Despite the growing ability of societies to extend life and prevent disease, health disparities persist (1). As researchers and policy makers have tried to answer the question of why some societies are healthier than others, attention has focused on individual-level factors such as biology (eg, genomics), psychology (eg, coping strategies), community (eg, place of residence, work environment), and society (eg, social and economic policies).

Although a strong association is well-established between lower levels of individual socioeconomic status (SES) and poor health outcomes from many diseases throughout the developed world (1), studies increasingly suggest that community social determinants (ie, the socioeconomic environment of a person's neighborhood) may influence health (2-7), independent of the person's SES (4,8-11). Regardless of personal socioeconomic position (12,13), characteristics of the physical, social, and service environments of neighborhoods and communities may influence the lives of people who live there (14), including their health (3).

Many studies that have examined the association of both individual and community social determinants with illness and death have been conducted primarily in large urban areas (5,15). The purpose of this study was to examine the relationship between individual and community SES with health-related quality of life (HRQOL) by race in a geographically diverse sample of community-dwelling non-Hispanic whites and African Americans in a southeastern state. Study participants were recruited from a representative cohort of adult primary care patients followed in 17 practices from rural and urban areas of North Carolina. Analyses were conducted on the entire sample and on subgroups of people who self-reported the 2 most common chronic conditions and leading causes of disability, arthritis and cardiovascular disease (CVD) (16).

Characteristics of study participants are displayed in Table 1. In the total sample, African Americans were significantly more likely to be younger, be female, and have a higher BMI, less education, and worse HRQOL (except for CDC HRQOL poor mental days) compared with whites. The block-group poverty-level tertiles were calculated separately for whites and African Americans and reflected higher poverty levels in African Americans. Patients within each of the disease-specific subgroups were significantly more likely to be older, have a higher BMI, and be less educated than the patients not in the disease subgroup. The disease-specific subgroups had significantly lower HRQOL than the whole group (with the exception of MCS for the CVD/hypertension group). The mental HRQOL showed an overall less negative effect of disease than did physical HRQOL.

Table 2 presents the adjusted means for the 5 HRQOL outcomes by education and poverty level for the white total group and the arthritis and CVD/hypertension subgroups. The highest level of education and the lowest poverty level are the referent categories. The adjusted means varied more within categories of education than within categories of poverty. For example, in the white total group, SF-12v2 PCS increased with education from 38.1 to 46.6 whereas PCS increased with decreasing poverty level from 43.8 to 46.0.

For the PCS in the total white group, groups with both the middle and low levels of education had significantly lower mean scores (poorer outcomes) than the group with the highest level of education, after adjusting for poverty level, age, BMI, sex, and family practice site. Also, for the PCS, both the middle and high poverty-level groups had significantly lower mean scores than the lowest poverty-level group, adjusting for education level and the other covariates (Table 2). In the whole population, mean SF-12v2 MCS scores were significantly poorer in the middle- and low-education groups, but we found no significant differences among the poverty-level groups.

All 3 of the CDC HRQOL healthy days measures showed significantly higher mean scores (poorer outcomes) in both the lowest education-level groups and the highest poverty-level groups, adjusting for each other and age, BMI, sex, and family practice setting (Table 2) in whites. Similar findings were identified in the subgroup analyses for participants who self-reported arthritis and those who self-reported CVD/hypertension (Table 2).

Table 3 shows the adjusted means for the 5 health status measures by education and poverty level for the African American total group and the 2 disease subgroups (arthritis and CVD/hypertension). The smaller sample size, particularly for African Americans in the arthritis subgroup, resulted in fewer findings of significant differences, but the general trend of the findings follows that seen for whites. Low individual levels of education and high poverty levels were significantly associated with poorer status in the SF-12v2, MCS, more poor physical and mental health days, and more limited activity days in the total group.

In this community-dwelling sample of patients living in rural and urban areas of North Carolina, both individual (education) and community (poverty level) SES measures were associated with physical and mental measures of HRQOL. A 1-day difference is considered meaningful for the CDC HRQOL healthy days scores at the individual level (26). All 3 healthy days scores had a difference of at least 1 day for both the individual and the community SES measures. The minimum clinically important difference for the SF-12v2 PCS and SF-12v2 MCS in a chronic disease has been reported as 1.26 and 2.28, respectively (27). In our study, the differences for the PCS and MCS exceeded these thresholds for both education and community poverty level. Thus, all differences reported here appear to be meaningful.

