Objective  To analyze the factors affecting the neurological outcome of inpatients with traumatic spinal cord injury (TSCI) in different hospitals, and explore the future prevention and treatment of TSCI.

  Methods  Multistage stratified cluster sampling method was used to extract the medical record information of TSCI patients hospitalized in 69 grade Ⅲ hospitals from January 1 to December 31, 2018. Wilcoxon and Kruskal-Wallis H rank sum test was used to analyze the changes in injury grade of TSCI patients at discharge and admission according to American Spinal Injury Association (ASIA) injury classification; the two-level logistic random intercept model was used to analyze the influencing factorls of the changes in injury grade of TSCI patients according to ASIA injury classification in different hospitals.

  Results  Among the 4404 TSCI patients, 1200 were in recovery group, accounting for 27.25%, and the difference in injury grade distribution at discharge and admission was significant. In the total variation of ASIA injury grade change, 24.12% was caused by hospital-level differences. In multivariate analysis, the grading improvement rate decreased by 0.90% for one-year increase of age. Compared with the lumbosacral segment. The grading improvement rates in the chest and neck segments were reduced by 28.90% and 21.50%, respectively. The grading improvement rate in patients with complete injury was 29.00% lower than that in patients with incomplete injury. The grading improvement rate in patients with fracture and dislocation was 33.50% lower than that in patients without these conditions. The grading improvement rate in patients with interval between injury and operation ≤3 days was 0.353 times higher than that in patients with the interval >3 days. The grading improvement rate in patients with rehabilitation training increased by 1.069 times. Compared with patients without surgery, the grading improvement rate in patients with decompression fixation increased by 1.565 times, and the rate in those with decompression fixation and fusion increased by 1.139 times.

  Conclusion  The effect of TSCI inpatient treatment was poor; the differences in age, injury and treatment characteristics at different hospital levels were all related to the prognosis of TSCI, indicating that it is urgent to develop precise TSCI prevention strategies and establish a regional integration of “emergency-treatment-rehabilitation” platform.