METHODS FOR THE EVALUATION OF SOCIAL EFFECTS ON BIRTH-WEIGHT - EXPERIENCES WITH SWEDISH POPULATION REGISTRIES

All births in Sweden in 1986, registered in the Medical Birth Registry , were analyzed with linked data from the November 1985 census in orde r to identify variables which could be used as proxy variables for the socio-economic situation of the women, using birth weight as outcome variable. Two phenomena appear to be strong socio-economic indices of significance for birth weight: whether the woman co-habited or not at the census and whether she had a long education (> 14 years) and an '' academic'' work. The two variables interacted, however: when the woman had a long education, cohabitation status played no significant role. Maternal education as judged from occupation seemed to play a more im portant role than paternal education or job. To some extent, the effec ts of the socio-economic variables could be explained by different smo king habits but a residual effect existed which may have different exp lanations. Cohabitation status and education may be used as estimates of socio-economic level of significance for birth weight and perhaps o ther pregnancy outcome in the analysis of, for instance, the effect of various occupational exposures on pregnancy outcome.