Development, validation, and application of a microsimulation model to predict stroke and mortality in medically managed asymptomatic patients with significant carotid artery stenosis.

OBJECTIVE: To develop a model to predict stroke-free survival and mortality over a multiyear time frame for a trial-excluded population of medically managed asymptomatic patients with significant carotid artery stenosis. METHODS: We calibrated, validated, and applied a Monte Carlo microsimulation model. For calibration we adjusted general-population mortality and stroke risks to capture these risks specific to asymptomatic carotid stenosis patients. For validation, we compared model-predicted and actual stroke-free survival curves and stroke counts from a population of comparable patients. For application, the validated model predicted stroke-free survival for a hypothetical medically managed arm of a recent single-arm carotid revascularization trial. RESULTS: For each month in the 60-month time frame, the model-predicted and actual calibration trial stroke-free survival curves were not statistically different (P > 0.62). In validation, the calibrated model's stroke-free survival curvematched the actual curve from an independent population; beyond 24 months, the model-predicted and actual curves were not statistically different (P > 0.32). We also compared model-predicted and actual number of strokes from the independent trial. The model predicted 187.25 strokes (95% confidence interval 161.49-213.01), while the actual number was 171.6, within 1.22 standard deviations of the simulated mean. CONCLUSIONS: Given the absence of medically managed populations in recent carotid stenosis trials, our model can estimate stroke-free survival and mortality data for these patients. The model may also estimate the effectiveness of novel medical and procedural therapies for stroke prevention. These effectiveness estimates can inform the development of policies, guidelines, or cost-effectiveness analyses when only single-arm trial data exist.

Duke Scholars

Author Gregory P. Samsa Biostatistics & Bioinformatics, Division of Biostatistics