Methods: The state trauma registry and administrative hospital discharge records were compared for each injured patient who died at a subset of Level One Trauma Centers in Massachusetts during FY2010 and FY2011. The trauma registry sets no limit on number of diagnoses while hospital discharge data sets a limit to 15 diagnoses.
Results: Records of the same 462 patients from the trauma registry and the administrative hospital discharge data were linked. The number of injury diagnoses in the hospital discharge data had a mean of 3,5 (STD 2.7) ranging from 1 injury to 14 injuries while the TR had a mean of 5.8 injuries (STD 4.6) ranging from 1 injury to 26 injuries. The trauma registry data had 2,855 injury diagnoses and 519 distinct anatomical injury types. In comparison, the HD data included 1,640 total injury diagnoses and 449 distinct injury types. Certain anatomic types of damage were less likely to be extensively enumerated in the hospital discharge data than in the trauma registry data, including the full extent of C1-C7 and T1-T12 cord level damage, closed and open fractures of the skull vault and base and specificity on subarachnoid, subdural, and extradural hemorrhaging.
Conclusion: The highly variable number of diagnostic fields in administrative hospital data state-by-state may affect the utility and validity of the data in severity adjusted injury disparity studies. Policy makers should consider the information gained in using trauma registry data and lifting the limits to administrative hospital discharge data to improve the utility of such data for injury research. This is especially true with the upcoming United States implementation of ICD-10-CM which expands diagnostic coding schema to include anatomic laterality of injuries and all components of the Glasgow Coma Score.