ABSTRACT
Objective
Although recognised as a leading cause of death and disability, it is unclear whether current treatment strategies have improved outcomes from blunt moderate to severe traumatic brain injuries (msTBI) due to a lack of available epidemiological data. This study sought to determine if in‐hospital mortality and early functional outcomes of msTBI patients have changed over 20 years. Early functional outcomes were determined through the surrogate measure of the proportion of patients discharged directly to home.
Methods
A longitudinal cohort study using The Alfred Health Trauma Registry data. Subjects were aged 16–70 years with blunt msTBI (Head Abbreviated Injury Scale 3–5) admitted to The Alfred Trauma Service between 1 January 2002 and 31 December 2021.
Results
A total of 8222 patients presented with msTBI over the period, with 692 deaths (8.5%). Unadjusted in‐hospital mortality decreased from 12.1% in 2002 to 4.0% in 2021 (OR 0.96, 95% CI 0.95–0.97, p < 0.001). Sensitivity analysis demonstrated a statistically significant reduction in the odds of death even after adjusting for injury severity (OR 0.96, 95% CI 0.93–0.99, p < 0.0001). The proportion of patients discharged to home increased from 30.6% to 60.8% over the period (OR 1.05, 95% CI 1.04–1.06, p < 0.001).
Conclusion
There was a reduction in in‐hospital mortality and an increase in the proportion discharged home for adults hospitalised with msTBI at The Alfred Hospital over the last two decades. These results foster optimism for the care of neurotrauma patients, but further research is required to identify the modifiable factors associated with these improved outcomes.
Keywords: epidemiology, longitudinal analysis, neurotrauma, Trauma, Trauma systems
1. Introduction
Traumatic brain injury (TBI) is the leading cause of injury‐related mortality and disability [1]. Annual estimates show that 69 million people experience TBI [2] and an estimated one million people die from TBI worldwide [1]. Compared to mild TBI with a case fatality rate of less than 1%, moderate to severe TBI (msTBI) has a markedly higher fatality of 20%–40% [3]. Survivors of msTBI also commonly experience life‐long disability and impaired functional status [1].
People who sustain msTBI typically present to major trauma services, which provide definitive care for the severely injured [4, 5]. Major trauma services operate within trauma systems, which provide organised and holistic care for the injured, spanning from preventative and prehospital services to rehabilitative care [5]. The development of trauma systems began in the United States in the 1970s after the realisation that preventable deaths were occurring due to suboptimal levels of care [4, 5]. Trauma system implementation has been effective in reducing injury mortality [4, 5, 6], demonstrating a 15%–20% increase in injury survival following trauma centre implementation [6] and up to a 10% decrease in mortality for severely injured patients in mature trauma systems [5]. However, it is difficult to assess whether trauma services have improved msTBI outcomes due to the limited epidemiological data worldwide [1, 7] and a lack of standardised data collection across different hospitals, states and countries [1].
The burden of TBI within Australia, including mild and msTBI, was detailed in the 2020–2021 Australian Institute of Health and Welfare (AIHW) report on head injuries. This reported 142,000 head injury hospitalisations and 2400 deaths between 2020 and 2021, with an average annual increase of 0.4% in deaths since 2011–2012 [8]. An Australian review of msTBI outcomes, as part of the Australian TBI National Data (ATBIND) project, reported a lack of change in msTBI mortality between 2015 and 2020 [9], reflecting the findings of the AIHW report [8].
Possible reasons for the AIHW and ATBIND findings of stagnating TBI mortality outcomes may involve the inclusion of populations with markedly poorer outcomes, including penetrating TBI [10, 11] and those over the age of 70 [12, 13]. Secondly, given that improvements in patient outcomes may lag behind changes in clinical practice, periods extending over decades may be needed to better identify longitudinal improvements [14]. Lastly, different states and territories in Australia have their own prehospital systems, and trauma centres have different management protocols and registry methodologies [15]. Hence, the AIHW and ATBIND project's nationally aggregated findings may not reflect the outcomes of individual trauma centres.
The aim of this study was to determine if in‐hospital mortality for blunt msTBI had changed at a single adult major trauma centre, The Alfred Hospital, over 20 years (2002–2021). The secondary aim was to investigate whether the proportion of patients discharged home following msTBI had changed over the period.
2. Methods
A longitudinal cohort study was conducted to investigate changes in adult blunt msTBI in‐hospital mortality and discharge destination outcomes over 20 years, using data from patients admitted to The Alfred Trauma Service from 1 January 2002 to 31 December 2021. All data were extracted from The Alfred Health Trauma Registry (AHTR).
