SOME RESULTS AND PROBLEMS ON THE ESTIMATION OF VITAL RATES IN A RURAL AFRICAN SETTING VIA MULTIPLE METHODS

This paper is one of a series concerning vital events registration and baseline demographic surveys conducted for the Danfa Rural Health and Family Planning Project in Ghana. Specifically, it presents the results of a recent experiment on the collection and analysis of data on births and deaths using the well-known Chandrasekhar-Deming multi-approach method. The site of the study was located l0-50 miles from Accra, an area with approximately 50,000 rural inhabitants. First, baseline data was gathered on age, sex, family relationship, education, religion, ethnic group, and pregnancy status. Follow-up surveys were carried out at 6-month intervals in order to increase the matching and accuracy of data on births, deaths, and migration. In designing their experiment, project personnel attempted to overcome the major difficulty of accounting for three broad classes of events inherent in any registration or survey. These classes of events include: (l) those that occur in the sample area to usual residents; (2) those that occur to residents when they are away from the area; and (3) those that occur to visitors to the sample area. A number of variables were selected for the matching of vital events: the house number, the name of the child/deceased, the sex of the child/deceased, date of birth/death, place of birth/death (and residential status), name of the mother (for births), and age at death for the deceased. Problems arose due to differences in receiving and recording information on names, and because of the large number of items to be compared and verified. In addition, problems involving memory lapse, mobility of individuals and households, and other factors suggest that some uniform method should be developed so that matching can be done with greater accuracy. In addition, the manual matching operations used in this experiment were time-consuming and delayed publication of results. Data from the registration and surveys are presented in tabular form and are discussed in the text.