Abstract
Purpose: People living with osteoarthritis (OA) often do not receive best evidence care. Coordinated OA management programs (OAMPs) have been implemented to address this global evidence-practice gap. An OAMP is defined as a package of care with the following: i) a personalized management plan; ii) with reassessment and progression; iii) using a minimum of 2 core treatments (education, exercise, weight control), and; iv) optional adjunctive therapies. Existing OAMP models differ in treatment mode, intensity, duration, the health professionals delivering care, and the healthcare systems and settings they operate within. Randomized trials (RCTs) and cohort studies assess the outcomes of different OAMPs, however, these models are unlikely to ever be compared in RCTs due to the huge expense and complicated logistics required. Prognosis research provides another method of comparing outcomes of different OAMP models. This study aimed to estimate the pain and self-reported function outcomes (at 12-, 26- and 52-weeks) of people with hip and/or knee OA who participated in international OAMPs. It also aimed to describe the characteristics of OAMP participants.
Methods: This study was undertaken by members of the OARSI Joint Effort Initiative (JEI), in collaboration with the OA Trial Bank (Erasmus MC, Netherlands). RCTs and clinical cohorts assessing OAMPs were identified through the JEI membership and literature searches. Eligible studies included data from an ongoing OAMP, in any real-world setting, with participants who were diagnosed with hip or knee OA, and longitudinal measures of patient-reported pain and function. The investigators of eligible studies were invited to complete data delivery agreements with the OA Trial Bank, share individual participant data (IPD), contribute to study design and authorship. Investigators ensured they had local ethics review board approval to contribute IPD to the OA Trial bank. Each dataset was converted to a common format to enable merging into one dataset. The IPD were evaluated to convert pain and function variables to standardized scales as appropriate. Pain scores were converted to a 0-100 point scale (100 worst). Function scores were converted to a 0-100 point scale (100 best). A generalized estimating equations (GEE) model analysis was performed to assess the change in pain and function from baseline across weeks 12, 26, and 52. The model specification was based on an unstructured correlation structure and robust standard errors. Pain and function estimates were adjusted by age, sex and body mass index (BMI). Data analyses were carried out using Stata 15 (StataCorp 2015) and SPSS 17.
Results: The investigators of 13 international OAMPs were invited to take part. IPD from 9 OAMPs were delivered: the OA Chronic Care Program, Ramsay Health OA Management Program, Joint Health Program, University of Wisconsin Health Knee and Hip Comprehensive Non-Surgical OA Management Clinic, Improved Management of Patients With Hip and Knee OA in Primary Health Care, Joint Academy, Amsterdam OA cohort, Management of OA In Consultations, and Collaborative model of care between Orthopaedics and allied healthcare professionals in knee OA. The characteristics of the OAMPs are summarised in table 1. The OAMPs were conducted in-person except for the Joint Academy that was implemented as an online OAMP. Individual participant data from 9819 participants were analyzed. The cohort studies were missing large amounts of data, as expected in clinical practice. The characteristics of OAMP participants are summarised in Table 2. The majority of OAMP participants reported the knee as their index joint, their mean age ranged between 62- 67 years, 58-74% were female, 25-48% were working and mean BMI indicated they were overweight at baseline. Pain was most commonly assessed using a Numeric Rating Scale or validated questionnaires e.g. the Knee Injury and OA Outcome Scale (KOOS). Function was mostly assessed using validated questionnaires such as the KOOS. The pain and fuction measured in the original datasets are reported in Table 1. The changes in pain and function of the OAMP participants from baseline across weeks 12, 26, and 52 are summarised in Table 3. There were reductions in pain scores and improvements in function scores seen across all programs at the majority of timepoints.
Conclusions: We established the first data bank of IPD from different international OAMPs. Analysis of the IPD demonstrated modest improvements in pain and function across the programs at all timepoints. The most rapid improvements were made by week-12, however, these gains were maintained at week-52. In future work this project will use IPD meta-analysis to identify prognostic factors of people with OA who participate in OAMPs.