Early life conditions and mortality later in life in Scania, Sweden 1814-1894: new ways of measuring conditions during early life

Kent Johansson, Lund University
Jan Beise, Max Planck Institute for Demographic Research

Health in later life is often claimed to be partly determined already during intra-uterine life and infancy (Barker (1998, 2001)). To confirm this ‘Barker’ hypothesis empirically, lengthy population records are needed, which makes it hard to test this hypothesis using modern data compilations, but several studies based on historical population reconstitutions have confirmed such links (Bengtsson (1997:15-19), Bengtsson & Lindström (2000), Alter (2000), Bengtsson, Broström & Lindström (2002), Johansson (2004)). However, the actual early life nutrition conditions and/or disease load conditions are unknown and must be approximated, so the ability to approximate these conditions accurately is crucial for the quality of such an investigation. Also, distinguishing between early-life-effects from nutrition and from disease was recognized as highly important at an early-life-session at PAA2006 (Session 135 ‘Early Life Experiences and Mortality’ at PAA 2006). This paper suggests improved and also new measures for distinguishing between effects from disease and from nutrition during early life, and between conditions in utero and in infancy. These measures are then used in Cox survival regression models with shared frailty for adult mortality in Scania, Sweden 1814-1894.