- Email: [email protected]
Does sleep quality mediate the association between neighborhood disorder and self-rated physical health?
Preventive Medicine 51 (2010) 275–278
Contents lists available at ScienceDirect
Preventive Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y p m e d
Does sleep quality mediate the association between neighborhood disorder and self-rated physical health? Lauren Hale a,⁎, Terrence D. Hill b, Amy M. Burdette b a b
Preventive Medicine, Graduate Program in Public Health, State University of New York, Stony Brook, NY 11794-8338, USA Sociology, Florida State University, USA
a r t i c l e
i n f o
Available online 30 June 2010 Keywords: Sleep Neighborhood Physical health
a b s t r a c t Objectives. We examine the association between perceived neighborhood disorder and self-rated physical health. Building on previous research, we test whether this association is mediated by sleep quality. Methods. We use data from the 2004 Survey of Texas Adults (n = 1323) to estimate a series of ordinary least squares regression models. We formally assess mediation by testing for significant changes in the effect of neighborhood disorder before and after adjusting for sleep quality. Results. We find that residence in a neighborhood that is perceived as noisy, unclean, and crime-ridden is associated with poorer self-rated physical health, even with controls for irregular exercise, poor diet quality, smoking, binge drinking, obesity and a host of relevant sociodemographic factors. Our results also indicate that the relationship between neighborhood disorder and self-rated physical health is partially mediated by lower sleep quality. Conclusion. Targeted interventions designed to promote sleep quality in disadvantaged neighborhoods may help to improve the physical health of residents in the short-term. Policies aimed at solving the problem of neighborhood disorder are needed to support sleep quality and physical health in the long-term. © 2010 Elsevier Inc. All rights reserved.
Introduction Residents of neighborhoods characterized by social disorganization (e.g., socioeconomic disadvantage) and disorder (e.g., noise, crime, and dilapidation) tend to exhibit poorer physical health than residents of more advantaged neighborhoods (Hill et al., 2005; Ross and Mirowsky, 2001). While prior research has contributed to our understanding of the physical health consequences of neighborhood context, additional research is needed to explain these patterns. Building on previous work, we test whether the association between neighborhood disorder (perceptions of neighborhood noise, cleanliness, and crime) and selfrated physical health is mediated by sleep quality. Research shows that residence in disadvantaged and urban neighborhood environments is associated with low sleep quality and shortened sleep duration (Hale and Do, 2007; Hill et al., 2009). It is also well known that poor sleep quality, short sleep duration, and sleep disorders can contribute to a range of physical health risks, including, for example, hypertension, cardiovascular disease, stroke, metabolic dysregulation, and cognitive decrements. If living in a disadvantaged neighborhood environment can disrupt sleep, and poor sleep can undermine physical health, the association between perceived neighborhood disorder and self-rated physical
⁎ Corresponding author. Fax: +1 631 444 3480. E-mail address: [email protected] (L. Hale). 0091-7435/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2010.06.017
health could be at least partially mediated or explained by sleep quality. Methods Data We use data from the 2004 Survey of Texas Adults, a statewide probability sample of 1504 community-dwelling adults (18 and over) residing in Texas (Musick, 2004). Sampling was conducted using a modified random digit dialing design, with a household-level cooperation rate of 37% and a respondent-level cooperation rate of 89%. Each computer-assisted telephone interview lasted approximately 30–35 min and was also available in Spanish. All analyses are weighted to match the sample to the Texas population. Listwise deletion reduced the analytic sample from 1504 to 1323. Measures Self-rated physical health, our focal outcome variable, is a widely used measure of general physical health status that is strongly correlated with physician diagnoses and mortality risk. Respondents were asked: “How would you rate your physical health at the present time? Would you say it is excellent, very good, good, fair, or poor?” Response categories ranged from (1) poor to (5) excellent. Neighborhood disorder, our focal predictor variable, is measured as the mean response to three items designed to assess perceptions of problems in the neighborhood, including social disorganization and disorder, ambient hazards, and structural disrepair: “There is a lot of crime in my
