Abstract：
OBJECTIVE: Central cord syndrome (CCS) is expected to become the most common traumatic spinal cord injury, yet its optimal management remains unclear. This study aimed to evaluate variability in nonoperative vs operative treatment for CCS between trauma centers in the American College of Surgeons Trauma Quality Improvement Program, identify patient- and hospital-level factors associated with treatment, and determine the association of treatment with outcomes.
METHODS: Adults with CCS were identified from the Trauma Quality Improvement Program database (2014-2016). Mixed-effects modeling with a random intercept for trauma centers was used to examine the adjusted association of patient- and hospital-level variables with nonoperative treatment. The random-effects output of the model assessed the risk-adjusted variability in nonoperative treatment across centers. Outlier hospitals were identified, and the median odds ratio was calculated. The adjusted effect of nonoperative treatment on mortality, morbidity, and hospital length of stay (LOS) was examined at the patient and hospital level by mixed-effects regression.
RESULTS: Three thousand, nine hundred twenty-eight patients across 255 centers were eligible; of these, 1523 (38.8%) were treated nonoperatively. Older age, noncommercial insurance (odds ratio [OR] 1.26, 95% CI 1.08-1.48, P = .004), absence of fracture (OR 0.58, 95% CI 0.49-0.68, P < .001), severe head injury (OR 1.41, 95% CI 1.09-1.82, P = .008), and comatose presentation (1.82, 95% CI 1.15-2.89, P = .011) were associated with nonoperative treatment. Twenty-eight hospitals were outliers, and the median odds ratio was 2.02. Patients receiving nonoperative treatment had shorter LOS (mean difference -4.65 days). Nonoperative treatment was associated with lesser in-hospital morbidity (OR 0.49, 95% CI 0.37-0.63, P < .001) at the patient level. There was no difference in mortality.
CONCLUSIONS: Operative decision-making for CCS is influenced by patient factors. There remains substantial variability between trauma centers not explained by case-mix differences. Nonoperative treatment was associated with shorter hospital LOS and lesser inpatient morbidity.
METHODS: Adults with CCS were identified from the Trauma Quality Improvement Program database (2014-2016). Mixed-effects modeling with a random intercept for trauma centers was used to examine the adjusted association of patient- and hospital-level variables with nonoperative treatment. The random-effects output of the model assessed the risk-adjusted variability in nonoperative treatment across centers. Outlier hospitals were identified, and the median odds ratio was calculated. The adjusted effect of nonoperative treatment on mortality, morbidity, and hospital length of stay (LOS) was examined at the patient and hospital level by mixed-effects regression.
RESULTS: Three thousand, nine hundred twenty-eight patients across 255 centers were eligible; of these, 1523 (38.8%) were treated nonoperatively. Older age, noncommercial insurance (odds ratio [OR] 1.26, 95% CI 1.08-1.48, P = .004), absence of fracture (OR 0.58, 95% CI 0.49-0.68, P < .001), severe head injury (OR 1.41, 95% CI 1.09-1.82, P = .008), and comatose presentation (1.82, 95% CI 1.15-2.89, P = .011) were associated with nonoperative treatment. Twenty-eight hospitals were outliers, and the median odds ratio was 2.02. Patients receiving nonoperative treatment had shorter LOS (mean difference -4.65 days). Nonoperative treatment was associated with lesser in-hospital morbidity (OR 0.49, 95% CI 0.37-0.63, P < .001) at the patient level. There was no difference in mortality.
CONCLUSIONS: Operative decision-making for CCS is influenced by patient factors. There remains substantial variability between trauma centers not explained by case-mix differences. Nonoperative treatment was associated with shorter hospital LOS and lesser inpatient morbidity.
摘要：
目的：中央脊髓综合征（CCS）有望成为最常见的创伤性脊髓损伤，然而，其最佳管理仍不清楚。这项研究旨在评估美国外科医生学院创伤质量改善计划中创伤中心之间的CCS非手术治疗与手术治疗的差异，确定与治疗相关的患者和医院层面的因素，并确定治疗与结果的关联。
方法：从创伤质量改善计划数据库（2014-2016）中确定了患有CCS的成年人。使用创伤中心随机截取的混合效应模型来检查患者和医院水平变量与非手术治疗的校正关联。该模型的随机效应输出评估了各中心非手术治疗中风险调整后的变异性。确定了离群医院，并计算了中位比值比。非手术治疗对死亡率的调整效果，发病率，通过混合效应回归在患者和医院级别检查住院时间（LOS）。
结果：三千，255个中心的9128名患者符合资格；其中，1523例（38.8%）非手术治疗。年纪大了,非商业保险(赔率比[OR]1.26，95%CI1.08-1.48，P=.004)，无骨折（OR0.58，95%CI0.49-0.68，P<.001），重型颅脑损伤（OR1.41，95%CI1.09-1.82，P=.008），和昏迷表现（1.82，95%CI1.15-2.89，P=0.011）与非手术治疗相关。二十八家医院是离群值,中位比值比为2.02。接受非手术治疗的患者的LOS较短（平均差异-4.65天）。非手术治疗与患者的住院发病率较低相关（OR0.49，95％CI0.37-0.63，P<.001）。死亡率没有差异。
结论：CCS的手术决策受患者因素的影响。创伤中心之间仍然存在很大的差异，无法通过病例混合差异来解释。非手术治疗与较短的住院LOS和较低的住院发病率相关。
方法：从创伤质量改善计划数据库（2014-2016）中确定了患有CCS的成年人。使用创伤中心随机截取的混合效应模型来检查患者和医院水平变量与非手术治疗的校正关联。该模型的随机效应输出评估了各中心非手术治疗中风险调整后的变异性。确定了离群医院，并计算了中位比值比。非手术治疗对死亡率的调整效果，发病率，通过混合效应回归在患者和医院级别检查住院时间（LOS）。
结果：三千，255个中心的9128名患者符合资格；其中，1523例（38.8%）非手术治疗。年纪大了,非商业保险(赔率比[OR]1.26，95%CI1.08-1.48，P=.004)，无骨折（OR0.58，95%CI0.49-0.68，P<.001），重型颅脑损伤（OR1.41，95%CI1.09-1.82，P=.008），和昏迷表现（1.82，95%CI1.15-2.89，P=0.011）与非手术治疗相关。二十八家医院是离群值,中位比值比为2.02。接受非手术治疗的患者的LOS较短（平均差异-4.65天）。非手术治疗与患者的住院发病率较低相关（OR0.49，95％CI0.37-0.63，P<.001）。死亡率没有差异。
结论：CCS的手术决策受患者因素的影响。创伤中心之间仍然存在很大的差异，无法通过病例混合差异来解释。非手术治疗与较短的住院LOS和较低的住院发病率相关。
