Convergent validity of the patient reported outcomes measurement information system’s (PROMIS) physical function computerized adaptive test (PF-CAT) for the knee and shoulder injury sports medicine patient population

Abstract

Objectives: The Patient Reported Outcomes Measurement Instrument System (PROMIS) Physical Function Computer Adaptive Test (PF-CAT) has been developed by the National Institutes of Health as a means of providing validated outcomes scores. PF-CAT adapts to patient response by applying different validated questions based on prior responses to generate scores using between 4 to 12 questions, and has been validated in a variety of patient populations to assess physical function. The purpose of our study is to compare patient reported outcomes scores in a sports medicine (shoulder and knee injury) patient population by analyzing convergence validity of PF-CAT to IKDC, SST, ASES, and SANE scores. We also wanted to assess and compare time burden placed on patients in completing these outcomes scores. Methods: All patient visits from April through September 2014 presenting with either knee or shoulder complaints were included from a university-based sports medicine clinic, during which both PF-CAT as well as SANE, SST, and ASES legacy outcomes scores for shoulder injury patients, and SANE and IKDC outcomes scores for knee injury patients were obtained via tablet during the initial part of each visit. A total of 415 shoulder and 450 knee clinical evaluations qualified for inclusion in the study. Spearman correlation was used to evaluate pair-wise agreement among outcomes scores. The total scores from each outcome test and the T-scores from the PF-CAT were used in this correlation analysis. For interpretation, results greater than 0.5-0.69 were considered to have fair correlation, while values greater than 0.7 were considered as having good correlation. Significance was assessed at the 0.05 level. Median and interquartile ranges were used to summarize scores. The minimum score (floor effects) and maximum score (ceiling effects) occurrences were tallied and reported as percentages. A Fishers exact test was used to test for a difference in the number of times floor and ceiling values occurred. The Wilcoxon signed rank test was used to test for differences in completion times (measured in seconds). Results: Good correlation was found between IKDC and PF-CAT scores (r=0.7511; p<0.001), and fair correlation was found between SST (r=0.6777; p<0.001) and ASES (r=0.6262; p<0.001). Percent of patients’ scores that hit floor and ceiling values was 15% on SST (p<0.001), 2.5% on ASES (ceiling only, p=0.001), and 0% for PF-CAT shoulder patients. In knee injury patients, 0% of IKDC and 0.44% of PF-CAT scores hit floor or ceiling values (p<0.514). Completion times for PF-CAT averaged 81 seconds versus 359 seconds for IKDC which was significant (p<0.001). Average completions times for shoulder patients completing PF-CAT, SST, and ASES scores was 94, 192, and 157 (349 combined) seconds respectively (p <0.001). Conclusion: PF-CAT holds strong convergence correlation with IKDC outcome scores in a sports medicine knee injury population, and only fair convergence correlation with SST and ASES scores in a sports medicine shoulder injury population. In addition, PF-CAT avoided floor and ceiling effects equal to IKDC outcomes and better than SST and ASES outcomes. Finally, PF-CAT was completed in a shorter time period than IKDC scores as well as compared to SST and ASES times both separately and combined. The PROMIS PF-CAT is an acceptable outcomes tool to evaluate sports medicine knee injuries that places less time burden on patients to complete outcomes scores.