2.1. Setting
The Alfred Hospital is one of three major trauma centres in Victoria—two adult trauma centres (The Alfred and Royal Melbourne Hospital) and one paediatric trauma centre (Royal Children's Hospital). The Alfred is the largest major trauma service in Australia, accounting for approximately 1/7th of national major trauma presentations [16]. The Victorian State Trauma System (VSTS) was established in 2000, with an integrated system across all trauma‐receiving hospitals (with their own assigned levels or roles, including regional centres), to ensure major trauma patients reached a Level 1 trauma centre as quickly and as safely as possible [17, 18].
2.2. Definitions
Moderate to severe TBI is defined as an Abbreviated Injury Scale (AIS) score for the head region of 3, 4 or 5 [19, 20]. The AIS is an international injury scoring system that is anatomically based, scoring injuries by body region on a scale of 1 to 6 (1 = minor and 6 = maximal/non‐survivable) [21]. It is conducted retrospectively, and studies have shown that the AIS provides a standardised framework with high reliability when used by trained professionals [22]. This definition of TBI severity is chosen over the Glasgow Coma Scale (GCS) as GCS can often be an unreliable indicator of TBI severity and be affected by various confounders, such as sedation and observer variability [23, 24].
The AHTR collects data for all patients admitted under The Alfred Trauma Service, including minor (Injury Severity Score (ISS) ≤ 12) and major trauma patients (ISS > 12), which were both included in this study. Prior to 01/07/2010, the AHTR used AIS1998 coding for injury severity and Injury Severity Score (ISS) calculations, with major trauma cases defined as ISS > 15 [19]. Post 1 July 2010, the AHTR switched to the new AIS2005/2008 coding, under which major trauma was defined as ISS > 12 [19]. For this study, major trauma was defined as ISS > 12 throughout the entire study period for consistency. Patients with an ISS > 24 are defined to be critically injured [19].
The majority of msTBI occur in polytrauma (63.5% of total cases) as opposed to isolated cases (36.5%) [25]. Polytrauma is defined as subjects with injuries of AIS ≥ 3 in at least two different ISS body regions, capturing a cohort with more severe injuries, higher mortality and requiring more resource‐intensive management [26].
The first legitimate GCS was the score that was first recorded before the patient had received any interventions that may have impaired their GCS, including intubation, sedation and paralysing agents. Where possible, the first prehospital GCS was used.
2.3. Inclusion Criteria
Those aged over 70 years or with a penetrating type of msTBI were excluded from this study as a ‘one size fits all’ [12] investigation approach neglects the markedly worse prognoses and specific complexities of these populations [10, 11, 12, 13]. Penetrating msTBI has a much higher mortality due to its complex mechanism and pathophysiology and requires different management than that of blunt msTBI [10, 11]. Patients aged over 70 also have significantly worse outcomes due to concurrent comorbidities, increased anticoagulation usage and increased frailty [12, 13], as well as having different management to that of younger adults [2, 3, 13]. The paediatric population aged under 16 was excluded as The Alfred is an adult major trauma centre.
2.4. Outcomes
The primary outcome was in‐hospital mortality. The secondary outcome was hospital discharge destination—whether the patient was discharged to home directly from The Alfred. Discharge destination was used as a surrogate measure of early functional independence and to support the primary mortality results (i.e., to determine if improvements in mortality occurred at the expense of poorer short‐term disability outcomes, or if early functionality had improved despite a lack of change in mortality). Discharge to home suggested a higher level of initial functional independence, whilst discharge to rehabilitation services indicated poorer short‐term outcomes [27, 28]. Variables with less than 80% completeness for the 20‐year study period were excluded from analyses (Supplementary List 1). The proportion of missing data for each variable is reported.
2.5. Statistical Analysis
Symmetrically distributed continuous data were summarised as means and standard deviations (SDs) and compared with the unpaired t‐test. Non‐symmetrically distributed continuous data and ordinal data were summarised as medians and inter‐quartile ranges (IQRs) and compared with the Mann–Whitney U test. Categorical data were summarised as absolute numbers and percentages (%) and compared with the chi‐squared test. A p‐value less than 0.05 was considered statistically significant.
Unadjusted mortality and discharge destination rates were reported. Logistic regression analyses were used to examine unadjusted mortality and discharge destinations over the 20‐year period. The primary exposure variable for all regression analyses was calendar year (i.e., time, in years, from the commencement of the study period). A multivariable logistic regression analysis was used to examine independent associations between mortality and year, adjusting for all confounders. Confounding variables were variables that had a statistically significant association with both the outcome (death) and exposure (year), respectively, determined through univariable regressions. Outcomes were reported as odds ratios (ORs) with 95% confidence intervals (95% CIs). A p‐value of less than 0.1 was considered significant for all univariable regression analyses and a p‐value of less than 0.05 was considered significant for the multivariable analysis.
Finally, a sensitivity analysis was performed to determine the independent association between death, in hospital and time, in years, adjusted for concrete measures of injury severity. The initial multi‐variable regression was re‐conducted following the removal of variables considered to be modifiable factors (representing processes of care) to determine the real association between mortality and time. Outcomes were reported as ORs with 95% CIs and a p‐value of less than 0.05 was considered significant. Microsoft Excel and Stata 17 were used to analyse and graph data.
