国家神经系统疾病临床医学研究中心数据管理与共享系统

发表杂志：BMC Medical Research Methodology

发表类型: Article

影响因子(IF): 4.612

PMID: 35842606 Pubmed原文链接 全文下载

Stroke is the second leading cause of death worldwide and the leading cause of mortality and disability in China[1,2]. In previous studies in multiple countries, the recurrence rate of stroke 90 days after TIA/minor stroke is 18%-20%[3]. About 40% of stroke survivors have a poor prognosis (modified Rankin Scale [mRS] score ≥3) between 1 month and 5 years after stroke[4]. In the past clinical work, both patients and medical workers tend to pay more attention to the secondary prevention of non-minor stroke patients, ignoring the poor prognosis of minor stroke. According to the data of the Third China National Stroke Registry (CNSR-III), TIA and minor stroke (a National Institutes of Health Stroke Scale (NIHSS) score≤5) account for about 73% of acute ischemic stroke cases. Therefore, it is very important to predict the prognosis of patients with TIA and minor stroke, find risk factors, identify high-risk patients, and accurately carry out early intervention for patients with TIA and minor stroke.With the improvement of computer computing power, the advent of the era of big data, and the update of algorithms, machine learning has made good progress in the field of disease prediction, but there are few comparative studies on multiple tree model machine learning algorithms. This study used the CNSR-III database to focus on factors related to the 90-day prognosis of TIA and minor stroke patients and compared the predictive performance of machine learning models and the Logistic model to provide references for related research and clinical work.

We selected TIA and minor stroke patients from a prospective registry study(CNSR-III).Demographic characteristics,smoking history,drinking history(>=20g/day),physiological data,medical history,secondary prevention treatment,in-hospital evaluation and education,laboratory data,neurological severity,mRS score and TOAST classification of patients were assessed. Univariate and multivariate logistic regression analyses were performed to identify predictors associated with the poor prognosis (mRS≥3). The predictors were used to establish machine learning models and the traditional Logistic model, which were randomly divided into the training set and test set according to the ratio of 7:3. The training set was used to construct the prediction model, and the test set was used to evaluate the effect of the model. The evaluation indicators of the model included the area under the curve (AUC) of the discrimination index and the Brier score of the calibration index.