PATIENT PREFERENCES IN TYPE 2 DIABETES - DEVELOPING A STATED-PREFERENCE INSTRUMENT USING A COMMUNITY ENGAGEMENT APPROACH

Purpose: While stated-preference methods play an important role in patient-centered outcomes research (PCOR), there is a paucity of literature on the development of stated-preferences instruments, particularly with regards to appropriately engaging patients and community stakeholders. As part of a PCORI methods award, we sought to develop and implement a community-centered approach to develop a stated-preference instrument focused on type 2 diabetes treatments. 

Method: Using Pronovost’s four E’s (engage, education, execute, evaluate) we worked with patients and community stakeholders in Baltimore, Maryland to develop two stated-preferences instruments: a Best-Worst Scaling (BWS) and a Discrete-Choice Experiment (DCE). The Diabetes Action Board (DAB), an advisory board including patients, local community representatives, clinicians, methods experts and other stakeholders, guided our study. The local representatives facilitated the recruitment of people with type 2 diabetes for the qualitative pretesting of the instrument using cognitive interviewing. A DAB meeting was held to present results of the pretests and to evaluate success of the engagement process. 

Result: The DAB helped to shape the study, and with the selection of six attributes (each with three levels). Qualitative pretest (n=25) were then conducted in Baltimore among people with diabetes aged between 31-89 years and with varying times since diagnosis (<1 to 37 years). Community engagement was successful; 40 percent of participants were recruited directly through a family member or friend. Three instrument iterations were tested and interviews lasted between 20-60 minutes. Pretest participants thought attributes were relevant but stressed the use of non-scientific language. Only five participants (20 percent) identified omitted attributes. Sixteen percent of participants preferred the BWS because the simplicity of information. Twenty-eight percent of participants preferred the DCE because the intuitive choice between two items. Fifty-six percent did not have or did not mention a preference. During the DAB meeting, committee members expressed the wish to be more involved in shaping the different iterations of the instrument.  

Conclusion: This study incorporated patient and stakeholder engagement in the development of stated-preference instruments. This type of engagement is time and resource intensive; engagement and education efforts started several months before survey development began. With the rise of PCOR, engagement strategies need to be explored so engagement can be meaningful and the engagement process can be made more efficient.