The baseline data of included patients were collected by trained research coordinators following a standard data collection protocol that was developed by the steering committee. The study investigators and research coordinators were trained before the kick-off meeting. Trained research coordinators at each site identified eligible patients, obtained informed consent, enrolled consecutive patients, and collected data by face-to-face interviews with the patients.
Information including prehospital care, prestroke modified Rankin Scale (mRS), National Institutes of Health Stroke Scale (NIHSS) score, Age, Blood pressure, Clincial features, Duration of symptoms and presence of Diabetes (ABCD2) score was collected through a direct interview by trained research coordinators at admission. Aetiology classification of ischaemic stroke was performed according to the TOAST (Trial of Org 10172 in Acute Stroke Treatment) criteria.15 (link) Other data were extracted from medical records that include patient demographics, medical history, family history, previous medication, physical examination, primary diagnosis, laboratory tests and risk factor assessment. At discharge, the research coordinators extracted the auxiliary examination and recorded standard aetiological evaluation result, medication, vascular related operation and surgical procedures, final diagnosis, NIHSS and mRS score, economic burden, cerebrovascular events during hospitalisation. Details of information collected at admission and discharge are presented in online appendix S2.
An electronic data capture system (EDC) was developed and used for data collection. Each participating hospital site entered data using an electronic signature (unique username and password). Trained research coordinators can store data in Pads and then upload the data later. The system can remind the researchers of patients’ follow-up spot, provide feedback of uploaded data timely and facilitate data quality monitoring. All data elements from each patient were automatically checked for completeness, correct coding, value range and logical error through EDC. All data changes made had an electronic audit trail with electronic signature and date. All laboratory test and auxiliary examination results were uploaded to the EDC as pictures. Considering that researchers had varying levels of comfort with the mobile technology, paper-based case report forms were also offered as a supplement if necessary. Independent data monitoring was also performed through EDC by an independent contract research organisation throughout the study period. All data were de-identified before data analysis.