2.6. Ethics
This study was approved by The Alfred Hospital Human Research Ethics Committee (Project ID 482/23) and the Monash University Human Research Ethics Committee (Project ID 40078).
3. Results
There were 8222 msTBI presentations to The Alfred Hospital over the 20‐year study period, with 692 deaths (8.5%) (Table 1). Most subjects were male (77.9%), with similar proportions of males in 2002 and 2021. The mean age was 39.4 years (SD 16.1). This increased from 34.9 years (SD 0.8) in 2002 to 42.9 years (SD 0.8) in 2021. In 2002, transport‐related accidents were the most common cause of injury (63.2%), followed by falls (19.9%). By 2021, the proportion of transport‐related causes had decreased (45.2%) and the proportion of falls had increased (36.7%). Most subjects survived to hospital discharge (91.6%), with the proportion of survivors increasing from 87.6% in 2002 to 96.0% in 2021. A greater proportion of survivors were discharged home in 2021 (63.3% vs. 34.8% in 2002), with fewer discharged to rehabilitation in 2021 (28.2% vs. 64.5% in 2002).
TABLE 1.
Characteristics of all included subjects, 2002–2021.
| Variable | Measurement | All patients (n = 8222) | 2002 patients (n = 337) | 2021 patients (n = 403) | p a |
|---|---|---|---|---|---|
| Sex |
Male (%) Female (%) Other (%) |
6407 (77.9%) 1812 (22.0%) 3 (0.04%) |
255 (75.7%) 82 (24.3%) Nil |
322 (79.9%) 81 (20.1%) Nil |
0.17 |
| Age |
Mean SD |
39.4 16.1 |
34.9 0.8 |
42.9 0.8 |
< 0.001 |
| Injury cause |
Transport (%) Falls (%) Struck (%) Other (%) |
4047 (49.2%) 2475 (30.1%) 1352 (16.4%) 348 (4.2%) |
213 (63.2%) 67 (19.9%) 40 (11.9%) 17 (5.0%) |
182 (45.2%) 148 (36.7%) 62 (15.4%) 11 (2.7%) |
< 0.001 |
| AIS (Head) |
Median IQR |
4 3–4 |
4 3–5 |
3 3–4 |
< 0.001 |
| ISS |
Median IQR |
21 16–29 |
22 17–34 |
20 14–27 |
< 0.001 |
| Comorbidities |
Healthy (%) S/I non‐limiting (%) S/I limiting (%) S/I constant threat to life (%) Missing (%) |
3864 (47.0%) 1823 (22.2%) 1229 (14.9%) 267 (3.2%) 1039 (12.6%) |
177 (52.5%) 30 (8.9%) 21 (6.2%) 1 (0.3%) 108 (32.1%) |
184 (45.7%) 125 (31.0%) 72 (17.9%) 7 (1.7%) 15 (3.7%) |
< 0.001 |
| Trauma status |
Major (%) Minor (%) |
6873 (83.6%) 1349 (16.4%) |
291 (86.4%) 46 (13.7%) |
329 (81.6%) 74 (18.4%) |
0.08 |
| Death |
Survived (%) Deceased (%) Missing (%) |
7494 (91.6%) 692 (8.5%) 36 (0.44%) |
290 (86.1%) 40 (11.9%) 7 (2.1%) |
387 (96.0%) 16 (4.0%) 0 (0%) |
< 0.001 |
| Discharge destination b |
Home (%) Rehabilitation (%) Other (%) |
3589 (43.8%) 3422 (41.8%) 483 (5.9%) |
101 (30.6%) 187 (56.7%) 2 (0.6%) |
245 (60.8%) 109 (27.0%) 33 (8.2%) |
< 0.001 |
| Total LOS (days) |
Median IQR |
6 3–14 |
8.5 4–19 |
6 2–15 |
0.01 |
| Arrival pupillary response |
Equal and reactive (%) Fixed dilation of right (%) Fixed dilation of left (%) Bilateral fixed dilation (%) Missing (%) |
7221 (87.8%) 121 (1.5%) 123 (1.5%) 403 (4.9%) 354 (4.3%) |
290 (86.1%) 1 (0.3%) 5 (1.5%) 27 (8.0%) 14 (4.2%) |
349 (86.6%) 11 (2.7%) 9 (2.2%) 20 (5.0%) 14 (3.5%) |
0.02 |
| First legitimate GCS |
Median IQR |
14 15–10 |
13 14–7 |
14 15–12 |
< 0.001 |
Abbreviations: GCS—Glasgow Coma Scale, IQR—interquartile range, LOS—length of stay, SD—standard deviation, S/I—systemic illness.
p‐value < 0.05 was considered statistically significant.
Not including missing cases.
