The Patient-Reported Outcomes Measurement Information System (PROMIS) developed by the National Institutes of Health can estimate patient reported outcomes (PROs) with accuracy similar to that of the lengthier Routine Assessment of Patient Index Data 3 (RAPID3) instrument, according to study results published in Arthritis Research and Care.
The study included participants from the ArthritisPower registry (n=6154). Participants underwent 4 PROMIS computer adaptive testing (CAT) instruments (Pain Interference, Physical Function, Sleep Disturbance, and Fatigue), as well as RAPID3 and pain intensity and patient global assessment score. The researchers used different combinations of these variables to predict RAPID3 and create a new score called CAT-PROMIS RAPID3. They used kappa statistics and Bland-Altman 95% limits of agreements to determine the agreement between the observed and predicted RAPID3.
The researchers collected 11,275 observations from the 6154 study participants. Median assessment times ranged from 29 seconds for the PROMIS sleep disturbance to 116 seconds for RAPID3.
In single pairwise comparisons, the researchers found that the PROMIS CATS modestly correlated to one another and to RAPID3 (r approximately 0.4-0.7).
The PROMIS instruments, along with pain intensity and patient global scores, accounted for a high fraction of total variance of the RAPID3 score (R2=0.97).
The results indicated that Bland-Altman 95% limits of agreement showed minimal residual differences and no systematic biases between the observed RAPID3 and the CAT-PROMIS RAPID3.
“This [PROMIS] approach was able to be effectively substituted for the fixed-length, 10-question [Multidimensional Health Assessment questionnaire] MDHAQ and save time, reducing the burden of data capture to patients,” the researchers wrote.
Yun H, Nowell B, Curtis D, et al. Assessing RA disease activity with PROMIS measures using digital technology [published March 30, 2019]. Arthritis Care Res. doi:10.1002/acr.23888