Implementing osteoarthritis guidelines in UK Primary Care: Mosaics cluster randomised controlled trial

Purpose: Recommendations for osteoarthritis (OA) are not currently implemented in UK primary care. The MOSAICS study determined the clinical and cost effectiveness of a model OA consultation as a method of implementing core guideline recommendations for OA. Methods: Health survey; two arm cluster randomised controlled trial; 15,083 (53%) eligible responders 45 years and over from eight general practices; 525 patients consulting for peripheral joint pain during the 6m recruitment period of the cluster trial (4 intervention, 4 control practices). Intervention practices received practice updates on core OA recommendations (diagnosis; written information [patient generated OA guidebook], exercise and physical activity, healthy eating, pain management); General practitioners (GPs) received four hours of in¬practice training and simulated consultations on a model OA consultation; Practice Nurses received four days of education, workshops and training with simulated consultations for a model OA consultation; Participants randomised in the intervention arm received a consultation with the GP for joint pain, and as appropriate an OA guidebook and consultations for OA with a practice nurse (up to 4 visits). Clinical effectiveness was measured by the SF12 physical component score (primary outcome) at 6 months, OMERACT/OARSI responder criteria, joint pain intensity, pain self¬efficacy; Uptake of self-management measured by self¬reported Quality Indicators of OA care and patient enablement. Cost-consequence and cost utility analyses were also undertaken. Results: Mean (SD) practice size and number of GPs for intervention and control practices were 10240 (9174.8) & 6.0 (6.1) and 6983 (2060.7) & 5.2 (2.9) respectively. Of eligible participants, 288 were recruited from intervention practices and 237 from control practices. The mean age (SD) was 67.3 (10.5) and 59.6%, were Female. At three months self-reported consultations with a practice nurse for joint problems were n=70 (29%) in the intervention group compared with n=26 (13.5%) in the control arm. At six months there were no statistically significant differences in SFPCS, EQ5D or in other health outcomes between intervention and control, except for a borderline significant finding in reduction of knee pain intensity in the intervention arm compared with control (p=0.55). Uptake of self-management was statistically significantly greater (Odds Ratios) in the intervention arm compared with control for all core OA recommendations: provision of written information (e.g. exercise) (4.36, p< 0.001). Enablement mean (SD) scores were greater in the intervention arm compared with the control arm at 6 months and this was also statistically significant (3.21 (3.42) vs 2.28 (2.96), p=0.03). Visits to the orthopaedic surgeon were lower in the intervention arm compared with the usual care arm (p=0.02). Time off work and associated productivity cost were lower in the intervention arm. Conclusions: A model OA consultation can increase the uptake of OA recommendations in primary care but does not increase health status or quality of life. An integrated model with an emphasis on support for self¬management and lifestyle changes had a borderline significant impact on pain reduction in the knee. The model OA consultation also reduced uptake of NSAIDs and visits to an orthopaedic surgeon. Defining an integrated model consultation supported by training workshops can be an effective way of implementing clinical guidelines.