Objective: To comprehensively synthesize the evolution of health-economic evaluation models (HEEMs) of all osteoarthritis (OA) interventions, including preventions, core treatments, adjunct nonpharmacologic interventions, pharmacologic interventions, and surgical treatments.
Methods: The literature was searched within health-economic/biomedical databases. Data extracted included OA type, population characteristics, model setting/type/events, study perspective, and comparators; the reporting quality of the studies was also assessed. The review protocol was registered at the International Prospective Register of Systematic Reviews (CRD42018092937).
Results: Eighty-eight studies were included. Pharmacologic and surgical interventions were the focus in 51% and 44% of studies, respectively. Twenty-four studies adopted a societal perspective (with increasing popularity after 2013), but most (63%) did not include indirect costs. Quality-adjusted life years was the most popular outcome measure since 2008. Markov models were used by 62% of studies, with increasing popularity since 2008. Until 2010, most studies used short-to-medium time horizons; subsequently, a lifetime horizon became popular. A total of 86% of studies reported discount rates (predominantly between 3% and 5%). Studies published after 2002 had a better coverage of OA-related adverse events (AEs). Reporting quality significantly improved after 2001.
Conclusion: OA HEEMs have evolved and improved substantially over time, with the focus shifting from short-to-medium-term pharmacologic decision-tree models to surgical-focused lifetime Markov models. Indirect costs of OA are frequently not considered, despite using a societal perspective. There was a lack of reporting sensitivity of model outcome to input parameters, including discount rate, OA definition, and population parameters. While the coverage of OA-related AEs has improved over time, it is still not comprehensive.
© 2020, American College of Rheumatology.