Treatment of peripheral neuropathic pain: a simulation model.

OBJECTIVE: To describe the use of a generalizable stochastic-simulation model of the treatment of neuropathic pain associated with peripheral neuropathies. METHODS: We developed a model to simulate treatment outcomes in a hypothetical cohort of patients with peripheral neuropathies. Each patient was randomly assigned an average pretreatment daily pain score (on a 0-10 scale), based on an assumed distribution of mean pretreatment pain scores in the cohort. Patients were randomly assigned daily pain scores, based on their pretreatment average and an assumed distribution of daily pain scores around this mean. Treatment outcomes were then simulated using the expected mean change (vs. baseline) in pain scores. Model outcomes include the expected increase in days with no or mild pain (score < or = 3), days with > or = 30% and > or = 50% reductions in pain intensity, and days with 2- and 3-point absolute reductions in pain intensity. To illustrate its use, the model was estimated over a 12-week period using data from a recent clinical trial of a new antiepileptic (pregabalin). RESULTS: Treatment over 12 weeks (84 days) was projected to result in 26 (+/-0.4) (mean [+/-SE]) additional (vs. no treatment) days with no or mild pain, 33 (+/-0.5) days with a > or = 30% reduction in pain intensity, 28 (+/-0.4) days with > or = 50% reduction in pain intensity, and 34 (+/-0.5) and 30 (+/-0.5) days with > or = 2-point and > or = 3-point absolute reductions in pain intensity. CONCLUSIONS: When combined with data on health-state utilities and treatment costs, this new analytical tool can provide a foundation for formal cost-effectiveness evaluations of interventions for painful peripheral neuropathies.