How to embed a choice experiment in an online decision aid or tool: a scoping review (Preprint)

Abstract

BACKGROUND Decision aids empower patients to understand how treatment options match their preferences. Choice experiments, a valuable method to clarify values used within decision aids, present patients with hypothetical scenarios to reveal their preferences for treatment characteristics. Given the rise in research embedding choice experiments in decision tools and the emergence of novel developments in embedding methodology, a scoping review is warranted. OBJECTIVE This scoping review examines how choice experiments are embedded into decision tools and how these tools are evaluated, to identify best practices. METHODS This scoping review was conducted following best practices in line with the PRISMA extension for scoping reviews. The searchers were conducted on MEDLINE, PsycInfo, and Web of Science databases using key search terms. Data were extracted using data charting tables created in Excel. A narrative synthesis was used to summarize the data and illustrations were used to visualise the results using tables and figures. RESULTS Overall, 22 tools were included in the scoping review. The methodology, development and evaluation details of tools were extracted from 33 papers. These tools were developed for a variety of health conditions including musculoskeletal conditions, oncological conditions, and chronic conditions. Most tools (78%) originated in the USA. The primary purpose (91%) of these tools was to assist patients in comparing or choosing treatments. The most commonly included attributes in the choice tasks were efficacy and side effects. Adaptive conjoint analysis was the most frequent (10 tools) design approach. Conjoint analysis designs used a higher number of tasks (16 -20) while DCEs and adaptive conjoint analysis designs used low (6) to moderate (12) number of tasks. Sawtooth software was commonly used to embed choice tasks in the tools. After completing the choice tasks patients received tailored information in the form of attribute importance scores, highlighting which treatment characteristics mattered most to the patient based on their choices (16 tools), and/or a "best match" treatment recommendation aligned with the patient's preferences (5 tools). A high degree of heterogeneity was observed in the evaluation methodologies and outcome measures used to assess the decision tools. The decisional conflict scale emerged as the most frequently employed outcome measure. CONCLUSIONS This study highlights several methodological challenges that require further investigation. Future research should focus on determining the most effective methods for embedding choice tasks in decision tools, presenting balanced information, and selecting suitable outcome measures to evaluate these tools.