USING DISCRETE EVENT SIMULATION TO IMPROVE ACUTE STROKE CARE QUALITY MEASUREMENT

Purpose: Time from stroke onset to the administration of tissue plasminogen activator (tPA) is an important acute stroke care performance measure, supported by clinical and cost-effectiveness evidence showing the value of tPA within 3 hours of ischemic stroke. However, performance measurement should be adjusted for the operational characteristics of hospitals to avoid setting unrealistic benchmarks for smaller hospitals, particularly if performance affects payment. We sought to model how time-to-tPA is subject to hospital size and stroke-related resources, and proposed a model-based capacity-adjustment framework to promote fairness in performance measurement across different hospital types.

Method: We developed a discrete event simulation model to compare the time-to-tPA measures among four types of hospitals with varying sizes and stroke-related resources (non-stroke center, acute stroke-ready hospital, primary stroke center, and comprehensive stroke center). In the model, stroke patients arrive at an emergency department (ED) and sequentially go through triage, initial assessment, CT scanning and interpretation, and (if eligible) tPA administration. We used a Poisson process (lambda = 5.3) to model stroke patient arrival, and assigned an "onset-to-arrival time" based on a uniform distribution (0,24 hours). Patient waiting time at each step in the ED depends on the hospital-specific resource capacity such as the number of beds and CTs and the availability of out-of-hours scanning and interpretation. Eligible patients only receive tPA if the total time-to-tPA is <4.5 hours. All model parameters, such as arrival rates and process times, were estimated from published sources.

Result: Proportion of stroke patients who received tPA was 3.7%, 4.5%, 5.8%, and 8.0% for non-stroke center, acute stroke-ready hospital, primary stroke center, and comprehensive stroke center, respectively. Average door-to-needle-time among those treated was 75, 64, 60, and 52 minutes for the four types of hospitals. In order to achieve the recommended door-to-needle time of 60 minutes for 50% among patients who receive tPA on time, the hospital size should be >1023 beds for comprehensive stroke center designation.

Conclusion: Based on capacity features, larger comprehensive stroke centers can achieve better acute stroke performance on time-to-tPA measures compared to smaller hospitals and non-stroke centers. In order for time-to-tPA to be used in a pay-for-performance reimbursement policies, smaller hospitals that meet minimum time-to-treatment standards should be held to their best achievable performance target as opposed to one-size-fits-all targets.