Bridging the gap between evidence and what people value from osteoarthritis care in New Zealand using multi-criteria decision analysis (MCDA)

International clinical practice guidelines (CPG) for osteoarthritis (OA) consistently recommend core management strategies of exercise, weight-loss and self-management education. However, these interventions are not routinely delivered or taken up by people experiencing OA, resulting in a tendency to deliver low-value care. Tailoring the delivery of recommended OA care to the preferences of stakeholders in a health system may help support greater implementation of high-value OA care. However, little is known about the preferences of stakeholders for evidence and recommendations for OA care in a CPG. The aim of this research was to establish a framework to prioritise knee interventions for managing OA and evaluate the relative importance of these interventions across the healthcare sector in New Zealand (NZ) using multi-criteria decision analysis (MCDA). This research used a mixed-methods approach to develop the MCDA framework. The first stage involved focus group discussions to generate a thematic framework of what OA consumers, health care providers, policy-makers, Māori representatives and OA experts consider in their choice of knee OA interventions. A Delphi survey was used to verify the thematic analysis and rank the most important criteria concerning the characteristics of interventions; these findings informed the criteria selected in the second stage. In the second stage, interventions were indirectly prioritised by systematically combining preference data with intervention performance data on the criteria. First, a survey involving trade-offs was used to measure stakeholders’ preferences, represented as ‘weights’, within and between the criteria. Evidence for the performance of 15 recommended knee OA interventions were then extracted from a CPG for hip and knee OA, and rated on the criteria according to their level of achievement (e.g. high, medium, low). To prioritise the interventions, a total score for each intervention was calculated by summing the weights associated with the intervention ratings, after which they were ranked by importance. Associations between the weights and stakeholder groups were explored using regression analysis. Thematic analysis of data from six focus groups produced a framework comprising three overarching categories, consisting of characteristics of the: (i) intervention, (ii) consumer and (iii) health system. Participants identified and ranked nine characteristics of interventions; the most important eight were included in the MCDA framework. The choice-based survey revealed that stakeholders valued the intervention characteristics, in decreasing order of importance (weight): Recommendation (19.0%), Quality of evidence (17.7%), Effectiveness (15.0%), Duration of effect (13.2%), Risk of serious harm (12.8%), Risk of mild side-effects (9.4%), Cost (6.6%) and Accessibility (6.3%). Total scores for the 15 guideline-recommended interventions revealed that for first-, second- and third-line OA care respectively, all land-based exercise (total score= 71.7%), NSAIDs (topical) (74.2%) and total joint replacement (74.3%) were ranked first. For first-, second- and third-line OA care, the recommended core interventions of weight management and self-management education ranked between 11th and 15th (48.0% to 56.0%). Regression analysis identified only small differences in weights (≤5.7%; p