Long-term survival and five year hospital resource usage following traumatic brain injury in Scotland from 1997 to 2015: A population-based retrospective cohort study.

BACKGROUND

It is unclear if traumatic brain injury (TBI) results in excess mortality compared with head injury without injury to neural structures (HI). Because TBI populations exhibit significant demographic differences from uninjured populations, to determine the effect of TBI on survival, it is essential that a similarly injured control population be used. We aimed to determine if survival and hospital resource usage differ following TBI compared with HI.

METHODS

This retrospective population-based cohort study included all 25 319 patients admitted to a Scottish NHS hospital from 1997 to 2015 with TBI. Participants were identified using previously validated ICD-10 based definitions. For comparison, a control group of all 194 049 HI cases was also identified. Our main outcome measures were hazards of all-cause mortality for patients with TBI, compared with those with HI, over the 18-year follow-up period; and odds of mortality at one month post-injury. Number of days spent as inpatients and number of outpatient attendances per surviving month post-injury were used as measures of resource utilisation.

RESULTS

The adjusted odds ratio for mortality in the first month post-injury for TBI, compared with HI, was 7.12 (95% confidence interval [CI] 6.73-7.52; p < 0.001). For the remaining 18-year study period, the hazards of morality after TBI were 0.93 (CI 0.90-0.96; p < 0.001). During the five-year post-injury period, brain injury was associated with 2.15 (CI 2.10-2.20; p < 0.001) more days spent as inpatient and 1.09 times more outpatient attendances (CI 1.07-1.11; p < 0.001) compared with HI.

CONCLUSIONS

Although initial mortality following TBI is high, survivors of the first month post-injury can achieve comparable long-term survival to HI. However, this is associated with, and may require, increased utilisation of hospital services in the TBI group.