Development a Risk Scoring System for Predicting the Risk of Poststroke Epilepsy to Improve the Quality of Care among Patients with Stroke

Background and Objective. Stroke is a major cause of epilepsy in older adults and the elderly. About 10% of all stroke patients experience seizures, from stroke onset until several years later. After stroke, seizures and the subsequent development of epilepsy may impede patients’ recovery. Health care systems will encounter an increasing proportion of patients with poststroke epilepsy (PSE) because the population of stroke survivor increases. Due to a relatively small sample size of this population, currently no prognostic model has demonstrated its performance as a useful role. Although antiepileptic drugs (AEDs) are available for treating patients with PSE, clinical guidelines do not recommend routinely prescribing AEDs for preventing PSE occurrence because there is no sufficient evidence of PSE prevention by AEDs, and AEDs have high incidence of adverse effects and drug-drug interactions. From the public health perspective, avoiding risk factors, identifying high risk patients, and applying in individualized patient management before PSE occurrence become very important. Thus, the goal of the 3-year proposal is to develop a risk scoring system for predicting the risk of PSE among patients with acute stroke. Specific aims are following: ■ To develop a risk scoring system for predicting the risk of PSE based on claim-based data. ■ To develop a risk scoring system for predicting the risk of PSE based on both hospital-based and claim-based data. ■ To conduct a prospective cohort to valid the performance of PSE risk score. Materials and Methods. Patient-level data from National Health Insurance Research Database (NHIRD), Taiwan Stroke Registry (TSR) and Taipei Medical University (TMU) affiliation hospitals will be used to achieve the aims. The specific works are following: ■ 1st year - A retrospective cohort design will be performed to investigate and identify the most important risk factors associated with PSE based on NHIRD claims. The validation of the predictive model will also be performed. ■ 2nd year - A case-control study will be conducted to compare baseline characteristics of PSE case patients with that of non-PSE control patients in TMU affiliation hospitals. Besides, we will link the data of TSR and NHIRD to develop PSE risk score on both hospital- and claim-based data. The performance of different predictive models will be also validated internally. A prospective follow-up cohort will be prepared to assess the performance of different predictive models externally. ■ 3rd year -Acute stroke patients who admitted to TMU affiliation hospitals will be classified into high risk, medium risk and low risk group of PSE, and the status of PSE will be confirmed based on either medical records or telephone interview within one-year follow-up period. The most appropriate predictive model of PSE will be assessed in the prospective cohort.