Urban-rural variations in health in The Netherlands: does selective migration play a part?

Abstract

STUDY OBJECTIVE: Urban-rural health differences are observed in many countries, even when socioeconomic and demographic characteristics are controlled for. People living in urban areas are often found to be less healthy. One of the possible causes for these differences is selective migration with respect to health or health risk factors. This hypothesis is hardly ever empirically tested. This paper tries to assess the existence of selective urban-rural migration. DESIGN: Health indicators and health risk factors were measured in a 1991 population sample. Moves were registered between 1991 and 1995. Using logistic regression analyses, comparisons were made between, firstly, urban to rural movers and rural to urban movers and secondly, between movers and stayers. SETTING: Region surrounding the city of Eindhoven in south eastern part of the Netherlands. SUBJECTS: Data were used of 15,895 respondents aged 20-74 in 1991. By 1995 613 subjects had moved from urban to rural and 191 subjects from rural to urban. MAIN RESULTS: Bivariate nor multivariate analyses show hardly and differences between movers into urban and movers into rural areas. Bivariate analyses on movers and stayers show that movers are healthier than stayers. However, when socioeconomic and demographic variables are controlled for, movers appear to be less healthy, with the exception of the younger age groups. CONCLUSIONS: Areas that attract many migrants from and lose few migrants to other degrees of urbanicity will in the long run obtain healthier populations, because of demographic and socioeconomic characteristics. However, if these characteristics are accounted for, the opposite is true, with the exception of younger age groups. In extreme cases this may cause spurious findings in cross sectional research into the relation between urbanicity and health. Absolute numbers of migrants need to be very high, however, to make this noticeable at the aggregate level.