement by choosing “strongly disagree” (0) to “strongly agree” (10). The subscale score is calculated with the average score of items. A higher score indicates more dysfunctional beliefs and attitudes about sleep. Example items are: “When I sleep poorly on one night, I know it will disturb my sleep schedule for the whole week,” “Medication is probably the only solution to sleeplessness.” The DBAS showed good reliability with the Cronbach’s alpha of .79.²² The Cronbach’s alpha in this study was .86.

Habitual Sleep Patterns

The Pittsburgh Sleep Quality Index (PSQI) was used to measure sleep patterns and sleep quality.²³ The questionnaire measures seven components: sleep quality, sleep latency, sleep duration, habitual sleep efficiency (percentage of total time in bed spent in sleep), sleep disturbances, use of sleep medications, and daytime dysfunction. Each component has a range from 0 (no difficulty) to 3 (severe difficulty). The global score of PSQI ranges from 0 to 21, with higher scores indicating worse sleep quality. The PSQI has had strong test-retest reliability and, using a cutoff score of 5; the measure demonstrated a sensitivity of 89.6% and specificity of 86.5% in separating people with and without insomnia.^23,24 We categorized individuals with PSQI scores <5 as having good sleep quality. The global PSQI demonstrated adequate reliability with Cronbach’s alpha of .80.²⁵ The Cronbach’s alpha for the scale in this study was .72.

Statistical analysis

Descriptive statistics were conducted to summarize sample characteristics and study variables. Continuous variables were summarized as means and standard deviations, and categorical variables by frequencies and percentages. Pearson’s correlation test was conducted to assess the association between continuous variables; Spearman's rank correlation test was used for the correlation analysis between ordinal variables (e.g., education, employment status, household income). Bivariate analyses were conducted to examine differences in sociodemographics, clinical characteristics, neighborhood environment, sleep hygiene, and DBAS, using chi-square and t-tests between groups by sleep quality (PSQI scores <5; poor sleep quality vs good sleep quality). Hierarchical regression analysis was performed to examine the relationships among sleep-related cognition and behaviors (DBAS, sleep hygiene practice), neighborhood environment, and sleep quality. Sociodemographics and BMI were not entered into the final model because there were not any that were significantly associated with sleep quality in the bivariate analysis. Statistically significant changes in R² for each of the steps in the hierarchical regression analyses were calculated. Variance inflation factor (VIF) and its tolerance value (tolerance=1/VIF) were calculated to check for possible multicollinearity among independent variables. A VIF value less than 8 and a tolerance value greater than 0.2 has been reported to be acceptable.²⁶ The internal reliability of the instruments was examined using Cronbach's alpha. All hypothesis testing was 2-sided; type I error was controlled at the 0.05 significance level.

RESULTS

A total of 252 African American men and women participated in the study; 77.8% were female, and the mean age was 53.88 years (SD=14.17). About 33% were currently married, and 50% had a college degree or higher. About half of the participants were employed full-time, and 54% had an annual household income over $40,000. About 45.2% had cardiovascular disease (CVD) such as angina, heart attack, stroke or hypertension, 19.4% had type 2 diabetes, and 10.7% had depression. Mean BMI was 31.93 kg/m² (SD 6.48) and mean total body fat was 40.79% (SD 10.24). About 12% had normal BMI; 31% of participants were overweight (BMI 25-29.99 kg/m²); and 57.2 % met criteria for obesity (BMI ≥30 kg/m²) (Table 1).

Table 1.

Demographic and clinical characteristics of participants (N=252)

