Collecting data on pregnancy loss: a review of evidence from the World Fertility Survey

Abstract

Estimates of levels and differentials of pregnancy loss are presented for 40 developing countries participating in the World Fertility Survey (WFS) program. Judged against agreed-upon levels of spontaneous loss in human populations, WFS surveys measured from 50 to 80 percent of recognizable losses. The coverage of induced abortions appears to be much worse. Consistent with data from other sources and settings, the probability of loss is strongly correlated with maternal demographic characteristics: age, pregnancy order, pregnancy spacing, and pregnancy loss history. Despite incomplete coverage, the WFS data on pregnancy loss provide considerable, and largely unexploited, insight on the dynamics of the reproductive career.

PIP: In spite of an assumption that information on pregnancy loss obtained retrospectively would be deficient, World Fertility Survey (WFS) studies carried out in developing countries have collected data on this variable, principally as a means of improving data on live births. Data on pregnancy loss were deemed of importance only in those few countries (Korea and Tunisia) where levels of abortion were high and legal sanctions mild enough that reasonable validity could be assumed. When judged against agreed-upon levels of spontaneous pregnancy loss in human populations (60%, but a maximum recognizable level of 25%), it appears that WFS studies in 40 developing countries measured 50-80% of recognizable such losses. The coverage was most complete for the 5-year period immediately preceding the survey. Overall, spontaneous pregnancy loss appears to be a commonly reported phenomenon,however. In only 8 of the 40 countries analyzed did fewer than 20% of the women report having experienced at least 1 spontaneous loss (Ghana, Lesotho, Mauritania, Nigeria, Nepal, Fiji, Indonesia, and Korea), while in 3 countries over 40% reported at least 1 such loss (Jordan, Dominican Republic, and Guyana). Further analysis of WFS data on spontaneous pregnancy loss suggest several demographic correlates, including maternal age (markedly higher rates of loss over age 30 years than at ages 20-29 years), parity (the probability of loss rises after the 6th pregnancy in all countries), and interbirth interval (with rates of loss most extreme when the interval from previous termination to next conception is less than 6 months). WFS coverage of induced abortions is far less accurate. WFS data show less than 2% of pregnancies terminated by an induced abortion in all 40 countries except Korea, Tunisia, Trinidad and Tobago, and Jordan. There is evidence that many induced abortions are reported as spontaneous pregnancy losses, confounding data analysis. Despite methodological problems, refined attention to the collection of data on pregnancy losses is advised to enhance understanding of variations in fertility over the reproductive cycle.