The Relationship Between Neighborhood Characteristics and Self-Rated Health for Adults With Chronic Conditions

Abstract

Objectives. We sought to determine whether the association between neighborhood characteristics and health differs for people with and without a chronic condition.

Methods. We analyzed data from 2536 adults from the Los Angeles Family and Neighborhood Survey and evaluated the relationship between the presence of a chronic condition at the individual level, neighborhood socioeconomic status (SES), and self-rated health. We constructed multilevel models to evaluate the relationship between the neighborhood SES index and self-rated health for people with and without chronic conditions, after adjustment for other individual characteristics.

Results. Having a chronic condition was associated with substantially poorer self-rated health among participants in a deprived area than among those in a more advantaged area.

Conclusions. Residence in a disadvantaged neighborhood may be associated with barriers to the management of a chronic condition. Further work is needed to identify the specific characteristics of disadvantaged areas associated with poorer self-rated health for adults with chronic conditions.

Residence in a socioeconomically deprived neighborhood has been linked to all-cause mortality,1^–3 functional decline,4 poorer health status,5^,6 and higher incidence and prevalence of chronic conditions such as diabetes, cardiovascular disease, and cancer.2^,7^–16 People living in deprived neighborhoods are likely to experience multiple dimensions of poor environmental and social quality, including higher-priced yet lower-quality foods, high crime rates, poor-quality housing, limited transportation, toxic environments, and lower social cohesion and social support, all of which may contribute to poorer health.13^,17^–22 Adults with chronic conditions may be particularly vulnerable to these dimensions of neighborhood deprivation.

Models of chronic disease management, such as the Chronic Care Model23^–25 and the Disablement Framework,26 highlight the importance of community resources to the management of chronic conditions. Yet research on chronic conditions tends to emphasize clinical care, the health care system, and individual factors, and only infrequently examines the role of the neighborhood environment in the management of chronic disease.

There are, however, several mechanisms through which the neighborhood context may differentially affect the health of people with chronic conditions. Among adults with conditions such as diabetes, cardiovascular disease, arthritis, and asthma, adequate disease management often requires continuous clinical follow-up, self-care, and complex medication, dietary, and exercise regimens,24 all of which may be influenced by neighborhood factors such as available health care, access to exercise facilities and nutritious foods, and environments otherwise conducive to self-management. Thus, the characteristics of local areas, such as limited availability or accessibility of health services, infrastructure deprivation, environmental stressors, and social interactions that promote an unhealthy lifestyle19^,20^,27 may be associated with greater reductions in health status among adults with chronic conditions than among those without a chronic condition.

We conducted a cross-sectional analysis of the 2000–2001 Los Angeles Family and Neighborhood Survey (LAFANS) to examine whether the neighborhood socioeconomic environment contributes to differences in self-rated health among persons with and without a chronic condition. We chose to evaluate self-rated health because it is closely associated with several health outcomes, including morbidity28^,29 and mortality,30^,31 and determinants of health ratings have been shown to differ between people with and without chronic conditions.32^–34 We hypothesized that, independent of individual income or education, lower neighborhood socioeconomic status would be associated with lower self-rated health and that the association would be strongest for persons with a chronic condition.

METHODS

For the analyses, we used data from LAFANS Wave 1, a longitudinal study of families in a stratified probability sample of census tracts in Los Angeles County conducted in 2000–2001. The design of LAFANS is presented elsewhere.35^,36 Briefly, 1652 census tracts in Los Angeles County were stratified on the basis of the percentage of people living in poverty as obtained from 1997 estimates. Census tracts were classified as very deprived (90%–100% of residents living in poverty), deprived (60%–89%), and not deprived (1%–59%). In a representative sample of 65 tracts (20 very deprived, 20 deprived, and 25 not deprived), 40 to 50 dwelling units were sampled at random, with an oversample of households with children. Within each household, LAFANS randomly selected 1 adult (aged 18 years or older), who was interviewed in person. These analyses include data only from the randomly sampled adults in the LAFANS Wave 1 cohort.

