Electronic health records of stroke convalescent patients will be selected from the third affiliated hospital of Zhejiang Chinese Medical University from January 2012 to December 2019. Relevant clinical data will be extracted according to case data, and model-based decision tree analysis will be conducted to establish and evaluate the clinical prediction Model.

Cases will be screened according to inclusion and exclusion criteria.Patients' basic information, risk factors, types of stroke, TCM diagnosis, all treatment during hospitalization, neurological impairment evaluation, imaging examination results and other contents will be recorded.Finally, statistical processing will be carried out to establish the prediction model, draw ROC curve and evaluate the model.

Inclusion Criteria:

• Clinical diagnosis of stroke
• The incidence is between 30 days and 180 days
• Age ≥ 18 years old
• The first onset

Exclusion Criteria:

• Subarachnoid hemorrhage
• transient ischemic attack
• other intracranial lesions
• Non-atherosclerotic thrombotic cerebral infarction
• Surgery, trauma, congenital disability leading to limb dysfunction

The study population will be stroke convalescent patients from The Third Affiliated Hospital of Zhejiang Chinese Medical University from January 2012 to December 2019.