Recent research suggests that the characteristics of the residential neighborhood may have an impact on risk of ischemic heart disease.1 Past efforts in this domain mainly investigated the association between neighborhood socioeconomic position and ischemic heart disease after controlling for individual socioeconomic characteristics.2–9 This effect may be due in part to the impact of material infrastructures and services on health-damaging behavior and healthcare utilization1 and to environmental hazards (air pollution, noise exposure).10,11 However, besides effects of the physical environment, the literature on social support and ischemic heart disease,12 including a study showing a protective effect of neighborhood-based social support,13 suggests that structural characteristics that shape social functioning and the level of social cohesion in the neighborhood may be relevant to ischemic heart disease risk. Social interaction patterns in the neighborhood may be stressful,14 may condition the quality and amount of social support individuals receive from their neighborhood,13,15 and may harm their psychologic well-being.16 Such causal pathways could influence health behavior and risk of ischemic heart disease.
Overall, the literature provides little support to assess whether neighborhood characteristics that shape social interaction patterns are associated with ischemic heart disease independent of individual and neighborhood socioeconomic position. We investigated the independent effects of neighborhood socioeconomic position and residential stability (a well-known determinant of neighborhood social interaction patterns) on ischemic heart disease risk. In line with the early work of the Chicago school,17,18 it is widely accepted that high residential instability or population turnover in the neighborhood operates as a barrier to the development of interpersonal bonds and local associational ties, disrupting a community's network of social relations.19–21 Our hypothesis was that the smaller the percentage of stable residents in the local neighborhood, the lower the level of social cohesion in the environment15 and the lower the amount of social support individuals may receive from their community. In turn, emotional, informational, and instrumental support, and social control of close neighbors encouraging one to comply with medical recommendations, may affect health behavior and coronary health.12,13,22,23 We attempted to isolate the effect of residential stability on ischemic heart disease risk, an effect that might otherwise be partly incorporated in the neighborhood socioeconomic effect.
Regarding ischemic heart disease outcomes, some neighborhood studies have focused on incidence,3,4,8,24 whereas others considered mortality7,25–29; however, none has investigated these 2 distinct end points comparatively. Moreover, to our knowledge, no previous study has attempted to disentangle neighborhood from individual effects on survival after acute myocardial infarction (MI). Considering incidence, 1-day case-fatality, and long-term survival after MI, we examined the extent to which neighborhood effects on mortality resulted from disparities in incidence or in survival after MI. Based on the hypothesis that social support may be particularly critical in the period immediately after MI, and on previous literature suggesting that social support has a stronger role in enhancing survival after MI than on incidence,23 we expected a larger effect of residential stability on survival than on incidence (because residential stability effects may be mediated by the availability of social support).
Using Swedish data, our investigation expands on previous literature by examining whether residential stability as a possible determinant of social interaction patterns influences ischemic heart disease risk beyond individual and contextual socioeconomic circumstances and whether dissimilar contextual effects operate at various stages of the disease process, ie, on incidence, 1-day case-fatality, and long-term survival after MI.
With the assistance of Statistics Sweden and the Swedish Center for Epidemiology, a longitudinal database (LOMAS), including all inhabitants in Scania, Sweden (approximately one million), was assembled. We used the personal identification number assigned to each person in Sweden to link the following data sources: 1) the exact spatial coordinates of the households in 1 January 1991 and 1996; 2) yearly information on individual income from 1975 to 1995; 3) education and occupation data from the 1970 population census; 4) date and causes of death from 1996 to 2002 from the National Mortality Register; and 5) diagnoses made at the hospital from 1987 to 2002 from the National Inpatient Register. The LOMAS project was reviewed and approved by the Swedish Regional Ethical Review Board in Lund and the Data Safety Committee at Statistics Sweden.
We constituted a cohort with baseline of 1 January 1996 comprising all 73,547 inhabitants from the 3 main cities of Scania (Malmö, Helsinborg, and Lund) reaching age 50 to 64 years in 1996. To consider individuals free of ischemic heart disease at baseline, we excluded 2289 individuals (3.11%) who had received a diagnosis of ischemic heart disease (codes 410–414 in the International Classification of Diseases [ICD], 9th revision) in the 9 years before follow-up (ie, since the beginning of the National Inpatient Register). Moreover, to assess the effects of residential instability on stable residents, we restricted analyses to individuals who had been living in the same local neighborhood for at least 5 years before follow-up; 19,174 individuals who had recently moved into the neighborhood (27% of the population) were excluded. Most of the remaining residents had lived in the same neighborhood for much longer than 5 years (79% for 10 years or more). The final database comprised 52,084 individuals.