The main predictors in the analysis were the socioeconomic status of the neighborhood of residence and the presence of a chronic condition. We assigned each tract a neighborhood socioeconomic status (SES) index.3 The SES index is the unweighted average of 5 census variables (percentage of individuals 25 years or older without a high school degree, median family income, median home value, percentage blue collar individuals, and percentage unemployed individuals), with the direction reversed for some variables; it was constructed from census tract–level data obtained from Summary File 3 of the 2000 US Census.37 We used an unweighted sum because a principal components analysis indicated that all 5 components contributed equally to the first factor, which accounted for 93% of the variance. Both the composite and the individual census variables were evaluated in separate models, but because we were interested in assessing multiple dimensions of the neighborhood socioeconomic environment and the findings did not differ appreciably, we used the SES index in these analyses.

The SES index was evaluated in separate analyses as either a continuous variable or in the 3 categories from the original sampling frame (“very deprived,” “deprived,” and “not deprived” census tracts). The results were similar for the 2 strategies; we present the results obtained from the second strategy. To test alternative definitions of “neighborhoods,” we constructed the SES index scores at both smaller (census block group) and larger (census tract) levels and compared final versions of the models with these 2 different definitions. Because we found no substantial difference in the results and prior literature suggests a high correlation between census block and census tract indictors,38 we present the census tract–level analyses.

The primary individual predictor was a report of a physician’s diagnosis of 1 or more chronic conditions, as determined by the question, “Has a doctor ever told you that you have … ?” followed by a list of conditions that included hypertension, arthritis, diabetes, and a chronic lung problem (asthma, chronic bronchitis, or chronic obstructive pulmonary disease). These conditions, which were included in analyses either individually or as an unweighted sum of the conditions, were the ones most commonly reported by study participants and are conditions that generally require substantial self-care.

Because depression is associated with health status and quality of life,39 we also evaluated it as a separate covariate in the regression model. Depression was defined as either ever having received a physician’s diagnosis of “major depression” or screening positive for depressive symptoms on the short form of the Composite International Diagnostic Interview.40 Because not all the randomly sampled adults in the cohort underwent screening for depression, we present separate results that include depression as a covariate in the portion of the sample that was screened for depression.

The dependent variable was self-rated health, which was measured by the question, “How would you rate your overall health?” Response categories were poor, fair, good, very good, and excellent. Poorer self-rated health has been associated with mortality and functional limitations in longitudinal studies.41 Unadjusted analyses were conducted using analyses of variance (repeated-measures analysis of variance [ANOVA]) and t tests. We evaluated self-rated health as a dichotomous variable (fair or poor vs good, very good, or excellent) and constructed weighted logistic regression models. Because self-rated health has an ordinal response scale, we conducted a sensitivity analysis that evaluated the outcome as an ordered categorical variable (poor or fair, good, very good, and excellent). The poor and fair categories were combined because only 4% of the overall sample rated their health as poor. The score test42^,43 supported the proportional odds assumption (χ²₃₈ = 47, P = .15). Multilevel ordered logistic models with random intercepts were constructed using Predictive Quasi Likelihood approximation procedures in HLM 5 software (Scientific Software International, Lincolnwood, Ill). Calculation of the intraclass correlation coefficient in the null (or empty) model suggested that 16.7% of the variance in self-rated health was between census tracts. After control for sociodemographic and clinical variables, the variance in self-rated health between census tracts was reduced to 10.8%.

At the individual level, the model included the main predictor (the presence of a chronic condition) and individual covariates, including demographic characteristics (age, gender, race/ethnicity, household income, education, immigrant status), body mass index, and health behaviors (smoking, alcohol use, and a physician visit in the prior year). A separate model also included depression as a covariate. Many of these characteristics have been associated with health status in previous research.32^,39^,44^,45 At the tract level, the main predictor was the neighborhood SES index. The models also included cross-level interactions between the neighborhood SES index and the presence of a chronic condition. Separate models were also constructed for each of the most common chronic conditions (hypertension, arthritis, diabetes, and chronic lung disease); these models included an interaction term between the neighborhood SES index and the chronic condition. Another model included an unweighted sum of chronic conditions. The logistic models included sampling weights that take into consideration both nonresponse and the over-sample of poor households and households with children.36

A potential confounder in research on neighborhood effects is “residential selection”: people in better health may choose to live in more advantaged neighborhoods. Although we could not directly assess the effect of selection on health outcomes in these cross-sectional analyses, we were able to evaluate health ratings among people with the greatest residential stability (i.e., those who had lived at the same address for 5 or more years). We also evaluated interaction terms between the neighborhood SES index and either individual income or individual education. Because these individual-level interactions were not significant and did not appreciably alter our findings, they are not presented in our final results.

