Trends in Traumatic Spinal Cord Injury--Reply.

Trends in Traumatic Spinal Cord Injury To the Editor The use of the Nationwide Inpatient Sample (NIS) by Dr Jain and colleagues1 to estimate trends in the incidence of traumatic spinal cord injury in the United States raises several questions. Their finding of an incidence of 53 to 54 cases per 1 million inpatient hospitalizations appears only slightly lower than the cumulative incidence estimates from a study based on the Nationwide Emergency Department Sample (56.4 per 1 million), which used a more conservative approach to define acute traumatic spinal cord injury by including only persons with a principal diagnosis.2 To quantify traumatic spinal cord injury burden as accurately as possible would require a comprehensive nationwide survey similar to those used in prior regional, countylevel studies, which is an arduous task involving significant costs and resources.3,4 Hence, identifying a proxy approach to estimating nationwide traumatic spinal cord injury burden is essential. A primary concern with the use of data from inpatients only is the potential to substantially underestimate populationlevel incidence. Prior population-based regional studies suggest that up to 38% of individuals with traumatic spinal cord injury die before receiving definitive inpatient care.3,4 In addition, although the majority of patients presenting to the emergency department with such injuries are hospitalized at a shortterm hospital, others are not. Nearly 15% of survivors are not admitted to short-term facilities that form the basis of the NIS, instead undergoing transfers to long-term facilities (including specialty rehabilitation centers), being discharged, leaving against medical advice, or having no discharge destination recorded.2 Although patients who are not admitted to short-term facilities after emergency department presentation may have relatively minor injuries,2 it is still important that they be accounted for when discussing traumatic spinal cord injury incidence. For example, a patient with an incomplete L4 injury resulting from a fall may not be admitted to a short-term hospital but may require rehabilitation services in the future. Apart from injury severity, inpatient admission may also be influenced by insurance coverage, patient circumstances, family preferences, and physician recommendations. As such, using inpatient records to determine traumatic spinal cord injury incidence misses not only those who die prior to definitive inpatient care but also individuals with minor injuries and those who are unable to access inpatient care.