276
L. Hale et al. / Preventive Medicine 51 (2010) 275–278
neighborhood.” “My neighborhood is noisy.” “My neighborhood is clean.” Response categories ranged from (1) strongly agree to (5) strongly disagree, with reverse codes for “crime” and “noise.” Sleep quality, our focal mediator variable, is measured with the following question: “How would you rate your sleep quality overall for the past 30 days?” Response categories ranged from (1) poor to (5) excellent. Our multivariate analyses also include a range background and lifestyle factors that are known to be associated with both low sleep quality and poor physical health: Age, Gender, Race/Ethnicity, Citizenship Status, Interview Language, Education, Employment, Family Income, Financial Strain, Marital Status, Number of Children, Irregular Exercise, Poor Diet Quality, Smoking, Binge Drinking, and Obesity. Coding for these variables is noted in Table 1. Statistical procedures Our mediation analysis proceeds in three steps. We establish a baseline in Model 1 by regressing self-rated physical health on neighborhood disorder
Table 1 Weighted descriptive statistics, 2004 Survey of Texas Adults (n = 1,323). Range
Mean
Focal measures Neighborhood disorder Sleep quality Physical health
1–5 1–5 1–5
2.28 2.82 3.42
Lifestyle factors Irregular exercise Diet quality (fair or poor) Current smoker Binge drinker (5 or more drinks) Obese
0–1 0–1 0–1 0–1 0–1
0.40 0.37 0.24 0.10 0.23
18–94 0–1 0–1 0–1 0–1 0–1 0–1 0–1 0–4 0–1 1–6 0–1 0–4 0–1 0–4
41.54 0.51 0.47 0.11 0.29 0.13 0.15 0.17 1.26 0.56 3.03 .17 1.20 0.53 1.03
Background factors Age Male Non-Hispanic White Black Mexican Other race or ethnicity Non-citizen Spanish interview Education Employed Family income Income missing Financial strain Married Number of children
SD .8 1.2 1.1
Cronbach's alpha .58
17.3
1.2 1.5 1.2 1.2
Notes. Coding for background and lifestyle factors: Age (measured in continuous years); Gender (1 = male, 0 = female); Race/ethnicity (four dummy variables for African American, Mexican/Mexican American, other Hispanic, and other minority, with nonHispanic white serving as the reference category); Citizenship status (1 = non-citizen, 0 = U.S. citizen); Interview language (1 = Spanish language interview, 0 = English language interview); Education (0 = less than a high school degree to 4 = graduate degree); Employment (1 = currently employed, 0 = other work status); Family income (1 = $0–$14,900 to 6 = $85,000 or more); Financial strain (0 = no difficulty paying monthly bills to 4 = extreme difficulty); Marital status (1 = married, 0 = otherwise); and Number of children (0 to 4 or more). Missing values on family income were replaced with the mode of the original income measure ($35,000–$49,000); as a precaution, our models control for whether or not the respondent was missing on income (1 = missing income, 0 = otherwise). Irregular exercisers were respondents who did not (a) walk or (b) engage in moderate exercise (e.g., playing golf, dancing, and gardening) on five or more days per week or (c) engage in strenuous exercise (e.g., running, swimming, and lifting weights) on three or more days per week. Diet quality is measured with the following question: “Overall, how would you rate the quality of your diet?” Response categories ranged from (1) excellent to (5) poor. To indicate suboptimal diet quality this item was recoded (1) for fair or poor and (0) otherwise. Smoking is measured with the following question: “Are you a current smoker, a former smoker or have you never smoked?” This item is coded (1) for a current smoker and (0) otherwise. Binge drinking is measured with the following question: “On those days that you drank alcohol, about how many drinks did you usually have?” This item is coded (1) for five or more drinks per drinking occasion and (0) otherwise. Obesity is measured using self-reports of height and weight, and then converted to body mass index (BMI). We coded respondents with a BMI equal to or greater than 30 as (1) obese and (0) otherwise.
and all background factors. Because previous research has linked neighborhood disorder with irregular exercise, poor diet quality, smoking, binge drinking, and obesity, we adjust for health and lifestyle factors in Model 2. Finally, Model 3 adjusts for sleep quality. To formally assess our mediation hypothesis, we employ the Clogg statistic to test for significant changes in the effects of disorder across nested models (i.e., before and after adjusting for sleep quality). A statistically significant reduction in the magnitude of the coefficient for neighborhood disorder—from Model 2 to Model 3—would suggest mediation.