3.1. Unadjusted Univariable Analysis for the Association Between Death and Year
There was a threefold reduction in unadjusted mortality from 12.1% in 2002 to 4.0% in 2021 (Figure 1). The trendline mortality decreased from 11.8% to 5.8% over the 20‐year study period, representing a mean absolute decrease of 0.3% per annum. This translated to a 4.0% decrease in the log odds of death per annum (OR 0.96, 95% CI 0.95–0.97, p < 0.001).
FIGURE 1.
msTBI in‐hospital mortality (unadjusted and trendline mortality rates), 2002–2021. The solid line depicts the unadjusted mortality rate. The dotted line depicts the line of best fit (when death is regressed against year).
3.2. Multivariable Analysis for an Association Between Death and Year
There was a 3.0% decrease in the odds of death per annum after adjusting for all confounders in the multi‐variable logistic regression model (Supplementary Tables 1–3). However, this was not statistically significant (OR 0.97, 95% CI 0.94–1.01, p = 0.13), indicating that year did not have an independent impact on mortality when adjusting for all possible confounders.
3.3. Multivariable Regression With Sensitivity Analysis
After adjusting for indicators of injury severity (Table 2), there was a 4.0% decrease in the odds of death per annum (OR 0.96, 95% CI 0.93–0.99, p = 0.02). This indicated that mortality had decreased over the period after adjusting for changes to the injury severity of presentations.
TABLE 2.
Sensitivity analysis (using multi‐variable logistic regression) for indicators of injury severity.
| Variable | Final co‐variable | Rationale | Final OR | Final p | Final 95% CI |
|---|---|---|---|---|---|
| Year | Primary exposure | 0.96 | 0.02 | 0.93–0.99 | |
| Age | Yes | 1.04 | < 0.001 | 1.03–1.06 | |
|
ISS 25–50 51–75 |
Yes |
5.83 12.33 |
< 0.001 < 0.001 |
3.56–9.53 5.85–25.99 |
|
|
Injury cause Transport |
Yes | 0.53 | 0.001 | 0.36–0.77 | |
|
Comorbidities Systemic illness (limiting) Systemic illness (constant threat) |
Yes |
1.60 3.41 |
0.04 < 0.001 |
1.03–2.49 1.89–6.15 |
|
| Team activation | No | Indicator of processes of care (modifiable co‐variable) | |||
| Triage status | No | Indicator of processes of care (modifiable co‐variable) | |||
| Arrival pulse | No | Removed in backward stepwise regression (p = 0.517) | |||
| Arrival systolic blood pressure | No | Removed in backward stepwise regression (p = 0.159) | |||
| Arrival O2 fractionation | Yes | 1.01 | 0.002 | 1.01–1.02 | |
|
Arrival pupillary response Fixed dilation (R) Fixed dilation (L) Fixed dilation (bilateral) |
Yes |
2.96 2.61 18.19 |
0.003 0.01 < 0.001 |
1.45–6.05 1.21–5.64 11.04–29.96 |
|
|
First Legitimate GCS 7–9 5–6 4 3 |
Yes |
2.61 6.23 4.19 14.26 |
0.002 < 0.001 < 0.001 < 0.001 |
1.41–4.85 3.23–12.02 1.87–9.37 8.55–23.78 |
|
| Arrival intubation | No | Indicator of processes of care (modifiable co‐variable) | |||
| ED departure destination | No | Indicator of processes of care (modifiable co‐variable) | |||
|
Prehospital CPR Yes |
Yes | 5.94 | < 0.001 | 2.69–13.11 | |
|
Prehospital paralysing agent Yes |
Yes | 0.65 | 0.05 | 0.42–0.99 | |
| ICU length of stay (hours) | No | Indicator of processes of care (modifiable co‐variable) | |||
| ICU ventilation hours | No | Indicator of processes of care (modifiable co‐variable) |
Note: All boxes highlighted in red depict the variables that were considered indicators of processes of care (i.e., the modifiable co‐variables). All boxes highlighted in yellow depict variables that were removed during the forward and backward stepwise regressions due to having a p‐value greater than 0.1. The variables in the green boxes were kept in the final sensitivity analysis model (p < 0.05).
Abbreviations: CPR—cardiopulmonary resuscitation, ED—emergency department, GCS—Glasgow Coma Scale, ICU—intensive care unit, ISS—Injury Severity Score, L—left, LOS—length of stay, O2—Oxygen, R—right.
3.4. Subgroup Mortality
Figure 2 shows changes in mortality for subgroups from 2002 to 2021 (year‐adjusted odds of death and mortality rates in Supplementary Table 4). Changes in the unadjusted mortality rates between 2002 and 2021 were statistically significant across all subgroups (p < 0.001 for all subgroups).
