Development of a discrete choice experiment to understand patient preferences for diabetes and hypertension management in rural Uganda

• Main Authors: Sarah E Moor, MSc, Andrew K Tusubira, MPH, Ann R Akiteng, MPH, Evelyn Hsieh, MD, Christine Ngaruiya, MD, Tracy L Rabin, MD, Nicola L Hawley, PhD, Kasia J Lipska, MD, Mari Armstrong-Hough, PhD, Christine K Nalwadda, PhD, Rachel Nugent, PhD, Robert Kalyesubula, MMed, Isaac Ssinabulya, MMed, Jeremy I Schwartz, MD
• Format: Article
• Language: English
• Published: Elsevier 2020-04-01
• Series: The Lancet Global Health
• Online Access: http://www.sciencedirect.com/science/article/pii/S2214109X20301637

Background: In 2020, non-communicable diseases (NCDs) currently account for 40% of deaths in east Africa and are expected to surpass communicable diseases as the leading causes of death in sub-Saharan Africa by 2030. However, there remain many gaps in the provision of high quality, patient-centred NCD service delivery in low-income countries such as Uganda, especially in rural settings. We developed a discrete choice experiment (DCE) to understand how patients with common NCDs, such as hypertension and diabetes, make health-care-seeking decisions. A DCE is a quantitative tool that measures the weight of different factors that affect a decision. Participants are presented with two hypothetical scenarios to choose between. Here we report on the formative, qualitative phase of DCE development. Methods: For this DCE, choice sets consist of two health facilities described in terms of various attributes (factors) that affect decision-making. To develop a locally relevant DCE, we conducted formative qualitative research consisting of 18 in-depth interviews with patients seeking care for hypertension or diabetes, or both, at three health facilities in rural Nakaseke District, Uganda. We purposively selected participants so that each disease and facility were represented. Interviews explored how participants choose which facility to visit and challenges that she or he faces in accessing care or medicines, or both. A team of three researchers coded the interviews using a directed approach to reveal 11 potential attributes on which patients make decisions. After review of the data and analysis with local and international content experts, we narrowed this list to six attributes to be included in the DCE. Findings: The six selected attributes included: provision of education; availability of medicines; transport to the facility; costs associated with treatment; interactions with health-care providers; and presence of peer support groups. Attribute levels encompassed the range of experiences for each attribute as emerged from qualitative analysis. We then developed a full profile, fractional factorial DCE with three surveys. Interpretation: The process of developing this NCD-focused DCE relied on reference to the literature, primary qualitative data collection, and expert consensus to create a tool that would yield actionable data to improve NCD health service delivery in rural Uganda. Little is known about how rural patients in low-income and middle-income countries navigate the health-care system, nor is there much description of how to develop DCEs to investigate patient preferences in these countries. This work will guide public health officials in developing NCD service delivery options and inform investigators on DCE development. Funding: Yale Institute of Global Health Hecht Global Health Faculty Network Award Downs International Health Student Travel Fellowship