Electronically self-assessed functional capacity and exercise testing: A comparison of the Duke Activity Status Index and Patient-Reported Outcomes Measurement Information System tools

Abstract

Electronic screening tools, such as Patient-Reported Outcomes Measurement Information System (PROMIS) Physical Function Short-Form 12a (PF-SF12a), may aid in the assessment of functional capacity. However, PROMIS PF-SF12a has not been validated against exercise capacity, or compared with established questionnaires, including the Duke Activity Status Index (DASI). We compared the DASI and PROMIS PF-SF12a to the maximum metabolic equivalents (METs) achieved during exercise stress testing. DASI and PROMIS PF-SF12a were electronically administered to 100 adult patients (median age 56years, 61% male) immediately before exercise stress testing. DASI-predicted METs and PROMIS T score were calculated. Correlations with exercise METs with and without age adjustment were examined. Linear regression lines were derived and adjusted r2 statistic was calculated. We compared models with the Davidson-Mackinnon J test. The median (interquartile range) DASI-predicted METs, PROMIS Tscore, and exercise METs were 8.97 (7.61-9.89), 47.90 (43.33-52.40), and 10.10 (10.10-12.80), respectively. In unadjusted correlation analyses, PROMIS accounted for 26% of the variance in exercise METs compared with 38% with DASI. With age adjustment, the r2values increased to 0.36 (PROMIS) and 0.46 (DASI). In both unadjusted and age-adjusted analyses, inclusion of DASI improved prediction of exercise METs beyond PROMIS T score (P<.0001). In contrast, PROMIS T score did not improve exercise MET prediction compared with DASI alone (P>.10). Among patients undergoing clinically indicated exercise stress testing, DASI outperformed PROMIS PF-SF12a as a predictor of exercise METs.