An evaluation of survey data on birthweight and prematurity status.

The 1986 experimental Demographic and Health Survey in the Dominican Republic was one of the first retrospective, population-based surveys of a developing country to collect information on prematurity status as well as birthweight. We show that the relationships among birthweight, prematurity status, and infant mortality are consistent with corresponding patterns observed in other Latin American and Caribbean countries and in developed nations. Two notable irregularities in the birthweight data are the atypically high proportion of infants weighing 4,000 grams or more, and the high mortality rates among infants with missing birthweight. Problems with the prematurity status data include an unexpectedly low proportion of premature births, an unusually small proportion of low birthweight explained by premature births, and an extremely high neonatal death rate among premature compared to term infants. Inclusion of an explicit definition of "premature" may help avert similar problems in future health surveys.