Category	Mean (SD) or N (%)	Min - Max
Age (years)	53.88 (14.17)
Gender
Man	56 (22.2%)
Woman	196 (77.8%)
Education level
Hight School Graduate	126 (50.00%)
College graduate	65 (25.79%)
Graduate school or higher	61 (24.21%)
Employment status
Working Full Time	123 (48.81%)
Working Part Time	30 (11.90%)
Unemployed	99 (39.29%)
Annual household income
0-$19,999	38 (15.1%)
$20,000-$39,999	49 (19.4%)
$40,000-$59,999	38 (15.1%)
$60,000-$79,999	32 (12.7%)
$80,000-$99,999	27 (10.7%)
More than $100,000	39 (15.5%)
Refused/Don't Know	29 (11.5%)
Clinical factors
Number of comorbidities	0.89 (0.88)	0.00 – 4.00
CVD	114 (45.20%)
- Hypertension	100 (39.70%)
- Angina	22 (8.70%)
- Heart attack	6 (2.40%)
- Heart failure	7 (2.80%)
- Stroke	3 (1.20%)
Chronic lung disease	34 (13.49%)
Diabetes	49 (19.40%)
Depression	27 (10.70%)
Body mass index (BMI) (kg/m2)	31.93 (6.48)	20.61 – 53.12
BMI 18.5–24.9 (normal)	30 (11.9%)
BMI 25.0–29.9 (overweight)	78 (31.0%)
BMI 30.0–34.9 (class 1 obesity)	73 (29.0%)
BMI 35.0–39.9 (class 2 obesity)	41 (16.2%)
BMI >40.0 (class 3 obesity)	30 (11.9%)
Waist-hip ratio	0.88 (0.09)	0.68 – 1.22
Body fat (%)	40.79 (10.24)	12.80 – 59.50

NOTE. Cardiovascular disease (CVD)

About 72% of the participants scored 5 or higher on the PSQI global score, indicating that they had experienced poor sleep quality in the past month. The mean sleep duration was 6.47 hours (SD 1.33), and the mean sleep efficiency was 84.13% (SD 15.09) (Table 2). About 52% of the participants reported a dysfunctional level of sleep beliefs and attitudes (mean DBAS score >3.8).^5,27

Table 2.

Descriptions of sleep-related variables and neighborhood environment (N=252)

Variable	Mean	SD	Min	Max	Possible range
1. Pittsburgh Sleep Quality Index (PSQI), global	6.82	3.62	1.00	18.00	0–21
1.1 Perceived sleep quality	2.62	1.91	0.00	9.00	0–9
1.2 Daily disturbance	2.18	1.07	0.00	5.00	0–6
1.3 Sleep efficiency	2.03	1.62	0.00	6.00	0–6
- Sleep latency (minutes)	26.75	22.29	0.00	120.00	0–1440
- Sleep duration (hours)	6.47	1.33	2.40	10.00	0–24
- Sleep efficiency (%)	84.13	15.09	28.24	100.00	0–100
2. Dysfunctional Beliefs and Attitudes Sleep	3.85	1.80	0.00	8.94	0–10
2.1. Consequences of insomnia	4.18	2.44	0.00	10.00	0–10
2.2. Worry about sleep	3.51	2.16	0.00	9.50	0–10
2.3. Sleep expectations	5.98	2.73	0.00	10.00	0–10
2.4. Medication	2.57	2.00	0.00	8.67	0–10
3. Sleep Hygiene	23.10	10.47	2.00	55.00	0–133
4. Neighborhood Environment	2.28	0.62	1.00	4.42	1–5
4.1. Aesthetic quality	2.22	0.83	1.00	5.00	1–5
4.2. Walking environment	2.35	0.71	1.00	4.43	1–5
4.3. Availability of healthy foods	2.77	1.07	1.00	5.00	1–5
4.4. Safety	2.59	1.04	1.00	5.00	1–5
4.5. Social cohesion	2.47	0.81	1.00	5.00	1–5
4.6. Violence	1.43	0.61	1.00	4.00	1–4

Educational levels were associated with sleep hygiene behaviors and beliefs about sleep. Individuals with higher education tend to have better sleep hygiene behaviors (r=−.15, p<.05) and more functional beliefs about sleep (the DBAS subscales, consequences of insomnia, worry about sleep, medication [r=−.13 to −.17, p<.01] and DBAS total score [r=−.17, p<.01]). A greater number of chronic comorbid conditions was associated with a higher global PSQI score (poorer sleep quality) (r=.13, p<.05). Worse sleep hygiene behaviors (r=.48, p<.01) and a higher summary score on the DBAS (dysfunctional beliefs about sleep) (r=.51, p<.01) were also associated with a higher global score of PSQI (poorer sleep quality).

The summary score of the DBAS was also associated with sleep hygiene behaviors (r=.43, p<.01), indicating that individuals with more functional sleep beliefs are more likely to perform good sleep hygiene practices. Table 3 presents bivariate analysis among sleep-related variables and the neighborhood environment subscales.

