Citation

Maxwell L, Brahmbhatt H, Ndyanabo A, Wagman JA, Nakigozi G, Kaufman JS, Nalugoda F, Serwadda D, Nandi A. Soc. Sci. Med. 2018; 209: 25-32.

Affiliation

Department of Epidemiology Biostatistics & Occupational Health, McGill University, 1020 Pine Avenue West, Montréal, QC H3A 1A2, Canada; Institute for Health and Social Policy, McGill University, 1130 Pine Ave West, Montréal, QC H3A 1A3 Canada. Electronic address: arijit.nandi@mcgill.ca.

Copyright

(Copyright © 2018, Elsevier Publishing)

DOI

10.1016/j.socscimed.2018.04.050

PMID

29783092

Abstract

A systematic review of longitudinal studies suggests that intimate partner violence (IPV) is associated with reduced contraceptive use, but most included studies were limited to two time points. We used seven waves of data from the Rakai Community Cohort Study in Rakai, Uganda to estimate the effect of prior year IPV at one visit on women's current contraceptive use at the following visit. We used inverse probability of treatment-weighted marginal structural models (MSMs) to estimate the relative risk of current contraceptive use comparing women who were exposed to emotional, physical, and/or sexual IPV during the year prior to interview to those who were not. We accounted for time-fixed and time-varying confounders and prior IPV and adjusted standard errors for repeated measures within individuals. The analysis included 7923 women interviewed between 2001 and 2013. In the weighted MSMs, women who experienced any form of prior year IPV were 20% less likely to use condoms at last sex than women who had not (95% CI: 0.12, 0.26). We did not find evidence that IPV affects current use of modern contraception (RR: 0.99; 95% CI: 0.95, 1.03); however, current use of a partner-dependent method was 27% lower among women who reported any form of prior-year IPV compared to women who had not (95% CI: 0.20, 0.33). Women who experienced prior-year IPV were less likely to use condoms and other forms of contraception that required negotiation with their male partners and more likely to use contraception that they could hide from their male partners. Longitudinal studies in Rakai and elsewhere have found that women who experience IPV have a higher rate of HIV than women who do not. Our finding that women who experience IPV are less likely to use condoms may help explain the relation between IPV and HIV.

Copyright © 2018 Elsevier Ltd. All rights reserved.

Language: en

Keywords

Causal inference; Cohort; Condoms; Intimate partner violence; Long-acting and permanent methods; Marginal structural models; Reproductive coercion