STUDY DESIGN: This is a survey-based study.
OBJECTIVE: To investigate the practice of spinal cord injury (SCI) core data collection by Chinese physicians to measure the extent and accuracy of routine collection of elements contained in the International Spinal Cord Injury Core Data Set (ISCICDS).
SETTING: This study was conducted in a workshop in Peking University, China.
METHODS: During an SCI workshop, a survey questionnaire was administered to 48 physicians from 20 provinces of China. The questions were developed on the basis of the data elements within the ISCICDS including the following issues: date of birth, injury, acute admission and inpatient discharge, total hospitalized days, gender, injury etiology, vertebral injury, associated injury, spinal surgery, ventilatory assistance and place of discharge. In addition, data collection practice on neurologic examinations including date, neurological level, injury severity and frequency of examination were involved.
RESULTS: The self-reported practice of data collection regarding date of birth, acute admission and inpatient discharge, gender, vertebral injury, associated injury, spinal surgery and frequency of neurological examination are consistent with the information in the ISCICDS among the majority (⩾76%) of physicians. However, only gender, vertebral injury, associated injury and spinal surgery are completely consistent. The consistency percentages of other data elements ranged from 39.5 to 66.8%.
CONCLUSION: Apart from four data elements, which were collected consistently with the intention in the ISCICDS, the collection of other core data elements need to be documented according to the guidelines included in the ISCICDS to ensure consistency of practice among Chinese physicians and to support worldwide comparison of SCI data.Suggestion:Only four data elements are collected in complete accordance with the ISCICDS by Chinese physicians. ISCICDS guidelines for the remaining elements need to be more rigorously adhered to in order to promote consistency and comparability of data.