Table 3.

Correlations between neighborhood environment and sleep-related variables (n=252)

Variables	Neighborhood Environment
	Total score	Aesthetic quality	Walking environment	Availability of healthy foods	Safety	Social cohesion	Violence
Demographics
Age	−.07	−.13*	−.04	−.06	.13*	−.14*	−.05
Gendera	−.17*	−.14*	−.16*	−.09	−.12	−.10	−.16*
PSQI, global	.20**	.10	.22**	.10	.15*	.21**	.10
Perceived sleep quality	.16*	.07	.16*	.15*	.08	.17**	.09
Daily disturbance	.18**	.18**	.22**	.03	.09	.17**	.09
Sleep efficiency	.14*	.03	.15*	.03	.18**	.16*	.06
Sleep latency (minutes)	.12*	.06	.13*	.10	.07	.14*	.03
Sleep duration (hours)	−.12	−.00	−.16*	−.05	−.14*	−.15*	−.03
Sleep efficiency (%)	−.10	−.03	−.12	−.01	−.15*	−.09	−.03
DBAS total	.10	.08	.17**	.02	.01	.09	.02
Consequences of insomnia	.03	.05	.09	−.04	−.03	.02	−.04
Worry about sleep	.12	.09	.18**	.04	.02	.10	.04
Sleep expectations	−.02	−.03	.04	−.07	−.06	.03	−.02
Medication	.19**	.13*	.22**	.16*	.10	.16*	.08
Sleep hygiene	.17**	.10	.17**	.13*	.05	.16*	.14*

NOTE.

^**Correlation is significant at the 0.01 level (2-tailed);

^*Correlation is significant at the 0.05 level (2-tailed); Pittsburgh Sleep Quality Index (PSQI); Dysfunctional Beliefs and Attitudes Sleep (DBAS);

^aresults of point biserial correlation analysis

The summary score of the neighborhood environment scale was significantly associated with most PSQI components and the global score (sleep patterns and quality). Particularly, the subscales, neighborhood walking environment, and social cohesion were significantly associated with the PSQI, beliefs about sleep, and sleep hygiene behaviors. Age and gender were significantly associated with some of the neighborhood subscales (Table 3). However, both age (r=−.04, p=0.54) and gender (r_pb=0.02, p=0.74) were not significantly associated with the PSQI global score.

Table 4 presents the frequencies and mean differences in demographic characteristics, anthropometrics, chronic comorbid conditions, beliefs about sleep, sleep hygiene behaviors, and neighborhood environment by the sleep quality status (poor sleep quality: PSQI score 5 or higher). There was no significant difference in frequency or mean in sociodemographics or anthropometric characteristics by sleep status (good vs. poor sleep quality). Individuals with poor sleep quality were more likely to report poor social cohesion and poor overall neighborhood environment. Individuals with poor sleep quality also reported higher scores in poor sleep hygiene and dysfunctional beliefs about sleep than those with good sleep quality (Table 4).

Table 4.

Differences in selected variables by sleep quality (N=252)

