PSS8 CLINICIAN INPUT FOR THE DEVELOPMENT OF PATIENT-REPORTED OUTCOMES: HOW MANY INTERVIEWS ARE ENOUGH?

Abstract

Development of patient-reported outcome (PRO) questionnaires relies on qualitative research to demonstrate content validity, which can be supported by clinician input. However, limited information is available to help guide sample size decisions for clinician interviews to support PRO questionnaire development. This research evaluates how many clinician interviews are needed to elicit key disease-related signs/symptoms documenting concept saturation (whether new, relevant data would emerge in additional interviews), a technique used to evaluate adequacy of patient sample sizes. Nine qualitative studies conducted between 2013-2018 with clinician interviews across nine dermatological conditions were retrospectively analyzed. Saturation analysis was conducted for each study by comparing new concepts reported by clinicians in the final interview set (quartile or tertile depending on sample size) to concepts reported in earlier interviews. The emergence of fewer concepts in the final interview set suggests that little to no new relevant data may be elicited from further interviews. Forty-five clinicians (mean=5/study [SD=2.1, range=3–10]) working in private/academic clinical research and practice settings with 21.9 mean years of experience were interviewed. Clinicians reported seeing a median of 25 patients/week and reported 14 mean signs/symptoms per condition. Most studies (n=5, 55.6%) included five clinicians, three studies (33.3%) included 3-4 clinicians, and one study (11.1%) included 10 clinicians. In 77.8% of studies (n=7) including 3-10 clinicians, 100.0% of concepts were elicited before the final set of interviews. In the remaining two studies (both n=5 clinicians), 81.8% and 92.3% of concepts were elicited before the final interviews (1-2 new concepts elicited/study). Results indicate that a sample of approximately five clinicians may be sufficient to elicit relevant disease-related signs/symptoms to inform PRO questionnaire development in dermatological conditions. Clinicians can report on the experiences of numerous individuals they see; therefore, fewer clinicians may be needed compared to patient interview studies.