We derived the relative risks and the 95% confidence intervals by bootstrapping with replacement over 1000 repetitions.46^,47 The expected values for each category of chronic condition and neighborhood poverty were then calculated. The analyses were performed using the statistical packages HLM 5, SAS version 9.1 (SAS Institute Inc, Cary, NC), and Stata version 9.0 (Stata Corp, College Station, Tex).

RESULTS

The sample comprised 2536 adults (response rate=70%), 848 of whom reported 1 or more of the chronic conditions of interest. Compared with the randomly sampled adults included in LAFANS, nonrespondents did not differ by race/ethnicity, gender, income, or education but were more likely to be the head of household.36 Older people, women, Whites, and African Americans were all more likely to report a chronic condition (Table 1 ▶). People with chronic conditions also had lower educational attainment and annual household income. Compared with those who did not report a chronic condition, people with 1 or more chronic conditions were not more likely to live in the poorest census tracts, but in the aggregate, their tracts of residence had slightly lower median SES index scores and they were more likely to have lived in the same residence for 5 years or longer. Those with chronic conditions had higher rates of reporting poor or fair health than did those without a chronic condition (38% vs 14%, P<.001).

TABLE 1—

Sample Characteristics of Participants and Census Tracts in the Los Angeles Family and Neighborhood Survey: Los Angeles, Calif, 2000–2001

	No Chronic Conditions (n = 1688)	Any Chronic Condition (n = 848)	P
Demographic characteristics
Mean age, y (SD)	36 (12)	46 (17)	<.001
Female, %	55	66	<.001
Race/ethnicity, %
White	24	28	.03
Latino	60	47	<.001
African American	7	14	<.001
Asian/Pacific Islander	8	7	.46
Immigrant, %	62	46	<.001
Education, %
None–8th grade	9	14	<.001
9th–11th grade	25	24	.92
High school graduate	46	47	.99
College graduate	20	15	.001
Income, $
0	10	27	<.001
1–20 000	31	28	.13
20 001–40 000	26	19	<.001
>40 000	33	26	<.001
Health-related characteristics
Mean body mass index, kg/m2 (SD)	25 (5)	29 (7)	<.001
Any chronic condition, %
Hypertension	. . .	56	. . .
Arthritis	. . .	34	. . .
Diabetes	. . .	22	. . .
Asthma or chronic obstructive pulmonary disease	. . .	24	. . .
Smoking status, %
Current smoker	15	17	.21
Past smoker	21	30	<.001
Never smoked	64	53	<.001
Binge drinking in past 30 days, %	7	10	.05
Depression, a %	8.8	20.6	<.001
Health service use, %
Doctor visit in past 12 mo	70	84	<.001
Hospitalization in past 12 mo	13	25	<.001
Unadjusted self-rated health, %
Poor	1	9	<.001
Fair	13	29	<.001
Good	31	32	.77
Very good	28	19	<.001
Excellent	27	11	<.001
Census tract characteristics
Median socioeconomic index score (95% CI)	−0.58 (−0.63, −0.49)	−0.60 (−0.65, −0.50)	.04
Census tract SES index, %
Not deprived	26	22	.18
Deprived	24	26	.13
Very deprived	50	52	.87
Residence in census tract ≥5 years, %	36	49	<.001

Note. CI = confidence interval; SES = socioeconomic status.

^aOf participants (n = 1838) screened with the Composite International Diagnostic Interview.40

Characteristics of the neighborhoods of residence of the study participants are presented in Table 2 ▶. Participants were divided into very deprived, deprived, and not deprived census tracts. Across these 3 groups, we present the median SES index score and the 5 census variables that comprise the SES index.

