Optimising the current model of care for knee osteoarthritis with the implementation of guideline recommended non-surgical treatments: a model-based health economic evaluation.

AIMS OF THE STUDY:Structured exercise, education, weight management and painkiller prescription are guideline recommended non-surgical treatments for patients suffering from knee osteoarthritis. Despite its endorsement, uptake of guideline recommended non-surgical treatments remains low. It is unknown whether the implementation of these treatments into the current model of care for knee osteoarthritis would be cost-effective from a Swiss statutory healthcare perspective. We therefore aimed to (1) assess the incremental cost-effectiveness ratio of an optimised model of care incorporating guideline recommended non-surgical treatments in adults with knee osteoarthritis and (2) the effect of total knee replacement (TKR) delay with guideline recommended non-surgical treatments on the cost-effectiveness of the overall model of care. METHODS:A Markov model from the Swiss statutory healthcare perspective was used to compare an optimised model of care incorporating guideline recommended non-surgical treatments versus the current model of care without standardised guideline recommended non-surgical treatments. Costs were derived from two Swiss health insurers, a national database, and a reimbursement catalogue. Utility values and transition probabilities were extracted from clinical trials and national population data. The main outcome was the incremental cost-effectiveness ratio for three scenarios: "base case" (current model of care vs optimised model of care with no delay of total knee replacement), "two-year delay" (current model of care vs optimised model of care + two-year delay of total knee replacement) and "five-year delay" (current model of care vs optimised model of care + five-year delay of total knee replacement). Costs and utilities were discounted at 3% per year and a time horizon of 70 years was chosen. Probabilistic sensitivity analyses were conducted. RESULTS:The "base case" scenario led to 0.155 additional quality-adjusted life years (QALYs) per person at an additional cost per person of CHF 341 (ICER = CHF 2,203 / QALY gained). The "two-year delay" scenario led to 0.134 additional QALYs and CHF -14 cost per person. The "five-year delay" scenario led to 0.118 additional QALYs and CHF -501 cost per person. Delay of total knee replacement by two and five years led to an 18% and 36% reduction of revision surgeries, respectively, and had a cost-saving effect. CONCLUSION:According to this Markov model, the optimisation of the current model of care by implementing guideline recommended non-surgical treatments would likely be cost-effective from a statutory healthcare perspective. If implementing guideline recommended non-surgical treatments delays total knee replacement by two or five years, the amount of revision surgeries may be reduced.