PNS227 Response SETS within Patient Global Impression Measures: Challenges and Recommendations for Development

Abstract

Patient Global Impression (PGI) instruments are used to assess aspects of a patient’s experience with a given health condition or treatment. Respondents typically evaluate disease activity or severity, or change in health status, choosing among discrete response options on an ordinal scale with positive and negative extremes. Despite their simplicity, common issues in PGI translation and comprehension can interfere in the accurate assessment of a PGI’s target construct, as well as their ability to pool data collected across languages or countries. This research reviews these common problems in PGI translation and presents recommendations for resolving them. A convenience sample of 8 recent cognitive debriefing projects including PGIs was compiled and patient feedback on response options, suggestions for translation revisions, and terminology preferences was extracted for review. Overall, debriefing results from 85 language-country pairs were analyzed, each containing 5 patient interviews (N = 425). Our results indicate that 7-point PGI scales are problematic across many languages, and difficulty frequently arises in distinguishing between conceptually similar response options (e.g., “a little better” vs “moderately better”; “minimally” vs “mildly”). Regardless of response set length, common response options such as “moderate”, “worst imaginable”, and “quite a bit” were conceptually unclear across languages (particularly Indian and Nordic languages, as well as Russian variants), and more descriptive language was typically preferred by patients. Another frequent issue was misinterpretation of “global” in the instrument title, predominantly in variants of English and Romance languages. Based on these data, RWS Life Sciences recommends avoidance of large and conceptually similar response sets, along with standardization and simplification of the terminology used in PGIs. Unclear wording and overly dense response sets interfere in the collection of accurate data on patient experience, especially in the context of global studies on patients from diverse language and educational backgrounds.