Results Table 1 provides descriptive characteristics of the sample. The average respondent reports fairly low levels of neighborhood disorder and “good” physical health and sleep quality. With respect to lifestyle factors, we observe moderate to low rates of irregular exercise, fair or poor diet quality, smoking, binge drinking, and obesity. Model 1 of Table 2 shows that neighborhood disorder is inversely associated with self-rated physical health. Those respondents who live in neighborhoods they characterize as being unclean, unsafe, and noisy report poorer physical health than those who live in areas that are perceived as clean, quiet, and safe, even after adjustment for a range of background factors. This pattern is consistent across Models 2 and 3, with additional controls for lifestyle factors and sleep quality. In Model 3, the magnitude of the standardized regression coefficient for neighborhood disorder (β = .06) is comparable to other statistically significant predictors of self-rated physical health like current smoker (β = .07) and financial strain (β = .08), but smaller than sleep quality (β = .23), obesity (β = .15), and age (β = .21). Model 2 adds lifestyle factors to Model 1. These adjustments explain approximately 18% ([.11 − .09]/.11) of the association between neighborhood disorder and self-rated physical health. In Model 3, sleep quality accounts for an additional 11% ([.09 − .08]/.09) of the association between neighborhood disorder and self-rated physical health, which is a statistically significant reduction (t = 2.50, P b .05). Because the coefficient for neighborhood disorder remains statistically significant in Model 3, these results suggest that sleep quality is a partial mediator of the association between neighborhood disorder and self-rated physical health. Results from the full model show that other statistically significant (all at P b .01) potentially modifiable factors associated with poorer self-rated health include financial strain, unemployment, irregular exercise, smoking, poor diet, and obesity. Discussion Building on previous research, we used data from a statewide probability sample of Texas adults to formally test whether the association between neighborhood disorder and self-rated physical health is mediated by sleep quality. We found that residence in a neighborhood that is perceived as noisy, unclean, and crime-ridden is associated with a modest but statistically significant reduction in selfrated physical health. Our results also indicate that the relationship between neighborhood disorder and self-rated physical health is partially mediated by poor sleep quality. The observed association between neighborhood disorder and poor sleep quality is consistent with previous studies of disadvantaged neighborhood conditions and adverse sleep outcomes (Hale and Do, 2007; Hill et al., 2009). The association between neighborhood disorder and self-rated health also confirms prior work (Ross and Mirowsky, 2001). To the best of our knowledge, we are the first to empirically show that the association between neighborhood disorder and self-reported physical health is mediated by sleep quality. If the association between neighborhood disorder and selfreported physical health is mediated by sleep quality, what might account for this process? Researchers speculate that because sleep is
L. Hale et al. / Preventive Medicine 51 (2010) 275–278
277
Table 2 Weighted ordinary least squares regression of self-rated physical health, 2004 Survey of Texas Adults (n = 1323). Model 1
Focal measures Neighborhood disorder Sleep quality
Model 2
b
SE
β
−.11
.04
−.08
⁎⁎
Lifestyle factors Irregular exercise Diet quality (fair or poor) Current smoker Binge drinker (5 or more drinks) Obese Background factors Age Male Black Mexican Other race or ethnicity Non-citizen Spanish interview Education Employed Family income Income missing Financial strain Married Number of children Model statistics Model F Nested F R-squared
−.01 −.02 −.02 −.32 −.34 .06 .12 .09 .30 .05 −.10 −.13 −.00 .00
.00 .06 .10 .08 .09 .11 .11 .03 .06 .02 .08 .03 .06 .03
16.27
−.22 −.01 −.00 −.13 −.10 .02 .04 .09 .13 .07 −.03 −.14 −.00 .01
⁎⁎⁎
⁎⁎⁎ ⁎⁎⁎
⁎⁎ ⁎⁎⁎ ⁎ ⁎⁎⁎
⁎⁎⁎
.16
Model 3
b
SE
β
b
SE
β
−.09
.04
−.07
⁎
−.08 .21
.03 .02
−.06 .23
⁎ ⁎⁎⁎
−.24 −.39 −.21 .18 −.41
.06 .06 .07 .09 .07
−.11 −.17 −.08 .05 −.15
⁎⁎⁎ ⁎⁎⁎ ⁎⁎
−.24 −.25 −.19 .16 −.39
.06 .06 .07 .09 .06
−.11 −.11 −.07 .04 −.15
⁎⁎⁎ ⁎⁎⁎ ⁎⁎
−.01 −.04 −.06 −.29 −.29 .04 .07 .06 .31 .03 −.16 −.10 −.00 .03
.00 .06 .10 .08 .09 .10 .11 .03 .06 .02 .07 .03 .06 .03
−.19 −.02 −.02 −.12 −.09 .01 .02 .06 .14 .04 −.05 −.10 −.00 .03
⁎⁎⁎
−.01 −.07 −.03 −.28 −.23 .03 .06 .05 .28 .03 −.18 −.07 −.03 .04
.00 .06 .09 .08 .09 .10 .10 .03 .06 .02 .07 .03 .06 .02
−.21 −.03 −.01 −.12 −.07 .01 .02 .05 .13 .04 −.06 −.08 −.01 .05
⁎⁎⁎
20.84 29.28 .24
⁎⁎⁎
⁎⁎⁎ ⁎⁎
⁎ ⁎⁎⁎ ⁎ ⁎⁎⁎
⁎⁎⁎ ⁎⁎⁎
24.77 78.51 .29
⁎⁎⁎
⁎⁎⁎ ⁎
⁎⁎⁎ ⁎ ⁎⁎
⁎⁎⁎ ⁎⁎⁎
Notes. Shown are unstandardized OLS regression coefficients (b), standard errors (SE), and standardized coefficients (β). The regression coefficients represent the change in units (1– 5) in self-rated health status with each unit change in the explanatory variable. ⁎ P b .05. ⁎⁎ P b .01. ⁎⁎⁎ P b .001.