FIGURE 2.
msTBI mortality by subgroups, 2002–2021: (a) Sex; (b) Age Groups; (c) Head Abbreviated Injury Scale (AIS); (d) Injury Severity Score (ISS); and (e) isolated msTBI versus polytrauma msTBI. Note: Graphs are not scaled between each other, and the y‐axis must be noted for the appropriate mortality rate when comparing between graphs. Please refer to the Supplementary Materials for the detailed percentage figures, odds ratios, and p‐values.
3.4.1. Sex
There was no statistically significant difference in the odds of death for females (OR 1.0, p = 0.663) compared to males. There was no difference in their mean absolute reduction in mortality (0.3% p.a. for both).
3.4.2. Age
The odds of death increased with age but were only statistically significant for the two oldest age groups—ages 50–59 (OR 1.4, p = 0.046) and 60–70 (OR 1.9, p < 0.001). Younger age groups had the same mean absolute decrease in mortality of 0.3% p.a., but the two oldest age groups had a higher mean reduction of 0.4% p.a. (ages 50–59) and 0.5% p.a. (ages 60–70).
3.4.3. Head AIS
Compared to the reference group of head AIS of 3, subjects with head AIS of 4 had slightly increased odds of death (OR 2.5, p < 0.001) and those with head AIS of 5 had significantly increased odds of death (OR 25.9, p < 0.001). Those with head AIS of 5 had a much greater mean absolute decrease in mortality of 0.6% p.a. over the 20 years, compared with head AIS of 3 and 4 (0.05% p.a. and 0.1% p.a. reduction, respectively).
3.4.4. Severity of Injuries
Compared to the reference group of ISS ≤ 12, subjects with ISS 25–50 had greater odds of death (OR 43.7, p < 0.001) and those with the highest ISS > 50 had the greatest odds of death (OR 191.2, p < 0.001). Those with ISS > 50 had a higher mean absolute decrease in mortality of 0.3% p.a., followed by ISS 25–50 group with 0.2% p.a. decrease. The non‐critically injured groups with ISS ≤ 24, had a much lower annual absolute decrease of 0.01% p.a.
3.4.5. Isolated msTBI Versus Polytrauma Including msTBI
Compared to those with isolated msTBI, subjects presenting with msTBI as a component of polytrauma had increased odds of death (OR 1.7, p < 0.001). Both isolated msTBI subjects and polytrauma subjects had a similar mean absolute decrease in mortality (0.2% p.a. isolated TBI, 0.3% p.a. polytrauma subjects).
3.5. Discharge Destinations
The proportion of patients discharged to home increased twofold from 30.6% in 2002 to 60.8% in 2021, whilst the proportion discharged to rehabilitation decreased twofold from 56.7% in 2002 to 27.0% in 2021 (Figure 3). The proportion discharged to other destinations, including nursing homes, special accommodation and transfers to other hospitals, increased from 0.6% to 8.2% over the period. The dip in the proportion discharged home and the corresponding spike in the proportion discharged to other destinations from 2012 to 2014 was likely due to incomplete data collection, as the default categorisation is ‘other’ when discharge data was incomplete. The trendline proportion discharged to home increased by a relative 76%, from 31.4% in 2002 to 55.3% in 2021, and the trendline proportion discharged to rehabilitation decreased by a relative 45%, from 53.9% in 2002 to 29.9% in 2021 (Figure 3).
FIGURE 3.
msTBI discharge destinations, 2002–2021. The solid line depicts the unadjusted discharge proportion rates. The dotted line depicts the line of best fit (when discharge destination is regressed against year).
There was a 5.0% increase in the unadjusted odds of discharge to home (OR 1.05, 95% CI 1.04–1.06, p < 0.001; Supplementary Figure 1) as compared to discharge to all other destinations (including in‐hospital death) over the study period.
4. Discussion
This study investigated in‐hospital mortality and discharge destination of blunt moderate to severe TBI for adults presenting to The Alfred Hospital over a 20‐year period (2002–2021). During the first two decades of the century, unadjusted mortality reduced by threefold to 4%, with the greatest declines seen in patients with the most severe head injuries and those over the age of 50. The odds of death per annum reduced by 4% per annum after adjusting for changes to the injury severity of presentations over the period. A surrogate marker of functional independence at discharge also improved, with a two‐fold increase in the proportion of subjects discharged to home over the 20 years.
4.1. Mortality
The findings of this study vary from recent publications related to msTBI mortality associated with aggregate Australian data [8, 9]. O'Reilly et al. had expressed concerns regarding a lack of statistically significant change in msTBI mortality from 2015 to 2020 [9]. The shorter study period, the inclusion of subgroups with higher fatality rates, including penetrating type TBIs and elderly populations, as well as the use of multicentre aggregate data in O'Reilly et al.'s study may explain the differences in findings.