Category	Total	Good sleep quality (n=70)	Poor sleep quality (n=182)
	N (%)/Mean (SD)	N (%)/Mean (SD)	N (%)/Mean (SD)	χ2 or t	p
Age	53.88 (14.17)	54.44 (15.49)	53.66 (13.67)	0.39	.69
Gender
Man	56 (22.2)	13 (18.6)	43 (23.6)	0.75	.49
Woman	196 (77.8)	57 (81.4)	139 (76.4)
Education level
≤Hight school graduate	126 (50.00)	30 (42.9)	96 (52.7)	1.98	.21
≥College graduate	126 (50.00)	40 (57.1)	86 (47.3)
Current work
Working Full Time	123 (48.81)	35 (50.0)	88 (48.4)	1.05	.59
Working Part Time	30 (11.90)	6 (8.6)	24 (13.2)
Unemployed	99 (39.29)	29 (41.4)	70 (38.5)
Annual household income
0-$39,999	87 (34.5)	21 (30.0)	66 (36.3)	0.93	.82
$40,000-$79,999	70 (27.8)	21 (30.0)	49 (26.9)
More than $80,000	66 (26.2)	19 (27.1)	47 (25.8)
Refused to answer	29 (11.5)	9 (12.9)	20 (11.0)
Number of comorbidities
0	100 (39.7)	28 (40.0)	72 (39.6)	0.79	.68
1–2	140 (55.6)	40 (57.1)	100 (54.9)
3–4	12 (4.8)	2 (2.9)	10 (5.5)
Body mass index (kg/m2)	31.93 (6.48)	32.30 (6.95)	31.78 (6.30)	0.56	.57
Waist-hip ratio	0.88 (0.09)	0.88 (0.09)	0.88 (0.09)	−0.17	.87
Body fat (%)	40.79 (10.24)	41.85 (10.07)	40.37 (10.30)	1.03	.31
Neighborhood environment
Aesthetic quality	2.22 (0.83)	2.14 (0.87)	2.26 (0.81)	−0.97	.33
Walking environment	2.35 (0.71)	2.22 (0.62)	2.40 (0.73)	−1.80	.07
Availability of healthy foods	2.77 (1.07)	2.65 (1.09)	2.82 (1.06)	−1.12	.26
Safety	2.59 (1.04)	2.38 (1.04)	2.67 (1.03)	−1.97	.05
Social cohesion	2.47 (0.81)	2.30 (0.73)	2.53 (0.83)	−2.05	.04*
Violence	1.43 (0.61)	1.34 (0.51)	1.47 (0.64)	−1.66	.09
Total score	2.28 (0.62)	2.15 (0.59)	2.33 (0.63)	−2.05	.04*
Sleep hygiene	23.10 (10.47)	17.07 (8.62)	25.41 (10.21)	−6.05	.00**
DBAS
Consequences of insomnia	4.18 (2.44)	3.09 (2.25)	4.60 (2.38)	−4.60	.00**
Worry about sleep	3.51 (2.16)	2.18 (1.62)	4.02 (2.13)	−7.38	.00**
Sleep expectations	5.98 (2.73)	5.74 (3.02)	6.07 (2.62)	−0.85	.39
Medication	2.57 (2.00)	1.80 (1.55)	2.86 (2.08)	−4.40	.00**
Total score	3.85 (1.80)	2.84 (1.46)	4.24 (1.78)	−5.89	.00**

Note. Poor sleep quality means Pittsburgh Sleep Quality Index (PSQI) global score of ≥ 5; Comorbidities include hypertension, stroke, myocardial infarction, chronic lung diseases, diabetes, and depression. Dysfunctional Beliefs and Attitudes Sleep (DBAS);

^**Correlation is significant at the 0.01 level (2-tailed);

^*Correlation is significant at the 0.05 level (2-tailed).

In the final multivariate model, the linear combination of the predictors in the model was significantly related to sleep quality. The R² was 0.37, indicating the model explained roughly 37% of the variance in the sleep quality. Beliefs about sleep, sleep hygiene, and neighborhood environment were significant predictors of sleep quality, controlling for a number of chronic comorbid conditions. Better sleep quality was associated with more functional beliefs about sleep, better sleep hygiene behaviors, and better neighborhood environment. Table 5 presents the results of the regression analysis.

Table 5.

Results of multiple regression analyses of Pittsburgh Sleep Quality Index global score (N=252)

	Variable	R	R2	Adjusted R2	R2 Change	F	P (change)	β	p	VIF
1	Number of comorbidities	.139	.019	.015	.019	4.945	.027	.139	.027	1.000
2	Number of comorbidities	.597	.357	.349	.337	63.990	.000	.082	.109	1.010
	DBAS							.366	.000	1.234
	Sleep Hygiene	.606	.367	.357	.011	4.231	.041	.323	.000	1.236
3	Number of comorbidities							.077	.133	1.013
	DBAS							.363	.000	1.235
	Sleep Hygiene							.307	.000	1.259
	Neighborhood							.106	.041	1.032

Note. The number of comorbidities included CVD, chronic lung disease (COPD, asthma), diabetes, and depression. Dysfunctional Beliefs and Attitudes Sleep (DBAS); Pittsburgh Sleep Quality Index (PSQI)

