Patterns in Attribute Selection and Development Reporting in Patient Preference Studies Between 2007-2024: A Systematic Literature Review

Background Discrete choice experiments (DCE) are commonly used for understanding patient preferences. However, their validity relies upon appropriate attribute selection and development. The present review aimed to identify reporting patterns and gaps to inform future reporting of patient preference studies. Methods Ovid (MEDLINE and EMBASE) was used to search terms for “patients,” “discrete choice experiments,” and “attribute selection.” Two independent reviewers screened studies against PICOS eligibility criteria, focusing on formative attribute development methods and reporting quality in patient preference DCEs up to 2024. Only full-text journal publications detailing attribute selection and development in patient preference DCEs were included. A coding form was developed and used to capture reporting on formative research used for attribute selection or development. Data synthesis employed a narrative approach following PRISMA guidelines. Results Six categories of formative methods were identified across 28 studies: patient qualitative concept elicitation (n = 27; 96%), literature reviews (n = 21; 75%), expert consultation (n = 21; 75%), quantitative prioritization (n = 17, 61%), quantitative pilot DCE surveys (n = 13; 46%), and qualitative cognitive debriefing ( n =7; 25%). Most studies stated objectives (>85%), but methodological transparency decreased substantially for other elements, including rationale for choice of method, sample characteristics, sampling methods, data collection procedures data analysis, and results for formative research methods involving primary data collection. Most studies reported attribute lists following formative research (>75% across methods), but across formative methods fewer than half overall reported how formative findings informed attribute selection, level selection, or wording decisions. While 96% of studies included patients as research participants, only 7% reported engaging patients as research partners. Conclusions The findings revealed inconsistent reporting of formative methods for attribute selection and development. Reporting of formative method details and results was mixed, with particularly low levels of reporting for how formative research results informed attribute selection, level selection, and wording. Furthermore, few studies engaged patients as research partners, suggesting another key area for development in the field of patient preference research. Improved reporting standards are needed to support methodological clarity and strengthen the validity of patient preference research. Patient engagement in the development of studies may further strengthen the patient relevance of patient preference studies.