TABLE 2—

Characteristics of Census Tracts (From 2000 US Census) of Study Participants in the Los Angeles Family and Neighborhoods Survey: Los Angeles, Calif, 2000–2001

	Census Tract Status
	Very Deprived	Deprived	Not Deprived
No. of study participants	1339	652	632
No. of census block groups	128	52	46
No. of census tracts	37	15	13
Residence in neighborhood ≥5 years, %	37	40	49
Median socioeconomic index (range)	−0.85 (−1.72–0.37)	−0.094 (−1.44–0.42)	0.84 (−0.81–2.42)
Adults 25 years or older without a high school degree, %	62.2	31.9	8.5
Median family income, $ (range)	27 516 (13 750–61 917)	41 293 (16 619–67 561)	78 016 (23 615–145 361)
Median home value, $ (range)	147 900 (74 300–220 000)	178 000 (96 600–242 100)	274 100 (175 700–891 700)
Employed residents with blue collar occupations, %	62.2	43.0	20.8
Unemployed residents, %	13.6	8.1	5.8
Residence in census tract > 5 years, %	50.7	50.9	59.6
Chronic conditions, %
None	67	66	72
Any	33	34	28
Hypertension	20	20	16
Osteoarthritis	12	12	10
Diabetes mellitus	8	8	5
Asthma/chronic obstructive pulmonary disease	9	10	10
No. of chronic conditions
1	21	24	20
2	8	8	6
3	3	2	2
4	0.9	0.6	0.5
Mean body mass index, kg/m2	28	27	25
Depression, a %	15	17	10

^aOf participants (n = 1838) screened with the Composite International Diagnostic Interview.40

In multivariate analyses, having a chronic condition and the interaction terms between having a chronic condition and living in a deprived or very deprived census tract were all associated with lower self-rated health (Table 3 ▶). Other characteristics associated with lower health ratings were male gender, older age, immigrant status, less education, lower income, body mass index more than 30 kg/m², current tobacco use, binge drinking, and visiting a health care provider in the prior year.

TABLE 3—

Weighted Odds Ratios for Reporting Good, Very Good, or Excellent Health: Los Angeles Family and Neighborhoods Survey, Los Angeles, Calif, 2000–2001

	Full Model Without Depression, Odds Ratio (95% CI)	P	Full Model With Depression, a Odds Ratio (95% CI)	P
Age, y
< 30 (reference)	1.00		1.00
30–39	0.78 (0.63, 0.96)	.02	0.78 (0.63, 0.96)	.02
40–49	0.65 (0.52, 0.82)	<.001	0.65 (0.51, 0.82)	<.001
> 50	0.64 (0.50, 0.83)	.001	0.63 (0.49, 0.81)	<.001
Female	1.23 (1.04, 1.44)	.01	1.21 (1.03, 1.43)	.02
Race/ethnicity
White (reference)	1.00		1.00
Latino	0.81 (0.63, 1.04)	.09	0.80 (0.62, 1.03)	.10
African American	0.91 (0.67, 1.23)	.54	0.90 (0.67, 1.22)	.52
Asian/Pacific Islander	1.07 (0.76, 1.50)	.74	1.02(0.73, 1.43)	.91
Immigrant	0.66 (0.49, 0.74)	<.001	0.59 (0.48, 0.73)	<.001
Education
College (reference)	1.00		1.00
Some college	0.61 (0.48, 0.76)	<.001	0.61 (0.48, 0.76)	<.001
High school graduate	0.38 (0.29, 0.51)	<.001	0.39 (0.29, 0.52)	<.001
Less than high school	0.26 (0.18, 0.39)	<.001	0.26 (0.18, 0.38)	<.001
Income, $
40 000 (reference)	1.00		1.00
20 001–40 000	0.73 (0.59, 0.90)	.07	0.73 (0.58, 0.90)	.004
1–20 000	0.54 (0.43, 0.68)	<.001	0.55 (0.44, 0.69)	<.001
0	0.41 (0.32, 0.54)	<.001	0.42 (0.32, 0.55)	<.001
Body mass index, kg/m2
<25 (reference)	1.00		1.00
25–30	0.87 (0.72, 1.05)	.14	0.88 (0.73, 1.05)	.16
>30	0.73 (0.60, 0.90)	.005	0.74 (0.60, 0.90)	.003
Smoker
Never (reference)	1.00
Current	0.78 (0.63, 0.97)	.02	0.81 (0.65, 1.01)	.06
Binge drinker	0.81 (0.69, 0.95)	.002	0.81 (0.69, 0.96)	.01
Visit to a primary care provider in the past year	0.78 (0.65, 0.94)	.008	0.80 (0.67, 0.97)	.02
Any chronic conditionb	0.50 (0.38, 0.63)	<.001	0.52 (0.41, 0.67)	<.001
Neighborhood socioeconomic index
Nondeprived tracts (reference)	1.00		1.00
Deprived tracts	0.88 (0.67, 1.16)	.37	0.91 (0.69, 1.21)	.52
Very deprived tracts	0.90 (0.68, 1.19)	.47	0.91 (0.69, 1.20)	.50
Interaction between neighborhood deprivation and presence of a chronic condition
Chronic condition × nondeprived	1.00		1.00
Chronic condition × deprived	0.68 (0.47, 0.98)	.04	0.66 (0.46, 0.97)	.03
Chronic condition × very deprived	0.59 (0.43, 0.81)	.001	0.60 (0.44, 0.82)	.002
Depression	. . .		0.53 (0.42, 0.66)	<.001

