Sven G Bremberg1. 1. Department of Public Health, Karolinska Institute, Stockholm, Sweden and Public Health Agency of Sweden, Stockholm, Sweden sven.bremberg@mac.com.
Abstract
BACKGROUND: Studies of country-level determinants of health have produced conflicting results even when the analyses have been restricted to high-income counties. Yet, most of these studies have not taken historical, country-specific developments into account. Thus, it is appropriate to separate the influence of current exposures from historical aspects. METHODS: Determinants of the infant mortality rate (IMR) were studied in 28 OECD countries over the period 1990-2012. Twelve determinants were selected. They refer to the level of general resources, resources that specifically address child health and characteristics that affect knowledge dissemination, including level of trust, and a health related behaviour: the rate of female smoking. RESULTS: Bivariate analyses with the IMR in year 2000 as outcome and the 12 determinants produced six statistically significant models. In multivariate analyses, the rate of decrease in the IMR was investigated as outcome and a history variable (IMR in 1990) was included in the models. The history variable alone explained 95% of the variation. None of the multivariate models, with the 12 determinants included, explained significantly more variation. CONCLUSION: Taking into account the historical development of the IMR will critically affect correlations between country-level determinants and the IMR.
BACKGROUND: Studies of country-level determinants of health have produced conflicting results even when the analyses have been restricted to high-income counties. Yet, most of these studies have not taken historical, country-specific developments into account. Thus, it is appropriate to separate the influence of current exposures from historical aspects. METHODS: Determinants of the infant mortality rate (IMR) were studied in 28 OECD countries over the period 1990-2012. Twelve determinants were selected. They refer to the level of general resources, resources that specifically address child health and characteristics that affect knowledge dissemination, including level of trust, and a health related behaviour: the rate of female smoking. RESULTS: Bivariate analyses with the IMR in year 2000 as outcome and the 12 determinants produced six statistically significant models. In multivariate analyses, the rate of decrease in the IMR was investigated as outcome and a history variable (IMR in 1990) was included in the models. The history variable alone explained 95% of the variation. None of the multivariate models, with the 12 determinants included, explained significantly more variation. CONCLUSION: Taking into account the historical development of the IMR will critically affect correlations between country-level determinants and the IMR.
Authors: Nhung T H Trinh; Sophie de Visme; Jérémie F Cohen; Tim Bruckner; Nathalie Lelong; Pauline Adnot; Jean-Christophe Rozé; Béatrice Blondel; François Goffinet; Grégoire Rey; Pierre-Yves Ancel; Jennifer Zeitlin; Martin Chalumeau Journal: Lancet Reg Health Eur Date: 2022-03-01