Optimising primary care management of knee osteoarthritis (the partner study): lessons from a non-randomised pilot study

Purpose: The PARTNER study is evaluating a new model of service delivery to improve the management of knee osteoarthritis (OA) in people presenting to primary care (ACTRN12617001595303). A feasibility pilot study was undertaken from 1 May to 31 December 2017 to test key components of the study protocol before commencing the RCT. We aimed to test processes and identify issues that may affect implementation. We did not randomise, include a control group or assess participant outcomes. Methods: The pilot was implemented at two general practices in Victoria, Australia. Both practices, and all participating general practitioners (GPs) and patients, received the intervention as described in the protocol. Briefly, the intervention provided training and support for GPs on current best-practice management for people with knee OA, using a mixture of tailored online training and desktop IT support. Their recruited patients (≥45 years with knee pain) had an initial GP visit to confirm a clinical diagnosis of knee OA and were referred to a centralised, multidisciplinary care support team (CST), trained in evidence-based OA management and behaviour change support. The CST’s role was to support effective self-management behaviours. Four main components were examined: i) processes (e.g. recruitment for GPs and patients, data collection); ii) time or resource issues affecting delivery (e.g. costs, staffing levels); iii) unforeseen human or data issues affecting implementation (e.g. software and databases, general practice staff); and iv) scientific quality considerations (e.g. acceptability of GP training, fidelity of CST intervention, acceptability of educational materials, adverse events and participant satisfaction). We sought feedback from patients, GPs, general practice staff, our CST, and other trial staff through structured questionnaires, semi-structured phone interviews, data collection monitoring and analysis of trial documents. Results: We recruited 8 GPs and 12 patients. We did not encounter problems with the recruitment of either GPs or patients. One GP withdrew citing time constraints, and one patient withdrew as their pain had reduced. There were no adverse events. Three GPs (43%) had undertaken prior training in OA management in the previous 2 years, and most GPs reported moderate to high confidence in their ability to treat OA (mean 7/10). No GPs reported an interest in health coaching, although several had interests in weight-loss (n=3) and pain management (n=3). Patients enrolled (mean age 69.7 years, 58% female) had a mean pain score of 5.8/10 on a numerical rating scale. Ten patients had a previous diagnosis of knee OA from their GP, but only one reported having an existing OA management plan. The main co-morbidities included back pain (50%), hypertension (33%), and high cholesterol (25%). Feedback from most patients about their involvement was positive (n=9), with all reporting their participation as “simple and straightforward”. They reported having a better understanding of their OA and non-surgical options for self-management following the intervention. They were satisfied with the support received from the CST. Completion of training by GPs was our most significant intervention protocol issue. Feedback from GPs suggested this was due to a combination of time constraints (too long), limitations with our online-delivery platform, and some redundancy in our content. Delivery of the GP training has been restructured for the main trial. Conclusions: The pilot was helpful in demonstrating the feasibility of our methods and delivery of the intervention protocol. We incorporated all changes identified into the main trail protocol before commencement in September 2018.