Abstract

OBJECTIVE Previous studies of neurosurgical transfers indicate that substantial numbers of patients may not need to be transferred, suggesting an opportunity to provide more patient-centered care by treating patients in their communities, while probably saving thousands of dollars in transport and duplicative workup. This study of neurosurgical transfers, the largest to date, aimed to better characterize how often transfers were potentially avoidable and which patient factors might affect whether transfer is needed. METHODS This was a retrospective cohort study of neurosurgical transfers to an urban, tertiary-care, level I trauma center between October 1, 2017, and October 1, 2022. Prior to data analysis, the authors devised criteria to differentiate necessary neurosurgical transfers from potentially avoidable ones. A transfer was considered necessary if 1) the patient went to the operating room within 12 hours of arrival at the emergency department (ED); 2) a neurological MRI study was conducted in the ED; 3) the patient was admitted to the ICU from the ED; or 4) the patient was admitted to either neurology or a surgical service (including neurosurgery). Transfers not meeting any of the above criteria were deemed potentially avoidable. Patient and clinical characteristics, including diagnostic groupings from Clinical Classification Software categories, were collected retrospectively via electronic health record data abstraction and stratified by whether the transfer was necessary or potentially avoidable. Statistical differences were assessed with a chi-square test. RESULTS A total of 5113 neurosurgical transfers were included in the study, of which 1701 (33.3%) were classified as potentially avoidable. Four percent of all transferred patients went to the operating room within 12 hours of reaching the receiving ED, 23.4% were admitted to the ICU from the ED, 26.6% had a neurological MRI study performed in the ED, and 54.4% were admitted to a surgical service or to neurology. Potentially avoidable transfers had a higher proportion of traumatic brain injury, headache, and syncope (p < 0.0001), as well as of spondylopathies/spondyloarthropathies (p = 0.0402), whereas patients needing transfer had a higher proportion of acute hemorrhagic cerebrovascular disease and cerebral infarction (p < 0.0001). CONCLUSIONS This study demonstrates that a large number of neurosurgical transfers can probably be treated in their home hospitals and highlights that the vast majority of patients transferred for neurosurgical conditions do not receive emergency neurosurgery. Further research is needed to better guide transferring and receiving facilities in reducing the burden of excessive transfers.