PATIENT-LEVEL SIMULATION MODELS FOR COST-EFFECTIVENESS ANALYSIS IN MULTIPLE SCLEROSIS: A SYSTEMATIC REVIEW
Author(s)
Attila Imre, PharmD1, Balázs Nagy, PhD1, Rok Hren, PhD2.
1Center for Health Technology Assessment, Semmelweis University, Budapest, Hungary, 2Syreon Research Institute, Budapest, Hungary.
1Center for Health Technology Assessment, Semmelweis University, Budapest, Hungary, 2Syreon Research Institute, Budapest, Hungary.
OBJECTIVES: Patient-level simulation models (microsimulation, discrete event simulation, agent-based) offer structural advantages over cohort Markov models for evaluating treatment sequences, patient heterogeneity, and time-varying effects in multiple sclerosis (MS). Prior reviews of MS economic models have not focused specifically on patient-level frameworks. This systematic review identified and characterized patient-level simulation models applied in MS cost-effectiveness analysis.
METHODS: MEDLINE, Embase, and Web of Science were searched on 7 February 2026, with backward and forward citation chasing. Studies using a patient-level simulation framework to evaluate disease-modifying therapies for MS and reporting cost-effectiveness outcomes were eligible. The review followed PRISMA 2020. Extracted data included model structure, sources, validation approaches, software platforms, and disease dimensions represented within the models. Each study was assessed against the CHEERS 2022 checklist.
RESULTS: From 631 screened records, 14 publications describing seven distinct models met the inclusion criteria: two multi-publication families (ErasmusMC/iMTA microsimulation, seven publications; Montgomery discrete event simulation, two publications) and five independently developed frameworks. Eleven publications appeared after 2020. All models represented disability through the Expanded Disability Status Scale (EDSS) and incorporated relapses; none captured cognition, fatigue, vision, or psychological outcomes. R was the dominant software platform (nine studies). Eight of the 14 studies used the British Columbia MS database for natural history. Only three studies reported external validation, and only one made its model code publicly available. CHEERS 2022 assessment showed that the number of fully reported items ranged from 8 to 19 of 28, with no study reporting a health economic analysis plan, distributional effects, or (with one exception) patient involvement.
CONCLUSIONS: Patient-level simulation remains uncommon in MS economic modelling but has addressed decision problems, notably treatment sequencing with patient-specific history, that cohort Markov models accommodate less readily. Future work should strengthen validation reporting, expand outcomes beyond EDSS, increase code availability, and maintain structural diversity across independent frameworks.
METHODS: MEDLINE, Embase, and Web of Science were searched on 7 February 2026, with backward and forward citation chasing. Studies using a patient-level simulation framework to evaluate disease-modifying therapies for MS and reporting cost-effectiveness outcomes were eligible. The review followed PRISMA 2020. Extracted data included model structure, sources, validation approaches, software platforms, and disease dimensions represented within the models. Each study was assessed against the CHEERS 2022 checklist.
RESULTS: From 631 screened records, 14 publications describing seven distinct models met the inclusion criteria: two multi-publication families (ErasmusMC/iMTA microsimulation, seven publications; Montgomery discrete event simulation, two publications) and five independently developed frameworks. Eleven publications appeared after 2020. All models represented disability through the Expanded Disability Status Scale (EDSS) and incorporated relapses; none captured cognition, fatigue, vision, or psychological outcomes. R was the dominant software platform (nine studies). Eight of the 14 studies used the British Columbia MS database for natural history. Only three studies reported external validation, and only one made its model code publicly available. CHEERS 2022 assessment showed that the number of fully reported items ranged from 8 to 19 of 28, with no study reporting a health economic analysis plan, distributional effects, or (with one exception) patient involvement.
CONCLUSIONS: Patient-level simulation remains uncommon in MS economic modelling but has addressed decision problems, notably treatment sequencing with patient-specific history, that cohort Markov models accommodate less readily. Future work should strengthen validation reporting, expand outcomes beyond EDSS, increase code availability, and maintain structural diversity across independent frameworks.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE429
Topic
Economic Evaluation, Methodological & Statistical Research, Study Approaches
Disease
Neurological Disorders, Systemic Disorders/Conditions (Anesthesia, Auto-Immune Disorders (n.e.c.), Hematological Disorders (non-oncologic), Pain)