an adaptive behavior, neighborhoods that are characterized by noise (from neighbors, busy streets, and crowding), dilapidation (substandard housing), and crime (fear of victimization) may directly undermine the ability of residents to initiate and/or maintain sleep (Hale and Do, 2007; Hill et al., 2009). Studies also suggest that stressful neighborhood conditions could contribute to poor sleep quality through various psychological and physiological pathways. For example, perceptions of noise and crime could elicit short-term feelings of annoyance, fear, and hopelessness. These feelings could activate the stress response and trigger the release stress hormones (epinephrine and cortisol) that promote mental and physiological arousal (Van Reeth et al., 2000). There are several key pathways through which poor sleep might undermine physical health. Psychologically, there are numerous bidirectional links between sleep problems and mental health, depression and suicidal ideation (Taylor et al., 2005). Inadequate sleep can contribute to a range of risky healthy behaviors, such as smoking, heavy drinking, reckless driving, and physical inactivity (Strine and Chapman, 2005). Physiologically, disordered sleep can modify neuroendocrine activity, increase hypertension, impair insulin resistance, disturb immune function, and alter hormones that regulate appetite (McEwen, 2006; Spiegel et al., 2004). The present study has several limitations. The cross-sectional nature of the data makes it difficult to establish the causal order of many observed associations. For example, does poor quality sleep cause poor physical health, or is it the other way around? Our data are also limited to single-item measurements of self-rated physical health and sleep quality. Although these items have demonstrated
construct validity in previous research, reliability is a cause for concern. Additional research is needed to replicate our results with longitudinal data, more specific physical health outcomes (e.g., hypertension, cardiovascular disease, and stroke), and other established measures of physical health (e.g., the Health Status Questionnaire) and sleep quality (e.g., the Pittsburgh Sleep Quality Index). Despite these limitations, our results suggest that neighborhood disorder may contribute to poor physical health by undermining sleep quality. Based on these results, targeted interventions designed to promote sleep quality in disadvantaged neighborhoods (e.g., community-based sleep promotion and noise level ordinances) could help to improve the physical health of residents in the short-term. However, more complex policies aimed at solving the problem of neighborhood disorder are needed to support sleep quality and physical health in the long-term. Conflict of interest statement The authors have no conflicts of interest to report.
References Hale, L., Do, D.P., 2007. Racial differences in self-report of sleep duration in a population-based study. Sleep 30 (9), 1092–1099. Hill, T.D., Ross, C.E., Angel, R.J., 2005. Neighborhood disorder, psychophysiological distress, and health. J. Health Soc. Behav. 46 (2), 170–186. Hill, T.D., Burdette, A.M., Hale, L., 2009. Neighborhood disorder, sleep quality, and psychological distress: testing a model of structural amplification. Health Place 15 (4), 1006–1013. McEwen, B.S., 2006. Sleep deprivation as a neurobiologic and physiologic stressor: allostasis and allostatic load. Metabolism 55 (10 Suppl 2), S20–S23.
278
L. Hale et al. / Preventive Medicine 51 (2010) 275–278
Musick, M., 2004. Survey of Texas Adults. The University of Texas at Austin, Austin, TX. Ross, C.E., Mirowsky, J., 2001. Neighborhood disadvantage, disorder, and health. J. Health Soc. Behav. 42 (3), 258–276. Spiegel, K., Leproult, R., L'Hermite-Baleriaux, M., Copinschi, G., Penev, P.D., Van Cauter, E., 2004. Leptin levels are dependent on sleep duration: relationships with sympathovagal balance, carbohydrate regulation, cortisol, and thyrotropin. J. Clin. Endocrinol. Metab. 89 (11), 5762–5771.
Strine, T.W., Chapman, D.P., 2005. Associations of frequent sleep insufficiency with health-related quality of life and health behaviors. Sleep Med. 6 (1), 23–27. Taylor, D.J., Lichstein, K.L., Durrence, H.H., Riedel, B.W., Bush, A.J., 2005. Epidemiology of insomnia, depression, and anxiety. Sleep 28 (11), 1457–1464. Van Reeth, O., Weibel, L., Spiegel, K., Leproult, R., Dugovic, C., Maccari, S., 2000. Interactions between stress and sleep: from basic research to clinical situations. Sleep Med. Rev. 4 (2), 201–219.