We defined an incident event as either an ischemic heart disease death (assessed with ICD 9th revision as codes 410–414 or with ICD 10th revision as codes I20–I25 for the underlying or contributing causes of death) or a diagnosis of acute MI made at the hospital (ICD-9 code 410 or ICD-10 code I21). Ischemic heart disease deaths were also considered separately to define mortality outcomes.
Besides age and gender, we considered previous diseases of individuals, marital status, and information on individual socioeconomic position over the adult life, ie, education, occupation 25 years before baseline, and individual income averaged over the 20 years before baseline. Age was coded in 5-year categories (50–54, 55–59, 60–64 years). Regarding previous diseases, separate dummy variables indicated whether individuals had received hospital diagnoses of cancer (ICD-9 codes 140–239), diabetes (ICD-9 code 250), hypertension (ICD-9 codes 401–405), other heart diseases (ICD-9 codes 420–429), and cerebrovascular diseases (ICD-9 codes 430–438) in the 9 years before follow-up.
Marital status was assessed at baseline and dichotomized as married/cohabiting or living alone (ie, single, divorced, or widowed). Education was divided into 3 classes (≤7 years, 8–9 years, >9 years). Regarding occupation 25 years before, we distinguished between nonmanual, manual, and others. Rather than household income, only individual income was available (in 1975, 1980, 1985, 1990, and 1995). We converted each yearly income variable into a rank (from 1 to 100) to standardize income level across years and then computed the average of the 5 standardized income variables for each individual.
To investigate contextual effects on a local scale, we used the smallest administrative geographic units (NYKO parcels). To define contextual variables in neighborhoods with a minimum number of inhabitants, we used an algorithm to group parcels with few inhabitants with adjacent ones. This algorithm aggregates each parcel with less than 100 inhabitants age 50 to 79 years to the one of the adjacent parcels that 1) belongs to the same larger city area and 2) shares the longest common boundary. In this operation and in the calculation of neighborhood factors, we could consider only the 50- to 79-year age bracket of the population (because we were unable to geocode other individuals at the level of the local neighborhoods). In 1996, the median number of inhabitants of all ages in the resulting 648 neighborhoods was approximately 1050 (interquartile range: 720–1550), and the median number of individuals age 50 to 64 years retained in the final sample was 63 (interquartile range: 39–103). The median area size of neighborhoods was 0.12 km2 (interquartile range: 0.04–0.35).
The main contextual factors we considered were neighborhood socioeconomic position and residential stability. Neighborhood socioeconomic position was defined as neighborhood mean income of the 50- to 79-year-old residents at 1 year before baseline (neighborhood deprivation corresponding to a low socioeconomic position). Considering the neighborhood of all 50- to 79-year-old residents in 1996 and 5 years before, we computed neighborhood residential stability as the percentage of residents in 1996 who were living in the same neighborhood 5 years before.20 In the sample of individuals, the median proportion of 5-year stable residents in the neighborhood was 0.83 (interquartile range = 0.75–0.89; interdecile range = 0.65–0.93). Individual income and neighborhood factors were each divided into 4 equal-sized categories.
To take into account possible contextual confounders, we also considered neighborhood population density (number of 50- to 79-year-old residents per square kilometer) as a proxy of urban living conditions30 and straight-line distance from the exact place of residence to the city hospital (because survival may depend on access to emergency services).
To capture between-neighborhood variability in ischemic heart disease risk, we used multilevel Weibull survival models including a neighborhood-level random intercept. In a multilevel survival model, the between-neighborhood variance is the variance of a residual term measured on the log hazard ratio (HR) scale. Therefore, this variance is not expressed in any directly meaningful units. We aimed to facilitate assessment of the magnitude of between-neighborhood variations31–33 by expressing geographic heterogeneity on the hazard ratio scale with the interquartile hazard ratio16,30,34 and the median hazard ratio.31,34 The interquartile hazard ratio expresses the difference in ischemic heart disease mortality between the 25% of all individuals in neighborhoods with the lowest risk and the 25% of all individuals in neighborhoods with the highest risk.16 The median hazard ratio is the median value of the hazard ratio between the individual in the neighborhood with the lowest risk and the individual in the neighborhood with the highest risk when randomly picking out 2 individuals in different neighborhoods.31,34
Using Winbugs 1.4.135 to implement a Markov chain Monte Carlo approach, we estimated multilevel survival models separately for ischemic heart disease incidence and mortality. Observations were right-censored in case of death by another disease. We first estimated empty models, then included individual variables, and finally introduced the 2 contextual dimensions simultaneously. For all HRs, we calculated 95% credible intervals (CIs). Complementary analyses indicated that the city of residence had no impact on ischemic heart disease once the individual and contextual variables were accounted for.
