Date of Award

January 2022

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Jeremy I. Schwartz


Non-communicable diseases account for 33% of deaths in Uganda and are expected to surpass communicable diseases as the leading causes of death in sub-Saharan Africa by 2030. However, chronic care service delivery in low-income countries, especially in rural settings, remains sub-optimal. We developed a discrete choice experiment (DCE) to explore how respondents with common chronic conditions - hypertension (HTN) and diabetes mellitus (DM) - make healthcare-seeking decisions related to these conditions. DCEs are tools to quantify preferences for goods and services. Participants are asked to choose between sets of two hypothetical scenarios that differ in terms of particular characteristics. Their selections reveal the relative importance of each “attribute”, or characteristic, and the extent to which people will consider trade-offs between characteristics. Specific efforts must be made in the development process to ensure the local relevance of a DCE to the population to whom it is administered. However, there is a lack of detail in the literature on the step-by-step process and decisions that one must consider. We present the six steps we followed to generate a DCE: 1) formative work; 2) attribute selection; 3) attribute level selection; 4) DCE design selection; 5) determination of attribute level combinations; and 6) assessment and enhancement of tool comprehensibility. This process created a DCE with six attributes: getting to the facility, interactions with healthcare providers, availability of medicines for condition, costs of treatment, patient peer support groups; and education at the facility. We administered theDCE to 496 adults with HTN and/or DM sampled from three health facilities within Nakaseke District, Uganda and mixed logit modeling was performed to determine relative utility estimates. Respondents preferred traveling 5 or 10 km over traveling 2 km (ß 0.65; P<0.001; ß 0.52; P<0.001, respectively). Preference for facilities with low staff turnover and healthcare providers with whom they have a friendly relationship was also demonstrated (ß 1.05; P <0.001). Respondents preferred free monthly treatment (ß 1.14; P<0.001) compared to paying 20,000 Ugandan Shillings (UGX) (approximately US$5.60) monthly, but preferred paying 5,000 UGX or 10,000 UGX over free treatment (ß 1.14; P<0.001, ß 0.31; P=0.003, respectively). Results demonstrated a preference for receiving one month’s worth or more than one month’s worth of medicines compared to none (ß 1.31; P<0.001; ß 1.15; P<0.001, respectively), but did not demonstrate any preference between receiving no or some medicine from the facility (ß -0.11; P=0.3710). Lastly, respondents preferred health facilities with peer support groups compared to facilities without them (ß 0.27; P = 0.007) and preferred receiving no facility-based health education compared to some (ß -0.34; P<0.001). This study revealed significant preferences for health facilities based on the availability of medicines, costs of treatment, and interactions with healthcare providers. Peer support groups were shown to be preferred by patients with HTN only, while respondents with DM only or HTN and DM demonstrated a willingness to travel to farther healthcare facilities. Future work can further elucidate the thought processes underlying these less intuitive preferences. Understanding these patient preferences can inform intervention design to optimize healthcare service delivery for patients with chronic conditions among rural populations in Uganda and other low-resource settings.


This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.