Contents lists available at ScienceDirect
Preventive Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y p m e d
Does sleep quality mediate the association between neighborhood disorder and self-rated physical health? Lauren Hale a,⁎, Terrence D. Hill b, Amy M. Burdette b a b
Preventive Medicine, Graduate Program in Public Health, State University of New York, Stony Brook, NY 11794-8338, USA Sociology, Florida State University, USA
a r t i c l e
i n f o
Available online 30 June 2010 Keywords: Sleep Neighborhood Physical health
a b s t r a c t Objectives. We examine the association between perceived neighborhood disorder and self-rated physical health. Building on previous research, we test whether this association is mediated by sleep quality. Methods. We use data from the 2004 Survey of Texas Adults (n = 1323) to estimate a series of ordinary least squares regression models. We formally assess mediation by testing for significant changes in the effect of neighborhood disorder before and after adjusting for sleep quality. Results. We find that residence in a neighborhood that is perceived as noisy, unclean, and crime-ridden is associated with poorer self-rated physical health, even with controls for irregular exercise, poor diet quality, smoking, binge drinking, obesity and a host of relevant sociodemographic factors. Our results also indicate that the relationship between neighborhood disorder and self-rated physical health is partially mediated by lower sleep quality. Conclusion. Targeted interventions designed to promote sleep quality in disadvantaged neighborhoods may help to improve the physical health of residents in the short-term. Policies aimed at solving the problem of neighborhood disorder are needed to support sleep quality and physical health in the long-term. © 2010 Elsevier Inc. All rights reserved.
Introduction Residents of neighborhoods characterized by social disorganization (e.g., socioeconomic disadvantage) and disorder (e.g., noise, crime, and dilapidation) tend to exhibit poorer physical health than residents of more advantaged neighborhoods (Hill et al., 2005; Ross and Mirowsky, 2001). While prior research has contributed to our understanding of the physical health consequences of neighborhood context, additional research is needed to explain these patterns. Building on previous work, we test whether the association between neighborhood disorder (perceptions of neighborhood noise, cleanliness, and crime) and selfrated physical health is mediated by sleep quality. Research shows that residence in disadvantaged and urban neighborhood environments is associated with low sleep quality and shortened sleep duration (Hale and Do, 2007; Hill et al., 2009). It is also well known that poor sleep quality, short sleep duration, and sleep disorders can contribute to a range of physical health risks, including, for example, hypertension, cardiovascular disease, stroke, metabolic dysregulation, and cognitive decrements. If living in a disadvantaged neighborhood environment can disrupt sleep, and poor sleep can undermine physical health, the association between perceived neighborhood disorder and self-rated physical
⁎ Corresponding author. Fax: +1 631 444 3480. E-mail address: [email protected] (L. Hale). 0091-7435/$ – see front matter © 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2010.06.017
health could be at least partially mediated or explained by sleep quality. Methods Data We use data from the 2004 Survey of Texas Adults, a statewide probability sample of 1504 community-dwelling adults (18 and over) residing in Texas (Musick, 2004). Sampling was conducted using a modified random digit dialing design, with a household-level cooperation rate of 37% and a respondent-level cooperation rate of 89%. Each computer-assisted telephone interview lasted approximately 30–35 min and was also available in Spanish. All analyses are weighted to match the sample to the Texas population. Listwise deletion reduced the analytic sample from 1504 to 1323. Measures Self-rated physical health, our focal outcome variable, is a widely used measure of general physical health status that is strongly correlated with physician diagnoses and mortality risk. Respondents were asked: “How would you rate your physical health at the present time? Would you say it is excellent, very good, good, fair, or poor?” Response categories ranged from (1) poor to (5) excellent. Neighborhood disorder, our focal predictor variable, is measured as the mean response to three items designed to assess perceptions of problems in the neighborhood, including social disorganization and disorder, ambient hazards, and structural disrepair: “There is a lot of crime in my
