An intergenerational and lifecourse study of health and mortality risk in parents of the 1958 birth cohort: (II) mortality rates and study representativeness

Abstract

We recently traced and flagged parents for the participants in the National Child Development Survey (NCDS, 1958 cohort). This paper evaluates the representativeness of the study population and assesses our success in identifying the biological parents. Intergenerational cohort study. Parents for participants in the NCDS (born 1 week in March 1958) were traced and flagged for mortality follow-up (14 334 fathers, 15 076 mothers). Standardized mortality rates (SMRs) were calculated using data from England, Scotland and Wales during corresponding time periods. By 31 December 2003, 6808 fathers and 4148 mothers (born 1890-1943) had died (569 918 and 645 354 years of follow-up, respectively). The overall mortality rate in this parent population was lower compared with the age-, gender-, period- and area-standardized reference rates (SMRs of 83 for fathers and 86 for mothers). Mortality rates for biological parents were higher if cohort members had had non-biological parent figures during the childhood surveys (SMRs of 135 for fathers and 374 for mothers). Parental smoking (in 1974) was strongly associated with lung cancer mortality among biological parents [HR 6.1, 95% confidence intervals (CI) 4.6-8.1 for fathers; HR 15.0, 95% CI 9.7-23 for mothers) but not among non-biological parents (HR 2.0, 95% CI 0.8-5.5; HR 1.8, 95% CI 0.4-7.9, respectively) which demonstrates that the tracing of the biological parents had been successful. Mortality is markedly reduced in a cohort of parents compared with the general population. The validity of identification of biological parents is demonstrated by the strong association between smoking and lung cancer.