Evaluation of Test-Retest Reliability for the Patient-Reported Outcomes Measurement Information System Upper
Hernan Roca1, Gretchen Maughan-Egbert1, Brittany N Garcia1
1Department of Orthopaedics, University of Utah, Salt Lake City, UT.
Purpose:
Utilization of the Patient-Reported Outcomes Measurement Information System (PROMIS) upper extremity (UE) and pain interference (PI) computer adaptive tests (CAT) has grown in the hand surgery literature. However, these instruments were derived using data from a general population rather than cohorts with UE pathology. As such, our primary purpose was to evaluate the test-retest reliability of the UE and PI CATs in a sample of hand surgery patients.
Methods:
At a tertiary academic center, we recruited adult (≥18 years) orthopedic hand surgery patients who completed the PROMIS UE and PI CAT 1-7 days prior to their office appointment, plus a subsequent response to both instruments in clinic. Patients were included only if their upper extremity symptoms were stable between questionnaires, as assessed by an anchor question (eg, no change in scores is expected). Intraclass correlation coefficients (ICCs) were used to assess test-retest reliability. Bland-Altman Plots were created to visualize the proportion of patients whose scores differed by a value greater than the minimally clinically important difference for both instruments (4.1 for UE CAT, 4.3 for PI CAT).
Results:
The 88 included patients had a mean age of 54 ± 18 years, and 56% (56/88) were women. There was a mean of 4 ± 2 days between surveys. The UE CAT demonstrated an ICC of 0.87 (95% confidence intervals [CI], 0.80-0.91), consistent with "good" to "excellent" reliability. The PI CAT ICC was 0.84 (95% CI, 0.76-0.89), consistent with "good" reliability. However, 34% (30/88) and 23% (20/88) of patients had UE CAT and PI CAT score differences that exceeded minimally clinically important difference thresholds.
Conclusions:
For a hand surgery cohort (population level), the UE CAT demonstrated "good" to "excellent" test-retest reliability, and the PI CAT demonstrated "good" reliability. However at the patient level, score differences were clinically relevant in approximately one of three patients for UE CAT and one of four patients for the PI CAT. We conclude that these instruments are reliable at the population level, but should not be used at the level of individual hand surgery patients.
Type Of Study/Level Of Evidence:
Therapeutic II.