276
L. Hale et al. / Preventive Medicine 51 (2010) 275–278
neighborhood.” “My neighborhood is noisy.” “My neighborhood is clean.” Response categories ranged from (1) strongly agree to (5) strongly disagree, with reverse codes for “crime” and “noise.” Sleep quality, our focal mediator variable, is measured with the following question: “How would you rate your sleep quality overall for the past 30 days?” Response categories ranged from (1) poor to (5) excellent. Our multivariate analyses also include a range background and lifestyle factors that are known to be associated with both low sleep quality and poor physical health: Age, Gender, Race/Ethnicity, Citizenship Status, Interview Language, Education, Employment, Family Income, Financial Strain, Marital Status, Number of Children, Irregular Exercise, Poor Diet Quality, Smoking, Binge Drinking, and Obesity. Coding for these variables is noted in Table 1. Statistical procedures Our mediation analysis proceeds in three steps. We establish a baseline in Model 1 by regressing self-rated physical health on neighborhood disorder
Table 1 Weighted descriptive statistics, 2004 Survey of Texas Adults (n = 1,323). Range
Mean
Focal measures Neighborhood disorder Sleep quality Physical health
1–5 1–5 1–5
2.28 2.82 3.42
Lifestyle factors Irregular exercise Diet quality (fair or poor) Current smoker Binge drinker (5 or more drinks) Obese
0–1 0–1 0–1 0–1 0–1
0.40 0.37 0.24 0.10 0.23
18–94 0–1 0–1 0–1 0–1 0–1 0–1 0–1 0–4 0–1 1–6 0–1 0–4 0–1 0–4
41.54 0.51 0.47 0.11 0.29 0.13 0.15 0.17 1.26 0.56 3.03 .17 1.20 0.53 1.03
Background factors Age Male Non-Hispanic White Black Mexican Other race or ethnicity Non-citizen Spanish interview Education Employed Family income Income missing Financial strain Married Number of children
SD .8 1.2 1.1
Cronbach's alpha .58
17.3
1.2 1.5 1.2 1.2
Notes. Coding for background and lifestyle factors: Age (measured in continuous years); Gender (1 = male, 0 = female); Race/ethnicity (four dummy variables for African American, Mexican/Mexican American, other Hispanic, and other minority, with nonHispanic white serving as the reference category); Citizenship status (1 = non-citizen, 0 = U.S. citizen); Interview language (1 = Spanish language interview, 0 = English language interview); Education (0 = less than a high school degree to 4 = graduate degree); Employment (1 = currently employed, 0 = other work status); Family income (1 = $0–$14,900 to 6 = $85,000 or more); Financial strain (0 = no difficulty paying monthly bills to 4 = extreme difficulty); Marital status (1 = married, 0 = otherwise); and Number of children (0 to 4 or more). Missing values on family income were replaced with the mode of the original income measure ($35,000–$49,000); as a precaution, our models control for whether or not the respondent was missing on income (1 = missing income, 0 = otherwise). Irregular exercisers were respondents who did not (a) walk or (b) engage in moderate exercise (e.g., playing golf, dancing, and gardening) on five or more days per week or (c) engage in strenuous exercise (e.g., running, swimming, and lifting weights) on three or more days per week. Diet quality is measured with the following question: “Overall, how would you rate the quality of your diet?” Response categories ranged from (1) excellent to (5) poor. To indicate suboptimal diet quality this item was recoded (1) for fair or poor and (0) otherwise. Smoking is measured with the following question: “Are you a current smoker, a former smoker or have you never smoked?” This item is coded (1) for a current smoker and (0) otherwise. Binge drinking is measured with the following question: “On those days that you drank alcohol, about how many drinks did you usually have?” This item is coded (1) for five or more drinks per drinking occasion and (0) otherwise. Obesity is measured using self-reports of height and weight, and then converted to body mass index (BMI). We coded respondents with a BMI equal to or greater than 30 as (1) obese and (0) otherwise.
and all background factors. Because previous research has linked neighborhood disorder with irregular exercise, poor diet quality, smoking, binge drinking, and obesity, we adjust for health and lifestyle factors in Model 2. Finally, Model 3 adjusts for sleep quality. To formally assess our mediation hypothesis, we employ the Clogg statistic to test for significant changes in the effects of disorder across nested models (i.e., before and after adjusting for sleep quality). A statistically significant reduction in the magnitude of the coefficient for neighborhood disorder—from Model 2 to Model 3—would suggest mediation.
