Longitudinal Study of Neighborhood Predictors of CVD

BACKGROUND:

There are few comprehensive databases that allow for a longitudinal examination of how neighborhood-level factors may interact with individual characteristics to predict incident cases of cardiovascular disease (CVD) and all-cause mortality. This cross-institutional project joins investigators from Karolinska Institutet in Sweden and Stanford University, known for their expertise in immigrant health and social inequalities in CVD.

DESIGN NARRATIVE:

The study examines how neighborhood social characteristics (e.g., neighborhood socioeconomic status (SES), social disintegration, socioeconomic and ethnic segregation, social capital), physical environments (e.g., geocoded assets including goods and services contributing to health such as educational and recreational resources; barriers to health including pollution, industries, waste dumps, criminal activity, alcohol and fast food outlets) and individual factors (e.g., SES, cardiovascular disease {CVD} risk factors, country of birth and social networks) may interrelate to predict CVD [morbidity and mortality] and all-cause mortality. Data will be used from a newly created, comprehensive set of Swedish databases, MigMed and MigSALLS. MigMed (1990-2002) includes data for the entire Swedish population of 6 million women and men aged 25 and older, of whom 600,000 are first generation immigrants. Their addresses have been geocoded, yielding 9,677 neighborhoods, which will be reduced to a smaller number of defined units by cluster analyses. MigMed includes an annual assessment of individual-level sociodemographic and health indicators. Data will also be analyzed from MigSALLS (1988-2002), which includes more in-depth data from face-to-face interviews with a representative sample of approximately 18,000 women and men aged 25-74, of whom 2,000 are first generation immigrants. MigSALLS contains similar individual- and neighborhood-level factors as MigMed, as well as extensive information on factors that may mediate relationships between neighborhoods and CVD outcomes (e.g., CVD risk factors such as smoking, weight, physical activity, and blood pressure). Information from these two datasets will be matched to hospital and death records (300,000 deaths and 140,000 incidence cases of CVD expected between 1988-2002), thus creating one of the largest databases in the world involving men and women from diverse SES levels and countries of origin.