^aLimited to those screened using the Composite International Diagnostic Interview.40

^bChronic conditions are hypertension, arthritis, diabetes, and asthma/chronic obstructive pulmonary disease.

Table 4 ▶ shows the adjusted percentages of participants expected to report good, very good, or excellent health, categorized by chronic condition and census tract status. The reference group for all comparisons was those without a chronic condition who resided in nondeprived tracts. Among people without a chronic condition, the adjusted predicted percentage of people who reported their health as good, very good, or excellent was 92.4% for those residing in nondeprived census tracts, 80.2% for those in deprived tracts, and 73.2% for those in very deprived tracts. People with 1 or more chronic conditions had lower predicted ratings: good, very good, or excellent health was reported by 88.4% of those in nondeprived tracts, 63.3% of those in deprived tracts, and 52.4% of those in very deprived tracts. The differences between the adjusted means for persons with a chronic condition (compared with those without one) were significantly higher for residents in deprived or very deprived areas than for those of nondeprived areas. With the exception of asthma, similar relationships were observed for the each of the chronic conditions.

TABLE 4—

Adjusted Percentage of Respondents Expected to Report Good, Very Good, or Excellent Health, by Chronic Condition and Neighborhood Socioeconomic Status Index: Los Angeles Family and Neighborhood Survey, Los Angeles, Calif, 2000–2001

	% of Respondents by Census Tract Status
	Not Deprived	Deprived	Very Deprived
No chronic condition	92.4	80.2	73.2
Any chronic condition	88.4	63.3a	52.4b
Hypertension	83.2	63.8a	55.3b
Arthritis	81.2	61.5a	52.8b
Diabetes	79.0	65.4	51.2b
Asthma/COPD	63.9	52.7	47.4

Note. COPD = Chronic obstructive pulmonary disease. Adjusted for age, gender, race/ethnicity, income, education, and immigrant status.

^aFor this chronic condition, difference between those living in nondeprived versus deprived tracts is greater for people with the condition than for those without the condition (P < .05).

^bFor this chronic condition, difference between those living in nondeprived versus very deprived tracts is greater for people with the condition than for those without the condition (P < .05).

Sensitivity analyses conducted with multilevel ordered logistic regression models found the same associations between self-rated health and area of residence, the presence of a chronic condition, and the interaction term between the 2. Similar results were also observed when we evaluated the number of chronic conditions as opposed to just the presence of a chronic condition and for analyses restricted to people who had lived in the census tract for 5 or more years.

DISCUSSION

In this analysis of a population-based study of adults in Los Angeles County, for both people with a chronic condition and those without, the difference in self-rated health was greater among those living in low-SES census tracts than among those in high-SES census tracts. Not only were the presence of a chronic condition and lower individual SES (e.g., income and education) associated with lower health status, but having a chronic condition was associated with substantially poorer self-rated health among people in a deprived area than among those in a more advantaged area. Participants with chronic conditions had substantially lower individual income and were less educated than were the other study participants. Although a lower neighborhood SES index was associated with poorer health status after adjustment for individual income and education, it is important to consider that their higher level of socioeconomic disadvantage may have made participants with chronic conditions particularly vulnerable to neighborhood deprivation.