The improved outcomes may be explained by the maturation of the Victorian State Trauma System [29, 30, 31] and improved neurotrauma management at The Alfred Hospital. Subjects with the most severe head injuries, as well as both isolated msTBI and polytrauma groups, had seen a reduction in mortality. This may reflect that general trauma, resuscitation, neurosurgical and intensive care advances have been effective. Changes that were implemented during this time included the inclusion of blood products on Air Ambulance Victoria and more aggressive early resuscitation on patient arrival. However, it was outside the scope of this current study to determine causal explanations, and further investigations should examine which interventions and changes were most efficacious. Future studies should also explore mortality trends for the excluded penetrating, paediatric and elderly populations.
4.2. Discharge Destinations
Moderate to severe TBI is not only a major burden of death but a leading cause of disability [1, 32]. Improvements in msTBI mortality through intensive treatments should not occur at the expense of severe disability. Reassuringly, the results suggest that improvements to msTBI mortality did not result in worsened disability outcomes, with fewer patients requiring additional care in specialised services. The increase in the proportion discharged to home and other destinations may be, in part, attributed to earlier allied health input in the Intensive Care Unit (ICU) and on the wards, as well as the introduction of the National Disability Insurance Scheme (NDIS) in 2011.
Many studies emphasise the importance of examining disability outcomes after TBI [29, 31, 33], but there are currently few studies that explore trends in short‐term functionality. Discharge destination was used as a surrogate measure in this study. Future studies should explore changes to disability outcomes in detail. The collection of data on different domains of cognitive and motor functional outcomes prior to discharge, such as in Sanchez et al.'s study [31], may be a consideration for the AHTR and other trauma registries.
4.3. Strengths and Limitations
To the best of our knowledge, this is the first Australian comprehensive single‐centre study to identify trends in msTBI mortality and functional outcomes over two decades. A major strength was the large cohort and dataset used in the study. However, as a single‐centre study, these results may lack applicability to other Australian major trauma centres.
As an observational study, only postulations can be provided to explain changes in msTBI in‐hospital mortality, and there is a further need to determine causation. Other limitations include an inability to distinguish whether deaths had occurred directly from msTBI or as a combination with other factors such as haemorrhagic shock.
Discharge destination is not only influenced by neurological recovery but also by changes to healthcare and discharge practices over time. The introduction of the NDIS in 2011 and increasing emphasis on at‐home care and community‐based care may have, in part, contributed to the increased likelihood of discharge to home and other destinations during this period. Hence, further studies should consider means to examine longer‐term disability outcomes directly.
There are also data‐related limitations. The AHTR injury severity coding changed from AIS1998 to AIS2005/2008 from July 2010 onwards to reflect more recent terminologies of injury severity as per the Association for Advancement of Automotive Medicine updates. Whilst this change should not affect the study's primary mortality findings, current evidence suggests the incidence of msTBI cases captured by each score from AIS ≥ 3 may have decreased [34, 35]. This should be taken into consideration when viewing msTBI incidence results.
5. Conclusion
In this single Australian Level 1 trauma centre study at The Alfred Hospital, in‐hospital mortality in 16–70 year old adults with blunt moderate to severe traumatic brain injury reduced significantly during the first two decades of the century (2002–2021). The greatest improvements were seen in those with the most severe head injuries and the older sub‐population 50–70 of age. The proportion discharged directly to home after msTBI also increased.
Whilst some uncertainty has been expressed regarding the impact of current systems of care on msTBI outcomes, these findings foster optimism and support further investigations into neurotrauma care.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supplementary Table 1. Univariable regressions, testing all variables with the outcome variable (death).
Supplementary Table 2. Univariable regressions, testing association between the exposure variables and the primary exposure variable (year).
Supplementary Table 3. Multi‐variable regression analysis, adjusting for all possible confounding variables.
Supplementary Table 4. Year‐adjusted odds ratios of death and mortality rates for subgroups (unadjusted and trendline results), 2002–2021.
Supplementary Figure 1. logistic regression of discharge to home against year, 2002–2021.