DISCUSSION

In our study of African American men and women, the majority of participants (72%) reported poor sleep quality based on the PSQI global score; the mean sleep duration was 6.47 hours. This finding is consistent with other studies of multiracial groups in which African Americans reported poorer sleep quality and the shortest sleep duration than other racial/ethnic groups as measured by actigraphy or self-report PSQI.^2,28 In a study of 2,230 multiethnic men and women, African American men and women had the shortest sleep duration (5.76 and 6.26 h, respectively); White women had the longest sleep.² Although the mechanism for the sleep disparities remains unclear, potential influences are the relatively high prevalence of comorbid conditions among African Americans, genetic difference; psychosocial, behavioral and environmental factors: socioeconomic status, beliefs about sleep, sleep hygiene behaviors, and disadvantaged neighborhoods.^2,29 Along the same lines, after controlling for the number of comorbid conditions in our multivariable analysis, beliefs about sleep, sleep hygiene behaviors, and the neighborhood environment were significantly associated with sleep quality among African Americans. Many of these factors are modifiable to improve sleep.

The score on dysfunctional beliefs about sleep in our African American participants was as high as in other studies with people with insomnia.^5,27 Many of our participants had dysfunctional or unrealistic sleep expectations and excessive worrying over sleep loss (consequences of insomnia), which may exacerbate sleep disturbances by increasing physical and emotional arousal.⁴ On the other hand, our study participants may not necessarily have had insomnia per se, but most were overweight or obese and may also have had sleep apnea that may have contributed to awakening at night and insomnia symptoms.

Dysfunctional beliefs about sleep were also significantly associated with educational levels, a finding that suggests the need for targeted interventions for those with low educational attainment. To date, limited data are available from clinical trials of behavioral interventions such as cognitive behavioral therapy for insomnia (CBT-I) to improve sleep-related cognitions (beliefs about sleep) among AAs. In studies of general populations, CBT-I was effective for reducing dysfunctional beliefs about sleep and increasing more adaptive beliefs about sleep in middle-aged or older adults.^5,27 However, whether such changes in sleep beliefs led to sleep improvement differed by a degree of sleep disturbance severity and other potential confounders such as sleep hygiene.^27,30 In a study of Japanese patients, sleep hygiene behaviors did not necessarily influence sleep improvement due in part to other confounders.³⁰ Therefore, assessing individuals’ sleep hygiene behaviors without understanding their sleep beliefs may have limitations in providing a tailored sleep treatment.³⁰

Studies have suggested that health disparities which are also linked to sleep disparities among African Americans may result from disadvantaged neighborhood environments.^13,14 In our study, the physical neighborhood and social environment were associated with beliefs about sleep, sleep hygiene and habitual sleep patterns. However, potential social environmental influences on sleep beliefs and sleep practices among African Americans have received little attention.

We found that neighborhood safety and social cohesion had small but significant correlations with most of the PSQI (sleep patterns and quality) sub-components and sleep latency, sleep duration, and sleep efficiency. Participants with good sleep quality were more likely to report better social cohesion, suggesting that they trusted their neighbors and felt connected within the community. This may reflect differences in stress levels among participants who live in unsafe neighborhoods or less supportive social environment that may contribute to high levels of stress and poor sleep in their daily lives.³¹ However, it is also possible that the relationship is bidirectional. That is, those who suffer from poor sleep may be critical to their environments or have negative perceptions about their environments.

Although little is known about the relationship between sleep and neighborhood social environments, in a study with disaster-affected communities in Japan, people who had strong networks in the neighborhood reported fewer sleep difficulties.³² Stressful family or negative social relationships also have harmful effects on sleep.³¹ Similarly, studies regarding a wider range of social factors have shown that social support, social isolation, and social capital are important social determinants of an individuals’ sleep.^33,34 Furthermore, growing evidence shows that the benefits of social environments, including social support, social ties, and social integration on health are particularly evident among African Americans.^35,36 Therefore, future efforts to build positive social relationships among family, friends, and neighbors may be needed to mitigate vulnerability to sleep problems and related health outcomes in African Americans.

