Developing a discrete choice experiment in Malawi: eliciting preferences for breast cancer early detection services

Abstract

Racquel E Kohler,1 Clara N Lee,2 Satish Gopal,3 Bryce B Reeve,1 Bryan J Weiner,1 Stephanie B Wheeler11Department of Health Policy and Management, Gillings School of Global Public Health, 2Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA; 3UNC Project-Malawi, Tidziwe Center, Lilongwe, MalawiBackground: In Malawi, routine breast cancer screening is not available and little is known about women’s preferences regarding early detection services. Discrete choice experiments are increasingly used to reveal preferences about new health services; however, selecting appropriate attributes that describe a new health service is imperative to ensure validity of the choice experiment.Objective: To identify important factors that are relevant to Malawian women’s preferences for breast cancer detection services and to select attributes and levels for a discrete choice experiment in a setting where both breast cancer early detection and choice experiments are rare.Methods: We reviewed the literature to establish an initial list of potential attributes and levels for a discrete choice experiment and conducted qualitative interviews with health workers and community women to explore relevant local factors affecting decisions to use cancer detection services. We tested the design through cognitive interviews and refined the levels, descriptions, and designs.Results: Themes that emerged from interviews provided critical information about breast cancer detection services, specifically, that breast cancer interventions should be integrated into other health services because asymptomatic screening may not be practical as an individual service. Based on participants’ responses, the final attributes of the choice experiment included travel time, health encounter, health worker type and sex, and breast cancer early detection strategy. Cognitive testing confirmed the acceptability of the final attributes, comprehension of choice tasks, and women’s abilities to make trade-offs.Conclusion: Applying a discrete choice experiment for breast cancer early detection was feasible with appropriate tailoring for a low-income, low-literacy African setting.Keywords: breast cancer, early detection, patient preferences, discrete choice experiment, Malawi, qualitative interviews