To interpret possible differences in contextual effects between ischemic heart disease incidence and mortality, we used a multilevel survival model to investigate contextual influences on ischemic heart disease mortality among individuals who received a hospital diagnosis of MI between 1996 and 2002 (survival after MI among the 1080 MI cases). To gain additional insight, we split this outcome into 2 distinct components. First, we used a multilevel logistic model to compute odds ratios (ORs) for determinants of 1-day case-fatality (defined as death occurring the same day than the MI) among the 1080 MI cases. Second, we estimated a multilevel survival model for ischemic heart disease mortality among individuals still alive 28 days after MI (983 cases).
Population density and distance to the hospital were introduced in the various models to assess possible confounding effects.
Finally, as a sensitivity analysis, we reestimated the models for incidence, mortality, and survival in a population including both 5-year stable residents and people who had moved into their neighborhood over the past 5 years. Such models were further adjusted for the individual residential status—being a 5-year stable resident or not.
In our population, 2.9% of individuals (1507 of 52,084) had an incident ischemic heart disease event in the 7-year follow-up, and 1.1% (547 individuals) died of ischemic heart disease. Among the 1080 individuals with a hospital diagnosis of acute MI, 5.9% (64 individuals) died on the same day.
Multilevel survival models indicated that there were between-neighborhood variations in ischemic heart disease incidence and mortality (Table 1). Geographic variations were much larger for mortality than for incidence. We then introduced individual characteristics in the models. Having a previous diagnosis of diabetes, hypertension, other heart diseases, and cerebrovascular disease was associated with an increased risk of ischemic heart disease incidence and mortality (Table 2). Less educated persons, low-income individuals, and noncohabiting ones also had a higher incidence and mortality. Table 1 indicates that residual between-neighborhood variations were drastically reduced when the individual characteristics were included in the models. However, the between-neighborhood variance in models with individual variables only cannot be used to quantify possible neighborhood influences; part of the neighborhood effects is captured by the individual variables considered in the models.36
Considering the contextual variables, the correlation between neighborhood socioeconomic position and residential stability was moderate (correlation = 0.24, 95% CI = 0.17–0.31, with the 648 neighborhoods as units of analysis).
To assess whether ischemic heart disease risk was related to neighborhood characteristics, we first computed crude percentages of events. The unadjusted percentages of incident cases and deaths markedly increased with neighborhood deprivation and residential instability (Table 3). Residential stability resulted in greater disparities for mortality than for incidence (as suggested by the Cochran-Armitage trend test statistics given in the footnotes to Table 3). Larger differences were observed with the neighborhood socioeconomic variable than with residential stability for ischemic heart disease incidence, whereas the opposite was true for 1-day case-fatality of acute MIs. As shown with the Kaplan-Meier survival curves reported in Figure 1, differences in ischemic heart disease mortality among individuals still alive 28 days after MI were observed when stratifying either by neighborhood socioeconomic position or residential stability.
The 2 contextual variables were included simultaneously in the multilevel survival models for incidence and mortality. After adjustment for individual variables, ischemic heart disease incidence increased regularly with neighborhood deprivation (Table 4). Beyond effects of individual and neighborhood deprivation, ischemic heart disease incidence increased only weakly with residential instability. There was evidence of a stronger effect of residential stability on mortality than on incidence (the 95% CIs of the hazard ratios for residing in a neighborhood with a low vs high residential stability were almost nonoverlapping). After full adjustment, the dose–response effects of neighborhood deprivation and residential instability on ischemic heart disease mortality were of comparable magnitude (Table 4). Inclusion of individual and contextual variables into the multilevel models explained a large part of the between-neighborhood variability in incidence and mortality (Table 1).
