New Client-Centered Discontinuation Measures Using the Demographic and Health Survey Calendar Data

The main source of method discontinuation data comes from the calendar data of the Demographic and Health Surveys (DHS). DHS considers each episode of use contributed by a woman in the calendar to calculate method discontinuation. This means that episodes from women who contribute only one episode of contraceptive use are combined with episodes from women who contribute multiple episodes of use. This paper explores the DHS calculation of episode-based discontinuation and proposes new indicators that focus specifically on the client and puts her as the unit of analysis. First, we attempt to replicate the DHS calculation by applying weighted averages to episodes from women who contributed only one episode and from women who contributed multiple episodes of use. We then calculate three client-centered discontinuation indicators, using the five-year calendar data of the DHS. The methodology is illustrated by using data collected in Bangladesh, Kenya, and Indonesia. Starting with the very first method used and reported in the calendar and following use over 12 months, we develop a client-centered discontinuation indicator. We build on the calculation of the client-centered discontinuation indicator by: (1) excluding switching to result in all contraceptive discontinuation (complete discontinuation); and (2) limiting discontinuation to women still in need of contraception applying the DHS definitions of in need.