Results Table 1 provides descriptive characteristics of the sample. The average respondent reports fairly low levels of neighborhood disorder and “good” physical health and sleep quality. With respect to lifestyle factors, we observe moderate to low rates of irregular exercise, fair or poor diet quality, smoking, binge drinking, and obesity. Model 1 of Table 2 shows that neighborhood disorder is inversely associated with self-rated physical health. Those respondents who live in neighborhoods they characterize as being unclean, unsafe, and noisy report poorer physical health than those who live in areas that are perceived as clean, quiet, and safe, even after adjustment for a range of background factors. This pattern is consistent across Models 2 and 3, with additional controls for lifestyle factors and sleep quality. In Model 3, the magnitude of the standardized regression coefficient for neighborhood disorder (β = .06) is comparable to other statistically significant predictors of self-rated physical health like current smoker (β = .07) and financial strain (β = .08), but smaller than sleep quality (β = .23), obesity (β = .15), and age (β = .21). Model 2 adds lifestyle factors to Model 1. These adjustments explain approximately 18% ([.11 − .09]/.11) of the association between neighborhood disorder and self-rated physical health. In Model 3, sleep quality accounts for an additional 11% ([.09 − .08]/.09) of the association between neighborhood disorder and self-rated physical health, which is a statistically significant reduction (t = 2.50, P b .05). Because the coefficient for neighborhood disorder remains statistically significant in Model 3, these results suggest that sleep quality is a partial mediator of the association between neighborhood disorder and self-rated physical health. Results from the full model show that other statistically significant (all at P b .01) potentially modifiable factors associated with poorer self-rated health include financial strain, unemployment, irregular exercise, smoking, poor diet, and obesity. Discussion Building on previous research, we used data from a statewide probability sample of Texas adults to formally test whether the association between neighborhood disorder and self-rated physical health is mediated by sleep quality. We found that residence in a neighborhood that is perceived as noisy, unclean, and crime-ridden is associated with a modest but statistically significant reduction in selfrated physical health. Our results also indicate that the relationship between neighborhood disorder and self-rated physical health is partially mediated by poor sleep quality. The observed association between neighborhood disorder and poor sleep quality is consistent with previous studies of disadvantaged neighborhood conditions and adverse sleep outcomes (Hale and Do, 2007; Hill et al., 2009). The association between neighborhood disorder and self-rated health also confirms prior work (Ross and Mirowsky, 2001). To the best of our knowledge, we are the first to empirically show that the association between neighborhood disorder and self-reported physical health is mediated by sleep quality. If the association between neighborhood disorder and selfreported physical health is mediated by sleep quality, what might account for this process? Researchers speculate that because sleep is
L. Hale et al. / Preventive Medicine 51 (2010) 275–278
277
Table 2 Weighted ordinary least squares regression of self-rated physical health, 2004 Survey of Texas Adults (n = 1323). Model 1
Focal measures Neighborhood disorder Sleep quality
Model 2
b
SE
β
−.11
.04
−.08
⁎⁎
Lifestyle factors Irregular exercise Diet quality (fair or poor) Current smoker Binge drinker (5 or more drinks) Obese Background factors Age Male Black Mexican Other race or ethnicity Non-citizen Spanish interview Education Employed Family income Income missing Financial strain Married Number of children Model statistics Model F Nested F R-squared
−.01 −.02 −.02 −.32 −.34 .06 .12 .09 .30 .05 −.10 −.13 −.00 .00
.00 .06 .10 .08 .09 .11 .11 .03 .06 .02 .08 .03 .06 .03
16.27
−.22 −.01 −.00 −.13 −.10 .02 .04 .09 .13 .07 −.03 −.14 −.00 .01
⁎⁎⁎
⁎⁎⁎ ⁎⁎⁎
⁎⁎ ⁎⁎⁎ ⁎ ⁎⁎⁎
⁎⁎⁎
.16
Model 3
b
SE
β
b
SE
β
−.09
.04
−.07
⁎
−.08 .21
.03 .02
−.06 .23
⁎ ⁎⁎⁎
−.24 −.39 −.21 .18 −.41
.06 .06 .07 .09 .07
−.11 −.17 −.08 .05 −.15
⁎⁎⁎ ⁎⁎⁎ ⁎⁎
−.24 −.25 −.19 .16 −.39
.06 .06 .07 .09 .06
−.11 −.11 −.07 .04 −.15
⁎⁎⁎ ⁎⁎⁎ ⁎⁎
−.01 −.04 −.06 −.29 −.29 .04 .07 .06 .31 .03 −.16 −.10 −.00 .03
.00 .06 .10 .08 .09 .10 .11 .03 .06 .02 .07 .03 .06 .03
−.19 −.02 −.02 −.12 −.09 .01 .02 .06 .14 .04 −.05 −.10 −.00 .03
⁎⁎⁎
−.01 −.07 −.03 −.28 −.23 .03 .06 .05 .28 .03 −.18 −.07 −.03 .04
.00 .06 .09 .08 .09 .10 .10 .03 .06 .02 .07 .03 .06 .02
−.21 −.03 −.01 −.12 −.07 .01 .02 .05 .13 .04 −.06 −.08 −.01 .05
⁎⁎⁎
20.84 29.28 .24
⁎⁎⁎
⁎⁎⁎ ⁎⁎
⁎ ⁎⁎⁎ ⁎ ⁎⁎⁎
⁎⁎⁎ ⁎⁎⁎
24.77 78.51 .29
⁎⁎⁎
⁎⁎⁎ ⁎
⁎⁎⁎ ⁎ ⁎⁎
⁎⁎⁎ ⁎⁎⁎
Notes. Shown are unstandardized OLS regression coefficients (b), standard errors (SE), and standardized coefficients (β). The regression coefficients represent the change in units (1– 5) in self-rated health status with each unit change in the explanatory variable. ⁎ P b .05. ⁎⁎ P b .01. ⁎⁎⁎ P b .001.
