Diagnosing Preferences to Facilitate Shared Decision-Making: Development of a Novel Preference-Based Decision Support Tool for Adult Patients with Eosinophilic Esophagitis

OBJECTIVES: Eosinophilic esophagitis (EoE) is a chronic, allergic inflammatory condition that can result in esophageal fibrosis. Medical treatment options vary in treatment modality, efficacy, side effects, and regulatory approval for EoE specifically. Patients must select from several, non-dominant choices and desirability of a treatment depends on preferences. Management of EoE could therefore benefit from tools that enhance patient understanding of options, elicit preferences, and facilitate treatment conversations with clinicians. Using results from a previously administered discrete-choice experiment (DCE), we aimed to develop a novel prototype decision support tool for adult patients with EoE to “diagnose” a patient’s preference phenotype and facilitate clinical shared decision-making.

METHODS: Treatment preferences for treatment efficacy, regulatory approval, and treatment type were elicited using a DCE with 212 patients. Latent-class analysis identified 3 distinct classes of patients with similar preferences (i.e., patient-preference phenotypes). Using class-specific preference weights, we employed a Bayesian classification algorithm to construct a set of 3 choice questions that discriminate between these phenotypes to "diagnose" patient's membership in a known preference class with a high level of accuracy (90%-99%). Using responses to these preference-diagnostic questions and patient treatment history, we developed individualized patient reports that can be shared with clinicians.

RESULTS: The final prototype includes an education component presenting treatment options, questions capturing treatment history, questions diagnosing patient-preference phenotypes, and individualized reports summarizing patient's treatment experience and preferences that can be shared with their clinician. In pilot interviews, 4 EoE patients and 3 gastroenterologists have corroborated the accuracy of the preference diagnosis and usefulness of the tool for shared decision-making.

CONCLUSIONS: Early testing of the tool shows that the prototype can offer meaningful information to both patients with EoE and clinicians. We will further test preference-diagnostic accuracy, usability, and acceptability with patients and gastroenterologists to determine the best use case for the tool.