model with incremental cost-effectiveness ratios or quality-adjusted life-years as outcome measures. Specific information on the decision-analytic models, including the modeling approach, perspective, population studied, interventions and data sources were extracted from each article. RESULTS: 1834 citations were screened and 68 full-text articles retrieved. Overall, 31 model-based studies were identified for extraction: 29 cohort-based state-transition models, one decision tree model, and one other cost-effectiveness model. Comparators included drugs such as dopamine agonists, entacapone, and rasagaline, as well as surgically implanted devices such as deep brain stimulation and levodopa-carbidopa intestinal gel. Overall, 10 modeled only early PD, 7 focused on advanced PD, 10 included early and late stage PD, and 4 did not specify stage. The most commonly modeled health outcomes were Hoehn and Yahr stage progression and % off-time. Other outcomes include motor complications, fluctuations, dyskinesia, falls, and dementia. CONCLUSIONS: The models identified typically had simple frameworks. Many projected disease progression from short-term clinical trial data alone and did not leverage real world observational data now available with longer longitudinal follow-up. Clinical efficacy was often applied by affecting only the initial distribution across health states, which likely does not fully capture the benefits of a treatment. Future studies should explore developing individual patient simulations to be able to more realistically represent the heterogeneity observed in the clinical manifestations and progression rates of the disease, as well as capture the potential benefits and risks of symptomatic or disease modifying treatments.

No abstract