Use of patient-reported outcome measures (PROs) in rheumatology research is widespread, but use of PRO data to evaluate the quality of rheumatologic care delivered is less well established. This article reviews the use of PROs in assessing health care quality, and highlights challenges and opportunities specific to their use in rheumatology quality measurement. It first explores other countries' experiences collecting and evaluating national PRO data to assess quality of care. It describes the current use of PROs as quality measures in rheumatology, and frames an agenda for future work supporting development of meaningful quality measures based on PROs.

Objective

We assessed implementation of the 10-item Patient-Reported Outcomes Measurement Information System (PROMIS) physical function form (PF-10a) in routine practice in a racially and ethnically diverse population with rheumatoid arthritis (RA). Objectives were to determine feasibility of implementing PF-10a in the electronic health record (EHR) and PF-10a validity and longitudinal responsiveness.

Methods

Clinical and demographic data were abstracted from EHRs for all RA patients seen at a university-based rheumatology clinic between February 2013 and February 2015. We evaluated floor and ceiling (edge) effects and construct validity of PF-10a in a subgroup of patients with Health Assessment Questionnaire (HAQ) scores (n = 189). We used linear mixed-effects models to assess responsiveness of PF-10a to longitudinal changes in the Clinical Disease Activity Index (CDAI) for patients in the entire clinical cohort, with both scores recorded on at least 2 encounters (n = 326).

Results

Half of the patients were nonwhite, and 15% were non-English speakers. Over a 2-year period, PF10a was successfully implemented; 97% of patients and 89% of encounters had at least 1 measurement performed. PF-10a had fewer ceiling (defined as best) effects than the HAQ (8% versus 22%), and convergent validity was high (r = -0.85). PF-10a was sensitive to expected differences (older versus younger patients, more versus less active disease). Longitudinal changes in PF-10a were highly associated with changes in the CDAI score (P < 0.0001).

Conclusion

PF-10a was feasible to implement in a diverse RA population. It strongly correlates with the HAQ but has fewer ceiling effects and is responsive to changes in RA disease activity, suggesting its validity for use in routine clinical practice.

Objective

Patient-reported outcomes (PROs) are an integral part of treat-to-target approaches in managing rheumatoid arthritis (RA). In clinical practice, however, routine collection, documentation, and discussion of PROs with patients are highly variable. The RISE LC (Rheumatology Informatics System for Effectiveness Learning Collaborative) was established to develop and share best practices in PRO collection and use across adult rheumatology practices in the United States METHODS: The goals of the RISE LC were developed through site surveys and in-person meetings. Participants completed a baseline survey on PRO collection and use in their practices. RISE LC learning sessions focused on improving communication around PROs with patients and enhancing shared decision-making in treatment plans. During the coronavirus disease 2019 (COVID-19) pandemic, the RISE LC pivoted to adapt PRO tools for telehealth.

Results

At baseline, all responding sites (n = 15) had established workflows for collecting PROs. Most sites used paper forms alone. PRO documentation in electronic health records was variable, with only half of the sites using structured data fields. To standardize and improve the use of PROs, participants iteratively developed a Clinical Disease Activity Index-based RA Disease Activity Communication Tool to solicit treatment goals and improve shared decision-making across sites. The COVID-19 pandemic necessitated developing a tool to gauge PROs via telehealth.

Conclusion

The RISE LC is a continuous, structured method for implementing strategies to improve PRO collection and use in rheumatological care, initially adapting from the Learning Collaborative model and extending to include features of a learning network. Future directions include measuring the impact of standardized PRO collection and discussion on shared decision-making and RA outcomes.