The detrimental effects of poor neighborhood environments on sleep appear to be complex. For example, in a study of 1,896 multi-ethnic men and women, living in neighborhoods with poor walking environments was associated with greater severity of sleep apnea, independent of BMI, comorbidities and socioeconomic status, especially among men and obese individuals.³⁷ Other studies show that, overall, African Americans live in disadvantaged neighborhoods with higher crime rates and less access to healthy food or recreational facilities,^38,39 but that there are gender and age differences in perceptions of neighborhoods that may lead to different health outcomes, including sleep. In our sample, there were significant associations between gender and age with some perceived neighborhood environment factors. However, neither gender nor age was significantly associated with sleep quality. Studies have also shown that racial differences in neighborhood quality were substantial even after accounting for individual-level socioeconomic variables,³⁹ which suggests that individual-levels of socioeconomic status may not be used interchangeably with neighborhood environment for studying sleep disparities. Taken together, to understand sleep and sleep disparities among African Americans, it is important to study sleep within the context of both individual factors and physical neighborhood and social environment.

To our knowledge, this is the first study that has evaluated habitual sleep patterns, beliefs about sleep, sleep hygiene behaviors, and neighborhood environment comprehensively together in a high-risk group, African Americans. The community-based sample with diverse socioeconomic status within African Americans allowed us to assess the prevalence of sleep disturbances and its relationship with neighborhood environment within this population.

Despite these strengths, the study conclusions should be considered in light of several limitations. The majority of our participants were women who were overweight or obese, reflecting the national obesity prevalence in this group (80% of African American women in the US are overweight or obese).⁴⁰ Thus, the non-significant relationship among anthropometric characteristics (BMI, body fat, waist-hip ratio) and sleep-related variables may be due to the lack of variability in the BMI in our sample. It is also possible that obstructive sleep apnea, a disorder that was not evaluated in this study but is common in people who are overweight or obese, influenced the sleep variables in this study. Since our participants were recruited from churches, findings limit generalizability. Last, due to the nature of our cross-sectional study, causal directions among the studied variables─ for example, the relationship between perceived neighborhood environments and sleep measured by self- report ─ cannot be determined. Future longitudinal research with objective sleep measures using actigraphy or polysomnography and evaluation for specific sleep disorders (e.g., insomnia, sleep apnea) is needed to evaluate sleep and sleep-related cognition and behaviors among African Americans.

Future directions to reduce sleep health disparities and conclusion

Increases in population diversity and widening health disparity gaps in racial/ethnic groups continue to be a threat to public health. The importance of sleep for health and the growing magnitude of sleep problems in racial/ethnic groups require special attention.

In our study with African Americans, the rate of poor sleep quality was high. African Americans’ beliefs about sleep and sleep hygiene practice were associated with their sleep quality. We also found that one’s physical neighborhood and social environment may influence an individual’s beliefs about sleep, sleep hygiene and overall sleep quality. Given our study findings with the growing evidence, future programs to enhance sleep should consider the physical neighborhood, social environment and cultural context affecting sleep as well as individual factors.

Over the years, a large body of literature has demonstrated the importance of effective patient-healthcare provider relationships on health outcomes. There are still substantial efforts needed to address how to identify sleep problems and barriers to sleep improvement proactively and how to facilitate the conversation about common sleep problems among racial/ethnic minorities. Culturally tailored sleep health education for patients may also improve patient-healthcare provider relationships as well as sleep, and in turn, improve other health outcomes.²⁹ More importantly, sleep health education and relevant support for health care providers may facilitate effective treatment delivery with a timely diagnosis for underserved populations with sleep problems.

Improving physical neighborhood and social environment needs substantial efforts from various stakeholders, collaborative relationships with the community, and health policy involvement. Insufficient sleep or poor sleep quality may affect work productivity and public safety among workers. Public health implications for sleep health are significant in our society. Future research and programs for sleep health are needed to improve not only individual’s health and well-being but also overall benefits to society.

Abbreviations

AAs

African Americans

AME

African Methodist Episcopal

BMI

body mass index

CBT

cognitive behavioral therapy

CI

confidence interval

DBAS

Dysfunctional Beliefs and Attitudes about Sleep

PSQI

Pittsburgh Sleep Quality Index

SD

standard deviation

SHPS

Sleep Hygiene Practice Scale

VIF

variance inflation factors