Citation

Hanf M, Nacher M, Guihenneuc C, Tubert-Bitter P, Chavance M. BMJ 2013; 347: f6427.

Affiliation

Biostatistics, CESP Centre for research in Epidemiology and Population Health, U1018, Inserm; Villejuif, France.

Copyright

(Copyright © 2013, BMJ Publishing Group)

DOI

unavailable

PMID

24212105

Abstract

OBJECTIVE: To assess at country level the association of mortality in under 5s with a large set of determinants. DESIGN: Longitudinal study. SETTING: 193 United Nations member countries, 2000-09. METHODS: Yearly data between 2000 and 2009 based on 12 world development indicators were used in a multivariable general additive mixed model allowing for non-linear relations and lag effects. MAIN OUTCOME MEASURE: National rate of deaths in under 5s per 1000 live births RESULTS: The model retained the variables: gross domestic product per capita; percentage of the population having access to improved water sources, having access to improved sanitation facilities, and living in urban areas; adolescent fertility rate; public health expenditure per capita; prevalence of HIV; perceived level of corruption and of violence; and mean number of years in school for women of reproductive age. Most of these variables exhibited non-linear behaviours and lag effects. CONCLUSIONS: By providing a unified framework for mortality in under 5s, encompassing both high and low income countries this study showed non-linear behaviours and lag effects of known or suspected determinants of mortality in this age group. Although some of the determinants presented a linear action on log mortality indicating that whatever the context, acting on them would be a pertinent strategy to effectively reduce mortality, others had a threshold based relation potentially mediated by lag effects. These findings could help designing efficient strategies to achieve maximum progress towards millennium development goal 4, which aims to reduce mortality in under 5s by two thirds between 1990 and 2015.

Language: en