Selection for early surgery in asymptomatic mitral regurgitation: A Markov model
Background: Current guidelines propose mitral valve repair in asymptomatic chronic mitral regurgitation (MR) when the likelihood of repair is 90% or more. As this figure is not evidence-based, we sought whether the results of a decision-analytic model could facilitate the selection between early surgery (ES) and watchful waiting (WW) based on current guidelines. Methods: A Markov model was developed to reflect the anticipated health states in MR (pre-operative, post-operative, post-complication and death). Risks and transitions were informed by the literature. Implications of the strategies for survival, quality-adjusted life years (QALYs), cost and cost-effectiveness were calculated from a US healthcare provider perspective. Results: In the reference case (90% repair), QALY with ES was superior to WW (11.2 [0.4-21.3] vs 10.7 [95%CI: 1.0-21.3]) at an incremental cost-effectiveness of $54,659 ($45,030-$64,288) per QALY. Sensitivity analyses of health benefit showed the main variables influencing outcome were repair rate, operative mortality and risks of heart failure and death with medical management. At the registry repair rate (50%), outcomes of ES were worse than WW, and threshold analysis showed that a repair rate of 84% was required for ES to be superior. High medical risk (yearly heart failure risk 5.6 ± 6.6% and mortality 2.5 ± 4%) was the most favorable scenario for surgery; ES was more effective when mortality in the WW group was > 3.5%/year. Conclusion: A Markov model might be used to guide the selection of asymptomatic patients for mitral repair, based on local variations in risk and complications as well as repair rate.
|Keywords||Cost-effectiveness, Mitral, Outcomes, Probabilistic sensitivity analysis, Regurgitation, Surgery|
|Persistent URL||dx.doi.org/10.1016/j.ijcard.2011.08.048, hdl.handle.net/1765/30814|
Marwick, T., Scuffham, P.A., & Hunink, M.G.M.. (2011). Selection for early surgery in asymptomatic mitral regurgitation: A Markov model. International Journal of Cardiology. doi:10.1016/j.ijcard.2011.08.048