Measuring Service Quality and Assessing Its Relationship to Contraceptive Discontinuation: A Prospective Cohort Study in Pakistan and Uganda

• Published in: Global health science and practice Vol. 8; no. 3; pp. 442 - 454
• Main Authors: Chang, Karen T., Chakraborty, Nirali M., Kalamar, Amanda M., Hameed, Waqas, Bellows, Ben, Grépin, Karen A, Gul, Agha Xaher, Bradley, Sarah E.K., Atuyambe, Lynn M., Montagu, Dominic
• Format: Journal Article
• Language: English
• Published: Global Health: Science and Practice 01.10.2020
• Subjects: Original
• ISSN: 2169-575X; 2169-575X
• DOI: 10.9745/GHSP-D-20-00105

BACKGROUNDThe quality of contraceptive counseling that women receive from their provider can influence their future contraceptive continuation. We examined (1) whether the quality of contraceptive service provision could be measured in a consistent way by using existing tools from 2 large-scale social franchises, and (2) whether facility quality measures based on these tools were consistently associated with contraceptive discontinuation. METHODSWe linked existing, routinely collected facility audit data from social franchise clinics in Pakistan and Uganda with client data. Clients were women aged 15-49 who initiated a modern, reversible contraceptive method from a sampled clinic. Consented participants completed an exit interview and were contacted 3, 6, and 12 months later. We collapsed indicators into quality domains using theory-based categorization, created summative quality domain scores, and used Cox proportional hazards models to estimate the relationship between these quality domains and discontinuation while in need of contraception. RESULTSThe 12-month all-modern method discontinuation rate was 12.5% among the 813 enrolled women in Pakistan and 5.1% among the 1,185 women in Uganda. We did not observe similar associations between facility-level quality measures and discontinuation across these 2 settings. In Pakistan, an increase in the structural privacy domain was associated with a 60% lower risk of discontinuation, adjusting for age and baseline method (P<.001). In Uganda, an increase in the management support domain was associated with a 33% reduction in discontinuation risk, controlling for age and baseline method (P=.005). CONCLUSIONSWe were not able to leverage existing, widely used quality measurement tools to create quality domains that were consistently associated with discontinuation in 2 study settings. Given the importance of contraceptive service quality and recent advances in indicator standardization in other areas, we recommend further effort to harmonize and simplify measurement tools to measure and improve contraceptive quality of care for all.