Send to

Choose Destination
See comment in PubMed Commons below
Int J Epidemiol. 1993 Dec;22(6):965-75.

Mortality in a rural South African mission, 1837-1909: an historical cohort study using church records.

Author information

Centre for Epidemiological Research in Southern Africa, SA Medical Research Council, Tygerberg.


Three decadal birth cohorts (1837-1846), 1870-1879 and 1900-1909) each of approximately 500 individuals, were constructed retrospectively through the parish records of the Moravian Mission at Mamre in the Western Cape region of South Africa. Nominative data collection techniques were used to determine the infant mortality rates (IMR), quinquennial mortality rates (QMR) and life expectancies of the three cohorts. The quality of the data was investigated, specifically non-registration and attrition. Overall, male registration coverage was substantially better than that for females. Birth registration was best for the 1837-1846 cohort for males and females, with the 1900-1909 registration being next best. Infant death registration was most complete for males in these two cohorts, but was poor for females. Based on these data, the IMR for the cohorts born in 1837-1846, 1870-1879 and 1900-1909 (196, 182 and 128 per 1000 respectively for males and 160, 172 and 97 per 1000 respectively for females) appeared to be underestimates. There was some evidence of a downward trend for the IMR with time for males, but this was not statistically significant. For the three cohorts QMR did not differ statistically for males, but for females the third cohort was consistently lower than the other cohorts. The difference was small (well within the 95% confidence interval) but the third cohort ranked consistently below the other two. The life expectancies did not differ significantly between cohorts. The life expectancies at birth (range 34-40 years for males and 32-45 years for females) were probably overestimates due to biased IMR. The life expectancies at age 1 (range 41-44 for males and 37-49 for females) were considered to be more representative figures. Life expectancies at age 20 (37-45 years) were fairly stable over time except for females in the 1900-1909 cohort whose life expectancies were substantially higher than the figures for the earlier cohorts. All mortality indices investigated in this study consistently showed a lighter burden of mortality in historical Mamre compared to 'coloureds' in the Cape Colony at the turn of the century. This is probably the result of the better housing, environmental, social, economic and educational conditions on the missions relative to the rest of the Colony in the century after the emancipation of slaves.


The aim was 1) to examine the quality of birth and death registration of 3 Mamre, South African cohorts born between 1837-1846, 1870-1879, and 1900-1909; 2) determine the infant mortality rates (IMR), quinquennial mortality rates (QMR), and life expectancies (LE) of the cohorts; and 3) to determine the trends and compare these to other population estimates. Data was obtained from parish records of the Moravian Church at Mamre between 1830 and the present. The records included Casualis books, Baptismal registers, Notebooks, Burial orders, official birth and death registration lists, and a household survey of Mamre in 1986, which traced live members of the 1900-1909 cohort. Nominative methods were used to extract the data on individuals and families. The findings showed that the most complete source of birth and death information was the Baptismal Register; other supportive documents were available or the 1st and 3rd cohorts. The crude birth rate was assumed to be 47/1000 and population size was 1000, 1250, and 1100 for the 1st, 2nd, and 3rd cohorts. The ratio of observed to expected births was respectively .89, .86, and .91. There was preferential registration of boys in all cohorts. The median age of baptism increased over time; baptism rarely occurred after 2 months in the 1st cohort. There was a declining percentage of births born outside Mamre over time, and follow-up until death decreased over time. The ratio of male to female infant deaths was 1.62, 1.34, and 1.55 for the 1st, 2nd, and 3rd cohorts. Only in the 2nd were unbaptized births recorded (14.3%). There was a decline in infant deaths within 10 days following baptism from 48.5% (33), to 25.6% (20), to 29.4% (15). There were 498 deaths in the 1st cohort, 593 in the 2nd, and 501 in the 3rd for which birth and death entries were available. In this population, infant mortality for males declined, but not in a statistically significant amount. There was no female IMR pattern. The QMR curves by age showed high IMR with a decline between 1-5 years, and a gradual increase after 10 years. Females followed the same pattern except for those aged 25-29 years, which were high, and those aged 30-34 years, who had low rates, particularly in the 1900-1909 cohort. There were no secular trends in LE.

[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Loading ...
    Support Center