Acknowledgements
We would like to acknowledge Ms. Jane Ford (Alfred Health Trauma Registry Manager) and Ms. Zoe Cheung (Senior Trauma Registry Information Management Officer) for extracting and collating the initial data from the Alfred Health Trauma Registry. Open access publishing facilitated by Monash University, as part of the Wiley ‐ Monash University agreement via the Council of Australasian University Librarians.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Maas A. I. R. and Menon D. K., “Traumatic Brain Injury: Integrated Approaches to Improve Prevention, Clinical Care, and Research,” Lancet Neurology 16, no. 12 (2017): 987–1048. [DOI] [PubMed] [Google Scholar]
- 2. Dewan M. C., Rattani A., Gupta S., et al., “Estimating the Global Incidence of Traumatic Brain Injury,” Journal of Neurosurgery 130 (2018): 1–18. [DOI] [PubMed] [Google Scholar]
- 3. Andriessen T. M. J. C., Horn J., and Franschman G., “Epidemiology, Severity Classification, and Outcome of Moderate and Severe Traumatic Brain Injury: A Prospective Multicenter Study,” Journal of Neurotrauma 28, no. 10 (2011): 2019–2031. [DOI] [PubMed] [Google Scholar]
- 4. West J. G., Trunkey D. D., and Lim R. C., “Systems of Trauma Care. A Study of Two Counties,” Archives of Surgery 114, no. 4 (1979): 455–460. [DOI] [PubMed] [Google Scholar]
- 5. Gruen R. L., Gabbe B. J., Stelfox H. T., and Cameron P. A., “Indicators of the Quality of Trauma Care and the Performance of Trauma Systems,” British Journal of Surgery 99 (2011): 97–104. [DOI] [PubMed] [Google Scholar]
- 6. Mullins R. J. and Mann N. C., “Population‐Based Research Assessing the Effectiveness of Trauma Systems,” Journal of Trauma 47, no. 3 (1999): S59–S66. [DOI] [PubMed] [Google Scholar]
- 7. James S. L., Theadom A., Ellenbogen R. G., et al., “Global, Regional, and National Burden of Traumatic Brain Injury and Spinal Cord Injury, 1990–2016: A Systematic Analysis for the Global Burden of Disease Study 2016,” Lancet Neurology 18, no. 1 (2019): 56–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Australian Institute of Health and Welfare , Head Injuries in Australia 2020–2021 (AIHW, 2023), https://www.aihw.gov.au/reports/injury/head‐injuries‐in‐australia‐2020‐21. [Google Scholar]
- 9. O'Reilly G. M., Curtis K., Mitra B., et al., “Hospitalisations and in‐Hospital Deaths Following Moderate to Severe Traumatic Brain Injury in Australia, 2015–2020: A Registry Data Analysis for the Australian Traumatic Brain Injury National Data (ATBIND) Project,” Medical Journal of Australia 219 (2023): 316–324. [DOI] [PubMed] [Google Scholar]
- 10. Kazim S. F., Shamim M. S., Tahir M. Z., Enam S. A., and Waheed S., “Management of Penetrating Brain Injury,” Journal of Emergencies, Trauma, and Shock 4, no. 3 (2011): 395–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Santiago L. A., Oh B. C., Dash P. K., Holcomb J. B., and Wade C. E., “A Clinical Comparison of Penetrating and Blunt Traumatic Brain Injuries,” Brain Injury 26, no. 2 (2012): 107–125. [DOI] [PubMed] [Google Scholar]
- 12. Thompson H. J., McCormick W. C., and Kagan S. H., “Traumatic Brain Injury in Older Adults: Epidemiology, Outcomes, and Future Implications,” Journal of the American Geriatrics Society 54, no. 10 (2006): 1590–1595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Gardner R. C., Dams‐O'Connor K., Morrissey M. R., and Manley G. T., “Geriatric Traumatic Brain Injury: Epidemiology, Outcomes, Knowledge Gaps, and Future Directions,” Journal of Neurotrauma 35, no. 7 (2018): 889–906. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Morris Z. S., Wooding S., and Grant J., “The Answer Is 17 Years, What Is the Question: Understanding Time Lags in Translational Research,” Journal of the Royal Society of Medicine 104, no. 12 (2011): 510–520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Warren K.‐R. J., Morrey C., Oppy A., Pirpiris M., and Balogh Z. J., “The Overview of the Australian Trauma System,” OTA International: The Open Access Journal of Orthopaedic Trauma 2, no. S1 (2019): e018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Australia New Zealand Trauma Registry , Annual Report 1 July 2020 to 20 June 2021. Alfred Health (Australian New Zealand Trauma Registry, 2021), https://static1.squarespace.com/static/5b761ed3f93fd491065f7839/t/639a85aa5aac7405c9cd89ee/1671071179741/ATR_Annual‐Report_20‐21.pdf. [Google Scholar]
- 17. Atkin C., Freedman I., Rosenfeld J. V., Fitzgerald M., and Kossmann T., “The Evolution of an Integrated State Trauma System in Victoria, Australia,” Injury 36, no. 11 (2005): 1277–1287. [DOI] [PubMed] [Google Scholar]
- 18. Cameron P. A., Gabbe B. J., Cooper D. J., Walker T., Judson R., and McNeil J., “A Statewide System of Trauma Care in Victoria: Effect on Patient Survival,” Medical Journal of Australia 189, no. 10 (2008): 546–550. [DOI] [PubMed] [Google Scholar]
