Modeling the Cost-Utility of Treatment Sequences for Multiple Sclerosis

Objectives: Most patients with multiple sclerosis (MS) switch between disease-modifying therapies (DMTs) during their lifetime. Our aim was to develop an MS cost-utility model that takes treatment switching into account to provide a more realistic estimate of treatment benefit than previous models that assume lifetime use of 1 DMT. Methods: A treatment sequence model using a microsimulation framework with a lifetime time horizon and a societal perspective was developed in R. Clinical plausibility and decision rules for switching were defined in consultation with Dutch MS neurologists. The ability of DMTs to prevent relapses and delay disease progression was modeled by applying DMTspecific estimates derived from a network meta-analysis of randomized controlled trials to natural history data. A total of 2 treatment strategies were compared: a first-line DMT sequence (peginterferon-glatiramer-teriflunomide-interferon-betadimethyl fumarate) and an escalation DMT sequence (peginterferon-glatiramer-ocrelizumab-natalizumab-alemtuzumab). Scenario analyses explored impact of alternative sources of natural history data, societal versus healthcare perspective, and condition-specific versus generic utilities. Predicted short-term switches (,5 years) were externally validated with Dutch claims data on DMT use. Results: Short-term switches predicted by the model compared well with Dutch claims data. Transition from relapsing remitting MS to secondary progressive MS was delayed by the escalation sequence (24.7 vs 20.3 years on first-line sequence). Model results were sensitive to utility values and medical resource consumption was a large driver of uncertainty. Conclusions: This microsimulation model overcomes the limitation of previous models by modeling treatment sequences. Because it better reflects clinical reality, it facilitates incorporating cost-utility information in clinical guidelines.

Automated summary

A cost-utility model for multiple sclerosis considering treatment switching was developed to provide more realistic estimates of treatment benefits. Short-term treatment switches were predicted accurately, with the escalation DMT sequence showing delayed disease progression. Model results were sensitive to utility values and medical resource consumption.