To understand why a stronger residential stability effect was observed on mortality than on incidence, we estimated neighborhood effects on ischemic heart disease mortality after MI. As shown in Table 5, after individual-level adjustment, we observed no neighborhood socioeconomic influences but a strong and dose–response effect of residential stability, resulting in markedly higher ischemic heart disease mortality after MI in residentially unstable neighborhoods. However, 95% CIs were rather wide, partly due to the low number of MI patients included in this analysis. To gain additional insight, we split the post-MI mortality outcome into 2 distinct components, ie, 1-day case-fatality and ischemic heart disease mortality among individuals still alive 28 days after MI. For both of these distinct outcomes, we noted a strong and dose–response effect of residential stability with worse survival outcomes in residentially unstable neighborhoods but no impact of neighborhood socioeconomic position (Table 5).
The contextual effects observed on the various ischemic heart disease outcomes remained even after including population density and distance to the hospital in the models. We did not observe reduced odds of survival among people residing relatively far away from the city hospital (results not shown).
Similar results were obtained when reestimating the models for incidence, mortality, and survival in a population including both 5-year stable residents and those who had moved into their neighborhood within the past 5 years (data not shown in a table). For example, in the fully adjusted multilevel survival model for ischemic heart disease mortality, the hazard ratios for residing in neighborhoods with a midhigh, midlow, and low (vs high) residential stability were 1.17 (95% CI = 0.85–1.54), 1.40 (1.05–1.89), and 1.61 (1.20–2.10), respectively. Similarly, among the 1507 MI cases diagnosed at the hospital over the period, the odds ratios for 1-day case-fatality in neighborhoods with a midhigh, midlow, and low (vs high) residential stability were 2.12 (0.93–5.17), 2.74 (1.21–6.68), and 3.80 (1.61–9.54), respectively.
Our study expands on previous contextual epidemiology of ischemic heart disease by disentangling the impact of neighborhood deprivation and residential instability, and comparing contextual effects operating at the different stages of the disease process, ie, on incidence, 1-day case-fatality, and long-term survival after MI.
The main strengths of our study include the use of a large cohort of individuals, the geocoding of participants to local neighborhoods, the adjustment of models for past diseases and socioeconomic indicators throughout adulthood, and the fact that we assessed the effect of 5-year residential stability among people who resided in the neighborhood over the period.
However, there were limitations to our study. First, ischemic heart disease cases were not validated with a standard procedure,37 but were identified from the hospital and mortality registers. However, Sweden has a long tradition in the administration of registers, whose quality is regularly audited. Moreover, the existence of centralized national registers limits the risk of differential measurement bias38 that may affect findings on geographic variations. The validity of acute MI cases in the hospital register was judged to be acceptable for epidemiologic analysis.39 Less accuracy may be expected in the mortality register, but its use together with the hospital register ensures high coverage of the incident ischemic heart disease events.
Second, we had no information on traditional risk factors of cardiovascular diseases such as smoking, physical inactivity, overweight, hypertension, cholesterol, or family antecedents, and on more recently identified risk factors such as certain inflammatory and hemostatic variables.40 It is a drawback of the large database we had to use, because typical cohorts with precise information on risk factors would not be large enough to compare contextual influences on incidence and post-MI survival. However, traditional risk factors of ischemic heart disease should be viewed not as confounders, but as possible mediators of neighborhood effects. Various studies have shown that neighborhood socioeconomic position or more specific features of the environment are related to smoking behavior, physical activity, blood pressure, and weight status beyond effects of individual characteristics.1,41 Therefore, those risk factors, and the more recently identified inflammatory and hemostatic variables,42,43 qualify as possible mediators of neighborhood social environment effects on ischemic heart disease. The exact mediating role of each of these variables will have to be assessed in future research.
Third, we were not able to examine whether the residential stability effect was mediated by differences in social support available to the individuals. Finally, we could not exclude the possibility that the effect of residential instability on post-MI survival was attributable to differences in the severity of the MI. However, this may not be the most likely explanation because contextual effects were also found when restricting analyses to individuals still alive 28 days after MI who may have a more homogeneous baseline risk; moreover, it is not clear why residential instability would influence severity at incidence while having little impact on incidence.