- 19. Palmer C. S., Gabbe B. J., and Cameron P. A., “Defining Major Trauma Using the 2008 Abbreviated Injury Scale,” Injury 47, no. 1 (2016): 109–115. [DOI] [PubMed] [Google Scholar]
- 20. O'Reilly G. M., Curtis K., Kim Y., et al., “The Australian Traumatic Brain Injury National Data (ATBIND) Project: A Mixed Methods Study Protocol,” Medical Journal of Australia 217, no. 7 (2022): 361–365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Medicine AftAoA , About AIS (AAAM, 2026), https://www.aaam.org/abbreviated‐injury‐scale‐ais/about‐ais/. [Google Scholar]
- 22. Likkachai K., Wongwaisayawan S., Siriwes K., and Worasuwannarak W., “Abbreviated Injury Scale‐Guided Assessment of Traumatic Deaths: Postmortem CT Versus Autopsy,” Forensic Science International: Synergy 10 (2025): 100588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Udekwu P., Kromhout‐Schiro S., Vaslef S., Baker C., and Oller D., “Glasgow Coma Scale Score, Mortality, and Functional Outcome in Head‐Injured Patients,” Journal of Trauma 56, no. 5 (2004): 1084–1089. [DOI] [PubMed] [Google Scholar]
- 24. Fitzgerald M., Tan T., and Rosenfeld J. V., “An Inital Glasgow Coma Scale Score of 8 or Less Does Not Define Severe Brain Injury,” Emergency Medicine Australasia 34, no. 3 (2022): 307–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. O'Reilly G. M., Afroz A., Curtis K., et al., “The Determinants for Death in Hospital Following Moderate to Severe Traumatic Brain Injury in Australia,” Emergency Medicine Australasia 37, no. 1 (2025): e14562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Butcher N. E., D'Este C., and Balogh Z. J., “The Quest for a Universal Definition of Polytrauma: A Trauma Registry‐Based Validation Study,” Journal of Trauma and Acute Care Surgery 77, no. 4 (2014): 620–623. [DOI] [PubMed] [Google Scholar]
- 27. Myszenski A., Zhou Y., Abbas F. T., and Siddiqui A., “The Predictive Validity of Functional Outcome Measures With Discharge Destination for Hospitalized Medical Patients,” Archives of Rehabilitation Research and Clinical Translation 4, no. 4 (2022): 100231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Thorpe E. R., Garrett K. B., Smith A. M., Reneker J. C., and Phillips R. S., “Outcome Measure Scores Predict Discharge Destination in Patients With Acute and Subacute Stroke: A Systematic Review and Series of Meta‐Analyses,” Journal of Neurologic Physical Therapy 42, no. 1 (2018): 2–11. [DOI] [PubMed] [Google Scholar]
- 29. Sawadogo D., Moore L., Tardif P. A., Farhat I., Lauzier F., and Turgeon A. F., “Trends of Clinical Outcomes in Patients With a Traumatic Brain Injury (TBI) in Canada Between 2006 and 2012,” Injury 51, no. 1 (2020): 76–83. [DOI] [PubMed] [Google Scholar]
- 30. Graves W. C., Oyesanya T. O., Gormley M., et al., “Pre‐ and In‐Hospital Mortality for Moderate‐To‐Severe Traumatic Brain Injuries: An Analysis of the National Trauma Data Bank (2008–2014),” Brain Injury 35, no. 3 (2021): 265–274. [DOI] [PubMed] [Google Scholar]
- 31. Sanchez A. I., Krafty R. T., Weiss H. B., Rubiano A. M., Peitzman A. B., and Puyana J. C., “Trends in Survival and Early Functional Outcomes From Hospitalized Severe Adult Traumatic Brain Injuries, Pennsylvania, 1998 to 2007,” Journal of Head Trauma Rehabilitation 27, no. 2 (2012): 159–169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Maas A. I. R., Menon D. K., Manley G. T., et al., “Traumatic Brain Injury: Progress and Challenges in Prevention, Clinical Care, and Research,” Lancet Neurology 21, no. 11 (2022): 1004–1060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Fuller G., Bouamra O., Woodford M., et al., “Temporal Trends in Head Injury Outcomes From 2003 to 2009 in England and Wales,” British Journal of Neurosurgery 25, no. 3 (2011): 414–421. [DOI] [PubMed] [Google Scholar]
- 34. Hsu S.‐Y., Wu S.‐C., Rau C.‐S., et al., “Impact of Adapting the Abbreviated Injury Scale (AIS)‐2005 From AIS‐1998 on Injury Severity Scores and Clinical Outcome,” International Journal of Environmental Research and Public Health 16, no. 24 (2019): 5033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Salottolo K., Settell A., Uribe P., et al., “The Impact of the AIS 2005 Revision on Injury Severity Scores and Clinical Outcome Measures,” Injury 40, no. 9 (2009): 999–1003. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 1. Univariable regressions, testing all variables with the outcome variable (death).
Supplementary Table 2. Univariable regressions, testing association between the exposure variables and the primary exposure variable (year).
Supplementary Table 3. Multi‐variable regression analysis, adjusting for all possible confounding variables.
Supplementary Table 4. Year‐adjusted odds ratios of death and mortality rates for subgroups (unadjusted and trendline results), 2002–2021.
Supplementary Figure 1. logistic regression of discharge to home against year, 2002–2021.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.