Misfortune before Birth: Intrauterine Mortality in China, 1955-1987

Yong Cai, University of Washington
Feng Wang, University of California, Irvine

Of all pregnancies, a large proportion, in the range of 15 to 20 percent, end in intrauterine mortality, either in the form of spontaneous abortion or stillbirth. Whereas stillbirths are shown to vary to some degree by the level of prenatal care, which in turn is affected by both the socioeconomic characteristics of the mother and availability of medical assistance at the societal level, the prevalence of spontaneous abortion is believed to be mostly spontaneous, or natural, and therefore largely immune of the impact of the social environment. Exceptions do exist, of course. Shocks external to the mother’s body, such as large-scale and severe famines and drastic social disruptions, may well affect not only the mortality of those who are born, but those before their births as well. One example is found in the recent history of the world’s largest national population, China. During two such drastic social and economic disruptions, once the famine following the socialist Great Leap Forward around 1960, and then coinciding with the peak of the Cultural Revolution in 1967, reported prevalence of spontaneous abortion based on large-scale retrospective survey of pregnancy histories shot up by 30 to 50 percent. Reported occurrences of stillbirth also rose during these time periods. Our main goal in this paper is to find out whether, for these deaths before birth, there is a pattern of social stratification similar to that found for mortality after birth. We intend to examine whether the reported occurrence of prenatal mortality vary by the mother’s and her husband’s social characteristics, and how the spikes in prenatal mortality associated with social disruptions and distresses distributed themselves among populations of different social segments. We are interested in finding out, in other words, the extent to which intrauterine mortality is also a social outcome. We are both blessed and challenged to find out answers to the questions raised above. On the one hand, we have data from China’s Two-Per-Thousand Fertility and Birth Control Survey, which include nearly 1.4 million pregnancy histories retrospectively reported by close to half million Chinese women aged 15 to 57 in 1988. In addition to pregnancy outcomes, about 45,000 (or about 3 percent) of which ended in spontaneous abortions, the dataset also contains detailed information on the mother’s and her husband’s biological and social characteristics (age at pregnancy, order of pregnancy, existing birth, education and residence, e.g.). On the other hand, such a dataset based on retrospective survey is also subject to several serious limitations that may well comprise our findings. Among them are both selection (of survivors) bias and reporting bias, both of which pose serious challenges. In our analyses we shall bear in mind these and other limitations of the data. Our main analytical strategy is to use multivariate regression (logistic regression), with external social and economic shocks built in as explicit variables. We shall examine, in a series of multivariate analytical models, the extent to which the difference in intrauterine mortality (for spontaneous abortion and stillbirth separately) is associated with the socioeconomic background and demographic characteristics of the mother and her husband. Through the use of interaction terms between the external shocks and the personal characteristics of the couple, we shall detect the presence or absence of the differential impact of social distresses on mortality before birth.