In the previous literature, Sundquist et al reported an effect of neighborhood social disorganization24 and neighborhood “linking social capital”44 on ischemic heart disease incidence. However, the authors failed to adjust those effects for the well-documented impact of neighborhood deprivation. Kölegard45 attempted to disentangle the effects of material deprivation and social fragmentation as captured by the Townsend and Congdon indices on MI incidence. However, including a common component (home ownership), the 2 indices were too highly correlated (r = 0.87) to allow separation of their effects. Other studies have investigated area social capital or cohesion effects on cardiovascular disease mortality but were not specifically focused on ischemic heart disease.25,26
In our study, residential stability was selected as an objective determinant of the density of social connections between residents in the neighborhood.17,19,20 Additional analyses indicated that the individual propensity to move between 1991 and 1996 was not associated with individual income but regularly increased with neighborhood deprivation (results not shown). Therefore, residential instability could be seen to some extent as a contextual mediator of the neighborhood deprivation effect. Despite this relationship, we aimed to disentangle the effects of the 2 dimensions, because residential stability is not entirely determined by neighborhood socioeconomic position, and we were interested in isolating the residential stability effect from the overall neighborhood socioeconomic effect.
Comparing contextual effects on ischemic heart disease incidence and mortality, a striking difference was seen for the residential stability effect, which was stronger for mortality than for incidence. This pattern was attributable to the fact that 1-day case-fatality of MIs was markedly higher and long-term survival after MI markedly reduced in residentially unstable neighborhoods (those survival outcomes showed no association with neighborhood deprivation). In addition to the finding that residential instability may independently influence ischemic heart disease risk, our study expands on previous knowledge in showing that patterns of contextual effects may be different for incidence and survival after MI.
The effect of neighborhood deprivation on ischemic heart disease incidence may be partly mediated by the higher prevalence of health-damaging behavior and reduced use of healthcare services in disadvantaged neighborhoods.4 Such behavior may be conditioned by the availability of infrastructures1 and the values and opinions regarding health and health care prevailing in the neighborhood.46 Environmental hazards such as air pollution and noise may also play a role in the neighborhood socioeconomic effect.10,11,47
Regarding the residential stability effect, we verified that the associations were not confounded by population density or distance to the hospital. Because our models were adjusted for neighborhood socioeconomic deprivation, the effect of residential instability may not stem from neighborhood unpleasant conditions of living as causes of residential mobility. Rather, its effect may result from the consequences of the population turnover, which fragments the network of social connections between neighbors17–21 and thereby limits the support individuals may receive from their community.
The presence or absence of a neighborhood-based network of relationships may affect ischemic heart disease risk in different ways. First, immediate neighbors may provide instrumental support, ie, help to perform daily tasks or fulfill basic needs (emergency assistance, transportation to the doctor, pick up of medications at the pharmacy, and so on).48 Second, involvement in a network of neighbors is a source of informational support,49 ie, knowledge on cardiovascular diseases and risk factors, advice on what to do in case of acute problems, and information on healthcare resources available in the vicinity. Third, individuals may receive emotional support from their neighborhood community. By enhancing psychologic well-being, such support may help individuals in the post-MI acute period focus on health preservation and comply with medical recommendations.50 Finally, insertion in a neighborhood network may result in the social control of neighbors encouraging one to avoid health-damaging behavior and adhere to prescribed treatments.22 This interpretation that focuses on neighborhood-based social support does not exclude the possibility that other sources of support such as household members, other relatives, work colleagues, and aspatial networks of friends have an independent effect on ischemic heart disease risk. It is even possible that those nonneighborhood-based sources of social support contribute to between-neighborhood variations in ischemic heart disease risk.
Our interpretation that residential instability effects on survival are mediated by individual availability of social support is coherent with the literature showing that social support may have a stronger role in enhancing survival after ischemic heart disease than on incidence in itself.23 However, no conclusion can be made before this hypothesis of a mediating effect is tested empirically.
As an implication for future etiologic research, social support resources constitute another class of mediating factors to investigate to understand the black box of mechanisms between residential context and ischemic heart disease beyond those related to the availability of infrastructures and services1 and to the health knowledge and perceptions of individuals.46 Another consequence for contextual epidemiology of coronary disease is our demonstration that it is relevant to compare neighborhood effects at the different stages of the disease process (ie, on incidence and short- and long-term survival after MI).