an adaptive behavior, neighborhoods that are characterized by noise (from neighbors, busy streets, and crowding), dilapidation (substandard housing), and crime (fear of victimization) may directly undermine the ability of residents to initiate and/or maintain sleep (Hale and Do, 2007; Hill et al., 2009). Studies also suggest that stressful neighborhood conditions could contribute to poor sleep quality through various psychological and physiological pathways. For example, perceptions of noise and crime could elicit short-term feelings of annoyance, fear, and hopelessness. These feelings could activate the stress response and trigger the release stress hormones (epinephrine and cortisol) that promote mental and physiological arousal (Van Reeth et al., 2000). There are several key pathways through which poor sleep might undermine physical health. Psychologically, there are numerous bidirectional links between sleep problems and mental health, depression and suicidal ideation (Taylor et al., 2005). Inadequate sleep can contribute to a range of risky healthy behaviors, such as smoking, heavy drinking, reckless driving, and physical inactivity (Strine and Chapman, 2005). Physiologically, disordered sleep can modify neuroendocrine activity, increase hypertension, impair insulin resistance, disturb immune function, and alter hormones that regulate appetite (McEwen, 2006; Spiegel et al., 2004). The present study has several limitations. The cross-sectional nature of the data makes it difficult to establish the causal order of many observed associations. For example, does poor quality sleep cause poor physical health, or is it the other way around? Our data are also limited to single-item measurements of self-rated physical health and sleep quality. Although these items have demonstrated
construct validity in previous research, reliability is a cause for concern. Additional research is needed to replicate our results with longitudinal data, more specific physical health outcomes (e.g., hypertension, cardiovascular disease, and stroke), and other established measures of physical health (e.g., the Health Status Questionnaire) and sleep quality (e.g., the Pittsburgh Sleep Quality Index). Despite these limitations, our results suggest that neighborhood disorder may contribute to poor physical health by undermining sleep quality. Based on these results, targeted interventions designed to promote sleep quality in disadvantaged neighborhoods (e.g., community-based sleep promotion and noise level ordinances) could help to improve the physical health of residents in the short-term. However, more complex policies aimed at solving the problem of neighborhood disorder are needed to support sleep quality and physical health in the long-term. Conflict of interest statement The authors have no conflicts of interest to report.
References Hale, L., Do, D.P., 2007. Racial differences in self-report of sleep duration in a population-based study. Sleep 30 (9), 1092–1099. Hill, T.D., Ross, C.E., Angel, R.J., 2005. Neighborhood disorder, psychophysiological distress, and health. J. Health Soc. Behav. 46 (2), 170–186. Hill, T.D., Burdette, A.M., Hale, L., 2009. Neighborhood disorder, sleep quality, and psychological distress: testing a model of structural amplification. Health Place 15 (4), 1006–1013. McEwen, B.S., 2006. Sleep deprivation as a neurobiologic and physiologic stressor: allostasis and allostatic load. Metabolism 55 (10 Suppl 2), S20–S23.
278
L. Hale et al. / Preventive Medicine 51 (2010) 275–278
Musick, M., 2004. Survey of Texas Adults. The University of Texas at Austin, Austin, TX. Ross, C.E., Mirowsky, J., 2001. Neighborhood disadvantage, disorder, and health. J. Health Soc. Behav. 42 (3), 258–276. Spiegel, K., Leproult, R., L'Hermite-Baleriaux, M., Copinschi, G., Penev, P.D., Van Cauter, E., 2004. Leptin levels are dependent on sleep duration: relationships with sympathovagal balance, carbohydrate regulation, cortisol, and thyrotropin. J. Clin. Endocrinol. Metab. 89 (11), 5762–5771.
Strine, T.W., Chapman, D.P., 2005. Associations of frequent sleep insufficiency with health-related quality of life and health behaviors. Sleep Med. 6 (1), 23–27. Taylor, D.J., Lichstein, K.L., Durrence, H.H., Riedel, B.W., Bush, A.J., 2005. Epidemiology of insomnia, depression, and anxiety. Sleep 28 (11), 1457–1464. Van Reeth, O., Weibel, L., Spiegel, K., Leproult, R., Dugovic, C., Maccari, S., 2000. Interactions between stress and sleep: from basic research to clinical situations. Sleep Med. Rev. 4 (2), 201–219.