Estimation of survival probabilities for use in cost-effectiveness analysis: a comparison of a multi-state modelling survival analysis approach with partitioned survival and Markov decision-analytic modelling

摘要

Modelling of clinical-effectiveness in cost-effectiveness analysis typically involves some form of partitioned survival or Markov decision-analytic modelling. The health states progression-free, progression and death and the transitions between them are frequently of interest. With partitioned survival, progression is not modelled directly as a state; instead time in that state is derived from the difference in area between overall survival and progression-free survival curves. With Markov decision-analytic modelling, a priori assumptions are often made with regard to the transitions rather than using the individual patient data directly to model them. This article compares a multi-state modelling survival regression approach to these two common methods. As a case study, we use a trial comparing rituximab in combination with fludarabine and cyclophosphamide versus fludarabine and cyclophosphamide alone for the first line treatment of chronic lymphocytic leukaemia. We calculated mean Life Years and QALYs that involved extrapolation of survival outcomes in a trial. We adapted an existing multi-state modelling approach to incorporate parametric distributions for transition hazards, to allow extrapolation. The comparison showed that, due to the different assumptions used in the different approaches, a discrepancy in results was evident. The partitioned survival and Markov decision-analytic modelling deemed the treatment cost-effective with ICERs of just over £16,000 and £13,000 respectively. However, the results with the multi-state modelling were less conclusive with an ICER of just over £29,000. This work has illustrated that it is imperative to check whether assumptions are realistic as different model choices can influence clinical and cost-effectiveness results.