How Are Age, Gender, and Country Differences Associated With PROMIS Physical Function, Upper Extremity, and Pain Interference Scores?
Constantin Yves Plessen,Gregor Liegl,Claudia Hartmann,Marilyn Heng,Alexander Joeris,Aaron J Kaat,Benjamin D Schalet,Felix Fischer,Matthias Rose,AOBERT Consortium
DOI: https://doi.org/10.1097/CORR.0000000000002798
2024-02-01
Abstract:Background: The interpretation of patient-reported outcomes requires appropriate comparison data. Currently, no patient-specific reference data exist for the Patient-Reported Outcome Measurement Information System (PROMIS) Physical Function (PF), Upper Extremity (UE), and Pain Interference (PI) scales for individuals 50 years and older. Questions/purposes: (1) Can all PROMIS PF, UE, and PI items be used for valid cross-country comparisons in these domains among the United States, the United Kingdom, and Germany? (2) How are age, gender, and country related to PROMIS PF, PROMIS UE, and PROMIS PI scores? (3) What is the relationship of age, gender, and country across individuals with PROMIS PF, PROMIS UE, and PROMIS PI scores ranging from very low to very high? Methods: We conducted telephone interviews to collect custom PROMIS PF (22 items), UE (eight items), and PI (eight items) short forms, as well as sociodemographic data (age, gender, work status, and education level), with participants randomly selected from the general population older than 50 years in the United States (n = 900), United Kingdom (n = 905), and Germany (n = 921). We focused on these individuals because of their higher prevalence of surgeries and lower physical functioning. Although response rates varied across countries (14% for the United Kingdom, 22% for Germany, and 12% for the United States), we used existing normative data to ensure demographic alignment with the overall populations of these countries. This helped mitigate potential nonresponder bias and enhance the representativeness and validity of our findings. We investigated differential item functioning to determine whether all items can be used for valid crosscultural comparisons. To answer our second research question, we compared age groups, gender, and countries using median regressions. Using imputation of plausible values and quantile regression, we modeled age-, gender-, and country-specific distributions of PROMIS scores to obtain patient-specific reference values and answer our third research question. Results: All items from the PROMIS PF, UE, and PI measures were valid for across-country comparisons. We found clinically meaningful associations of age, gender, and country with PROMIS PF, UE, and PI scores. With age, PROMIS PF scores decreased (age ß Median = -0.35 [95% CI -0.40 to -0.31]), and PROMIS UE scores followed a similar trend (age ß Median = -0.38 [95% CI -0.45 to -0.32]). This means that a 10-year increase in age corresponded to a decline in approximately 3.5 points for the PROMIS PF score-a value that is approximately the minimum clinically important difference (MCID). Concurrently, we observed a modest increase in PROMIS PI scores with age, reaching half the MCID after 20 years. Women in all countries scored higher than men on the PROMIS PI and 1 MCID lower on the PROMIS PF and UE. Additionally, there were higher T-scores for the United States than for the United Kingdom across all domains. The difference in scores ranged from 1.21 points for the PROMIS PF to a more pronounced 3.83 points for the PROMIS UE. Participants from the United States exhibited up to half an MCID lower T-scores than their German counterparts for the PROMIS PF and PROMIS PI. In individuals with high levels of physical function, with each 10-year increase in age, there could be a decrease of up to 4 points in PROMIS PF scores. Across all levels of upper extremity function, women reported lower PROMIS UE scores than men by an average of 5 points. Conclusion: Our study provides age-, gender-, and country-specific reference values for PROMIS PF, UE, and PI scores, which can be used by clinicians, researchers, and healthcare policymakers to better interpret patient-reported outcomes and provide more personalized care. These findings are particularly relevant for those collecting patient-reported outcomes in their clinical routine and researchers conducting multinational studies. We provide an internet application ( www.common-metrics.org/PROMIS_PF_and_PI_Reference_scores.php ) for user-friendly accessibility in order to perform age, gender, and country conversions of PROMIS scores. Population reference values can also serve as comparators to data collected with other PROMIS short forms or computerized adaptive tests. Level of evidence: Level II, diagnostic study.