The DHS Program

Accumulating Contraceptive Calendars across Surveys (English) 

Document Type

Working Papers

Publication Topic(s)

Family Planning

Country(s)

Senegal, Egypt, Bangladesh, Peru

Language

English

Recommended Citation

Elkasabi, Mahmoud, and Shireen Assaf. 2021. Accumulating Contraceptive Calendars across Surveys. DHS Working Papers No. 181. Rockville, Maryland, USA: ICF.

Download Citation

RIS format / Text format / Endnote format

Publication Date

September 2021

Publication ID

WP181

Abstract:

This paper extends the Demographic and Health Survey (DHS) Working Paper No. 177 that presented a framework for accumulating retrospective data. In that paper, the accumulation technique, which is widely known as data pooling, was applied to birth history data in DHS surveys. Childhood mortality rates were calculated based on cumulated birth histories, which are also known as accumulated rates. The estimated confidence intervals of the accumulated rates were more narrow than their counterparts from the separate survey datasets. The accumulated birth histories reduced fluctuations in the time series for national and subnational mortality rates, and established time trends that were more stable and reliable than those based on single-survey data. In this paper, we examine another application of the accumulation technique on the contraceptive calendars of DHS surveys from Bangladesh, Egypt, Peru, and Senegal. We examined the consistency of data collected in the contraceptive calendar by comparing monthly contraceptive use across different surveys in overlapping calendars. The results showed some differences in the contraceptive prevalence rate based on the overlapping calendars from two consecutive surveys, particularly in Bangladesh and Egypt. The Peru and Senegal DHS showed more consistent contraceptive prevalence rate estimates between the two sources of data. Some differences were more prevalent for specific contraceptive methods. We accumulated calendar data across surveys of Peru and Senegal, where consistent contraceptive prevalence rate estimates were observed. The accumulated data had more narrow confidence intervals when compared to the trends in the separate datasets. However, the application on calendar data did not show promising potential similar to the application on birth histories as found in the DHS Working Paper No. 177.