1. Diez Roux AV. Residential environments and cardiovascular risk. J Urban Health
2. Lawlor DA, Davey Smith G, Patel R, et al. Life-course socioeconomic position, area deprivation, and coronary heart disease: findings from the British Women's Heart and Health Study. Am J Public Health
3. Sundquist K, Malmstrom M, Johansson SE. Neighbourhood deprivation and incidence of coronary heart disease: a multilevel study of 2.6 million women and men in Sweden. J Epidemiol Community Health
4. Diez Roux AV, Merkin SS, Arnett D, et al. Neighborhood of residence and incidence of coronary heart disease. N Engl J Med
5. Diez-Roux AV, Nieto FJ, Muntaner C, et al. Neighborhood environments and coronary heart disease: a multilevel analysis. Am J Epidemiol
6. Nordstrom CK, Diez Roux AV, Jackson SA, et al. The association of personal and neighborhood socioeconomic indicators with subclinical cardiovascular disease in an elderly cohort. The Cardiovascular Health Study. Soc Sci Med
7. Davey Smith G, Hart C, Watt G, et al. Individual social class, area-based deprivation, cardiovascular disease risk factors, and mortality: the Renfrew and Paisley Study. J Epidemiol Community Health
8. Sundquist K, Winkleby M, Ahlen H, et al. Neighborhood socioeconomic environment and incidence of coronary heart disease: a follow-up study of 25,319 women and men in Sweden. Am J Epidemiol
9. Chaix B, Rosvall M, Merlo J. Increase of neighborhood socioeconomic effects on ischemic heart disease mortality over the recent period: a multilevel survival analysis of two large Swedish cohorts. Am J Epidemiol.
10. Finkelstein MM, Jerrett M, Sears MR. Environmental inequality and circulatory disease mortality gradients. J Epidemiol Community Health
11. van Kempen EE, Kruize H, Boshuizen HC, et al. The association between noise exposure and blood pressure and ischemic heart disease: a meta-analysis. Environ Health Perspect
12. Rozanski A, Blumenthal JA, Kaplan J. Impact of psychological factors on the pathogenesis of cardiovascular disease and implications for therapy. Circulation
13. Chaix B, Isacsson SO, Rastam L, et al. Income change at retirement, neighbourhood-based social support and ischaemic heart disease: results from the prospective cohort study “men born in 1914.” Soc Sci Med.
2006. In press.
14. Perkins DD, Taylor RB. Ecological assessments of community disorder: their relationship to fear of crime and theoretical implications. Am J Community Psychol
15. Sampson RJ. What ‘community’ supplies. In: Ferguson RF, Dickens WT, eds. Urban Problems and Community Development
. Washington, DC: Brookings Institution Press; 1999:241–292.
16. Chaix B, Merlo J, Subramanian SV, et al. Comparison of a spatial perspective with the multilevel analytic approach in neighborhood studies: the case of mental and behavioral disorders due to psychoactive substance use in Malmö, Sweden, 2001. Am J Epidemiol
17. Shaw CR, McKay HD. Juvenile Delinquency in Urban Areas.
Chicago: University of Chicago Press; 1942.
18. Kasarda JD, Janowitz M. Community attachment in mass society. Am Sociol Rev
19. Sampson RJ, Byron Groves W. Community structure and crime: testing social-disorganization theory. Am J Sociol
20. Sampson RJ, Raudenbush SW, Earls F. Neighborhoods and violent crime: a multilevel study of collective efficacy. Science
21. Ross CE, Reynolds JR, Geis KJ. The contingent meaning of neighborhood stability for residents’ psychological well-being. Am Sociol Rev
22. Johnell K, Rastam L, Lithman T, et al. Low adherence with antihypertensives in actual practice: the association with social participation—a multilevel analysis. BMC Public Health
23. Vogt TM, Mullooly JP, Ernst D, et al. Social networks as predictors of ischemic heart disease, cancer, stroke and hypertension: incidence, survival and mortality. J Clin Epidemiol
24. Sundquist K, Theobald H, Yang M, et al. Neighborhood violent crime and unemployment increase the risk of coronary heart disease: a multilevel study in an urban setting. Soc Sci Med.