Chronic stress (related to crime, poor housing quality, and infrastructure deprivation) and lower or poorer availability of resources (such as access to health care, food availability, and transportation) may contribute to the larger negative association between neighborhood socioeconomic environment and health for adults with chronic conditions. Among adults with diabetes, for example, chronic stress has been associated with poorer glycemic control through 2 mechanisms: health behaviors, such as lower rates of adherence to medication or less physical activity, and neurohormonal pathways.48

An important mediator of the relationship between neighborhood deprivation and lower self-rated health may be functional status. A recent study, for example, indicates that the impact of functional limitation on quality of life is more that 4 times that of the chronic condition by itself.49 Inadequate neighborhood resources, including access to health care, safe places to exercise, healthy foods, and transportation, may directly contribute to functional impairment. These factors may pose a greater barrier for adults with chronic conditions, particularly those with existing disability, and lead to reduced physical activity, poorer dietary patterns, and lower rates of visits to health providers. Allostatic load, the cumulative burden associated with the body’s adaptation to chronic stress, may be another important mechanism through which neighborhood characteristics influence health outcomes for adults with chronic conditions.50

We observed different relationships for some of the chronic conditions, most notably chronic lung conditions. We cannot be certain why, for conditions such as asthma, chronic obstructive pulmonary disease, and chronic bronchitis, there is no difference in self-rated health status according to the neighborhood SES index; other factors may be involved. For example, residence near a freeway or other determinants of local air quality may be more important predictors of self-rated health among people with chronic lung problems.51 Another explanation for these findings may be the smaller sample size in this group; still another could be that asthma, COPD, and chronic bronchitis are heterogenous conditions, yet differences in their individual health ratings are masked by grouping them into 1 category.

Among the limitations of these analyses are that we used cross-sectional data and have no information on change in health status or other health outcomes. Additionally, people in worse health may be more likely to move to or remain in more deprived neighborhoods. However, restricting the analyses to people who had lived in the same area for 5 or more years and controlling for duration of residence in these census tracts produced no significant change in the results. Another limitation is that although we tested for alternate definitions of neighborhoods using both tract and block group data, census-derived characterizations of neighborhoods may not reflect aspects of the social and physical environment that influence the health and behaviors of individuals. However, prior work from LAFANS suggests that census-tract definitions of neighborhood size are highly correlated with respondents’ reports of the size and boundaries of their neighborhoods.52 Still, further work is needed to better understand the characteristics of areas that influence the health-related experiences of residents.

Another limitation is that the chronic conditions were self-reported, with no independent verification of the results. Data on the validity of self-reported data on chronic conditions suggest that accuracy of self-report (with either the medical record or physical examination used as the gold standard) varies by condition. Several validation studies have compared methods of identifying people with the chronic conditions included in our analyses. Although self-reported diabetes shows the strongest agreement with the medical record, there is lower, but generally good, agreement between self-report and medical record diagnoses of hypertension.53 In contrast, there is substantial underreporting of asthma54 and osteoarthritis.55^–57 Nonetheless, we found similar patterns for the association between neighborhood deprivation and health ratings for hypertension, arthritis, and diabetes.

Another potential limitation is that certain groups, including poorer or less educated people, may be less likely to report a chronic condition owing to lower rates of diagnosis or a lack of awareness of the condition. A recent study suggests that less education and more co-morbid conditions are associated with under-reporting of conditions including hypertension and diabetes,58 but the absolute differences observed for different levels of education were relatively small. Misclassification of people in more deprived census tracts with undiagnosed disease, however, may have led to an overestimation of the magnitude of the difference in self-rated health between those with and those without a chronic condition in deprived and very deprived tracts.

Our findings of a differential in health status for those with and those without a chronic condition by level of neighborhood deprivation suggest a need to identify the specific characteristics of the built and social environment that are associated with substantially lower health ratings among persons with chronic conditions in the most deprived neighborhoods. An important step is for researchers and health care organizations to systematically collect information on the environmental and social factors that may serve as barriers to those with chronic conditions. An understanding of how neighborhood influences health can enhance efforts to improve health through urban planning, housing policies, and modifying the food resource environment.

In the clinical setting, enhanced awareness of the association between neighborhood factors and health for people with chronic conditions may help health care providers identify adults who are most likely to be affected by disadvantaged neighborhood environments; they can then tell them where and how to obtain important services, such as better food, improved transportation, and low-cost exercise facilities. Further work is also needed to clarify the specific socioeconomic forces and structural characteristics that influence health outcomes in general and those that contribute to the differential observed for people with chronic conditions. Previous studies have shown that for adults with diabetes, intensive clinical and behavioral interventions benefit those who are less educated59 or have low literacy levels60 more than those who are more educated or literate. We have yet to determine whether intensive clinical, behavioral, or policy interventions can modify neighborhood effects on people with chronic conditions.