Overall, the magnitude of difference in HRQOL was greater for education compared with community poverty level. For example, for whites in any group, changes in the SF-12 PCS scores, going from low to high SES, ranged from 8.1 to 9.6 for education and from 1.7 to 2.2 for community poverty level. Although the independent effects of community socioeconomic context may be relatively small, the overall importance of community socioeconomic context to individual health may be more substantial, both because it affects all people in a community and because community context shapes the person's sense of control and all individual-level variables.

The association between lower levels of individual SES and poorer health outcomes has been documented repeatedly in various parts of the developed world (28-32). Individual-level SES can be examined according to a number of variables, including formal education level, which was used in this study, income, occupation, home ownership, race, and marital status. Formal education level is, in part, a marker for behavioral variables (22,33,34), such as self-management, problem-solving abilities, efficiency in use of medical services, capacity to cope with stress, social skills, psychological status, and economic skills, which singly or together enable people to more effectively prevent, overcome, or cope with adversity (35).

Although the body of public health research relating community factors to patterns of health and disease is well-established (36), the underlying hypothesis of this prior work proposes that factors operating at the level of the communities may affect the health outcomes of individuals. Studies in recent years have suggested that area-level or community-level variables may provide information that is not captured by individual-level variables (2,37,38). In our study, using census-based block-group poverty levels as a community-level SES indicator, we found that the community indicator was significantly associated with HRQOL after controlling for the individual indicator.

Our study has several limitations. Our data are all self-reported, and we had only education level as our measure of individual SES. Although we lacked data on income and occupation as individual SES measures, many studies have shown education to be a strong marker of individual SES. The effects of community socioeconomic level on health may be underestimated in our study because we used a crude measure of community SES, block-group poverty level from the census. Future research should begin to include information about self-defined communities or at least purposefully delineate community boundaries to more closely match theoretical constructs.

The US health care research agenda places priority on reducing disparities in health outcomes among people from different socioeconomic and ethnic groups by examining the mechanisms believed to affect health. Now, more than ever, the interest in a more explicit investigation of the complex issues about health disparities is increasing (39). However, the statement in Healthy People 2010 recognizing that communities, states, and national organizations will need to take a multidisciplinary approach to achieving health equity is often overlooked (40). The development and communication of effective actions to reduce health disparities depends on clarifying relationships between community variables (social context), individual variables (social position), and health outcomes.

This project was supported by a grant from the National Institute of Arthritis and Musculoskeletal and Skin Diseases Multidisciplinary Clinical Research Center, Rheumatic Diseases: P60-AR49465-01.

We thank Jennifer Milan Polinski and Shannon Currey for assistance in the establishment of the cohort. We thank the following participating family practices in the North Carolina Family Medicine Research Network for their assistance: Black River Health Services, Burgaw; Bladen Medical Associates, Elizabethtown; Blair Family Medicine, Wallace; Cabarrus Family Medicine, Concord; Cabarrus Family Medicine, Harrisburg; Cabarrus Family Medicine, Kannapolis; Cabarrus Family Medicine, Mt. Pleasant; Chatham Primary Care, Siler City; CMC-Biddle Point, Charlotte; CMC-North Park, Charlotte; Community Family Practice, Asheville; Cornerstone Medical Center, Burlington; Crissman Family Practice, Graham; Dayspring Family Medicine, Eden; Family Practice of Summerfield, Summerfield; Goldsboro Family Physicians, Goldsboro; Henderson Family Medicine Clinic, Henderson; Lumber River Family Practice, Lumberton; Moncure Community Health Center-Piedmont Family Health Services, Moncure; Orange Family Medical Group, Hillsborough; Person Family Medical Center, Roxboro; Pittsboro Family Medicine, Pittsboro; Prospect Hill Community Health Center, Prospect Hill; Robbins Family Practice, Robbins; and Village Family Medicine, Chapel Hill.