25. Franzini L, Spears W. Contributions of social context to inequalities in years of life lost to heart disease in Texas, USA. Soc Sci Med
26. Martikainen P, Kauppinen TM, Valkonen T. Effects of the characteristics of neighbourhoods and the characteristics of people on cause specific mortality: a register based follow-up study of 252,000 men. J Epidemiol Community Health
27. Borrell LN, Diez Roux AV, Rose K, et al. Neighbourhood characteristics and mortality in the Atherosclerosis Risk in Communities Study. Int J Epidemiol
28. LeClere FB, Rogers RG, Peters K. Neighborhood social context and racial differences in women's heart disease mortality. J Health Soc Behav
29. Diez Roux AV, Borrell LN, Haan M, et al. Neighbourhood environments and mortality in an elderly cohort: results from the cardiovascular health study. J Epidemiol Community Health
30. Chaix B, Rosvall M, Lynch J, et al. Disentangling contextual effects on cause-specific mortality in a longitudinal 23-year follow-up study: impact of population density or socioeconomic environment? Int J Epidemiol
31. Larsen K, Merlo J. Appropriate assessment of neighborhood effects on individual health: integrating random and fixed effects in multilevel logistic regression. Am J Epidemiol
32. Merlo J. Multilevel analytical approaches in social epidemiology: measures of health variation compared with traditional measures of association [Editorial]. J Epidemiol Community Health
33. Merlo J, Chaix B, Yang M, et al. A brief conceptual tutorial of multilevel analysis in social epidemiology—linking the statistical concept of clustering to the idea of contextual phenomenon. J Epidemiol Community Health
34. Merlo J, Chaix B, Ohlsson H, et al. A brief conceptual tutorial of multilevel analysis in social epidemiology—using measures of clustering in multilevel logistic regression to investigate contextual phenomena. J Epidemiol Community Health
35. Smith AFM, Roberts GO. Bayesian computation via the Gibbs sampler and related Markov chain Monte Carlo methods. J R Stat Soc Ser B Stat Methodol
36. Bingenheimer JB, Raudenbush SW. Statistical and substantive inferences in public health: issues in the application of multilevel models. Annu Rev Public Health
37. Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, et al. Myocardial infarction and coronary deaths in the World Health Organization MONICA Project. Registration procedures, event rates, and case-fatality rates in 38 populations from 21 countries in four continents. Circulation
38. Merlo J, Lindblad U, Pessah-Rasmussen H, et al. Comparison of different procedures to identify probable cases of myocardial infarction and stroke in two Swedish prospective cohort studies using local and national routine registers. Eur J Epidemiol
39. Validation of the Acute Myocardial Infarction Diagnosis Quality in the Hospital Patient Register 1987–1995
[Värdering av diagnoskvaliteten för akut hjärtinfarkt i patientregistret 1987 och 1995]. Stockholm, Sweden: Epidemiologiskt Centrum, Socialstyrelsen; 2000.
40. Danesh J, Collins R, Appleby P, et al. Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA
41. Schulz AJ, Kannan S, Dvonch JT, et al. Social and physical environments and disparities in risk for cardiovascular disease: the healthy environments partnership conceptual model. Environ Health Perspect
42. Loucks EB, Sullivan LM, Hayes LJ, et al. Association of educational level with inflammatory markers in the Framingham Offspring Study. Am J Epidemiol
43. Wilson TW, Kaplan GA, Kauhanen J, et al. Association between plasma fibrinogen concentration and five socioeconomic indices in the Kuopio Ischemic Heart Disease Risk Factor Study. Am J Epidemiol
44. Sundquist J, Johansson SE, Yang M, et al. Low linking social capital as a predictor of coronary heart disease in Sweden: a cohort study of 2.8 million people. Soc Sci Med
45. Stjarne MK, Ponce de Leon A, Hallqvist J. Contextual effects of social fragmentation and material deprivation on risk of myocardial infarction—results from the Stockholm Heart Epidemiology Program (SHEEP). Int J Epidemiol
46. Cubbin C, Winkleby MA. Protective and harmful effects of neighborhood-level deprivation on individual-level health knowledge, behavior changes, and risk of coronary heart disease. Am J Epidemiol
47. Chaix B, Gustafsson S, Jerrett M, et al. Children's exposure to nitrogen dioxide in Sweden: investigating environmental injustice in an egalitarian country. J Epidemiol Community Health
48. Rosengren A, Wilhelmsen L, Orth-Gomer K. Coronary disease in relation to social support and social class in Swedish men. A 15 year follow-up in the study of men born in 1933. Eur Heart J
49. Stafford M, Bartley M, Wilkinson R, et al. Measuring the social environment: social cohesion and material deprivation in English and Scottish neighbourhoods. Environ Planning A
50. Krumholz HM, Butler J, Miller J, et al. Prognostic importance of emotional support for elderly patients hospitalized with heart failure. Circulation
© 2007 Lippincott Williams & Wilkins, Inc.
This article has been cited