Traumatic brain injury (TBI) mortality rates are rising among several subgroups of the US population [1], necessitating a better understanding of concomitant morbidity trends. Nationwide estimates confirm nonfatal TBI is both common and a costly condition to treat [2]. Efforts to assess TBI injury trends are complicated by the removal of unspecified head injuries from the CDC’s hospital-based TBI surveillance case definition following the transition from ICD-9-CM to ICD-10-CM [3]. This removal was due to the increased diagnostic information found in morbidity data under the ICD-10-CM coding scheme, in an attempt to promote the use of more specific TBI-related diagnosis codes [3].

Within the ICD-9-CM TBI surveillance case definition, the unspecified head injury code (959.01) accounted for approximately 50 % of all flagged visits and had comparable sensitivity to concussion-related diagnosis codes for mild TBIs (23 %) [4]. Relative to the ICD-9-CM unspecified head injury code, the excluded ICD-10-CM code (S09.90) had improved positive predictive value, increasing from 20 % to 36–52 % [4], [5]. Wharton et al. (2025) found S09.90 codes were primarily used due to insufficient provider documentation and time constraints [6]. ED visits where loss of consciousness was not adequately detailed and those with “probable” or “suspected” TBIs were commonly identified using the S09.90 code [6].

Contrary to the stated goals of the ICD-10-CM nonfatal TBI surveillance case definition [3], unspecified head injuries remain the most prevalent ICD-10-CM head injury diagnosis code and their removal from the definition led to a marked decrease in the frequency of TBI-related emergency department (ED) visits flagged post-transition [7]. Differences in the magnitude and trend of TBI ED visits with and without unspecified head injury codes may be compounded by significant societal events such as the COVID-19 pandemic, which significantly reduced ED volume [8], [9]. However, little is known regarding the longer-term impact of the code’s removal on TBI injury surveillance. Thus, we investigated variation in trends of TBI ED visits in North Carolina (NC) across several interruption periods based on different surveillance case definitions.