Key results

This study, examining 10,587 patients for a time period of nearly a decade, is to our knowledge the largest study in Sweden reporting trends in trauma. During the study period, there was an increase in mortality in trauma patients with a TA and NISS < 15, no change in the rate of emergency interventions and a decline in ICU admissions in all groups except TA and NISS < 15. Penetrating trauma increased significantly from 12.4 to 19.6%, and there was an increase in assaults (both penetrating and blunt) from 14.7 to 21.4%.

InterpretationMortality

Overall mortality in the group of NISS > 15 was 13.0%– less than in studies from Denmark [12], the USA [13] and Japan [14] but comparable to a previous Swedish study [3]. The mortality also compares very well when looking at SweTrau’s yearly reports, were the mortality in the group NISS > 15 between the years 2013–2021 varies from 14 to 20%, with a mean of around 17%. However, the mortality in SweTrau reflects a mixture of university hospitals and regional hospitals, in contrast to our study’s two major university trauma centers, which could possibly explain the higher mortality rate in SweTrau. Further, our results are similar to an Australian study that showed no significant changes in mortality during the years 2006–2016 [15], with one worrisome exception in our study; the group with a TA and NISS < 15 where mortality has risen from 1.3 to 2.7%. One could hypothesize that the implementation of the SNTTC in 2017 could have led to a more stringent use of the highest trauma call (TA), possibly resulting in a change of presentation of injured patients with a TA and NISS < 15 to ED: s and thus explaining part of this observed increase in mortality among NISS < 15 patients. However, in a Swedish study from 2019 [16], it was shown that both the number of TA: s and the proportion of patients with ISS < 15 among TA: s were the same before and after the implementation of the new criteria, which do not support this idea. The increase in mortality for patients with a TA and NISS < 15 is visible from 2019 and onward and an important proportion was related to elderly patients with blunt trauma and a low NISS that died in the ED. The low NISS (1–3) in these patients is surprising since it does not necessarily explain a traumatic cardiac arrest from blunt trauma, but could rather be due to e.g. rib fractures caused by cardiopulmonary resuscitation. Taking all this into account, including patient age, one must consider the possibility of these patients having been mis-triaged as trauma patients whilst instead suffering a non-traumatic cardiac arrest, subsequently causing the patient’s motor vehicle crash or fall. A factor that somewhat contradicts this, however, is the quality improvement variable “morbidity & mortality conferences” (M&M) in SweTrau. All these patients have been discussed at M&M:s to determine if the patient is a trauma patient or not, and normally cases that are not primary trauma would have been excluded from the registry after such discussion. Moreover, the subanalysis of the mortality in TA and NISS < 15 displayed a significant increase in blunt trauma only. However, there were only nine deaths from penetrating trauma (of which seven occurred 2019–2021), why there may be a risk of a type II-error. A larger population might have shown a mortality increase in penetrating trauma as well. On closer examination of the TA and NISS < 15 group, there was a striking difference in the number of deaths from knife injuries, GSW and suicides from blunt trauma during the study period, resulting in twelve fatalities during the last three years (2019–2021) compared to only three during the first six years (2013–2018). In summary, deaths from penetrating trauma and suicides, together with the blunt trauma patients who died in the ED, constituted more than half of the deceased patients with a TA and NISS < 15 during 2019–2021, coinciding with the increase in mortality in this group. The change in the trauma panorama in Sweden, with a clear increase of both penetrating and blunt assaults, thus can have contributed to this outcome.

Emergency intervention & intensive care unit admissions

Emergency interventions did not change during the study period, underlined by the fact that laparotomy and thoracotomy in penetrating trauma showed no significant trends. The overall decrease in ICU admissions is however of interest, with an even sharper decline during the COVID-19 pandemic years of 2020–2021. This does not correspond to findings from other international studies [17, 18], where the ICU proportions of patients requiring ICU admission was unchanged during COVID-19. Furthermore, the number of trauma patients in our study actually increased during the years 2020–2021, unlike many other countries that practiced lockdown [17,18,19,20], and the increase in severe trauma cannot be explained in its entirety by a rise in population. This could be interpreted as a very concerning finding, indicating that trauma patients in Sweden were down-prioritized from the ICU during COVID-19. Nonetheless, as earlier discussed there was no increase in mortality during the study period except for the group with a TA and NISS < 15 and, on closer examination, there was no decrease in the proportions of ICU admissions in deceased patients in this group. This indicates that even if fewer trauma patients were indeed admitted to the ICU, patient safety was still maintained. The current data therefore suggests that a larger proportion of severe trauma could be managed at a somewhat lower care level with the same result.

Mechanisms of injury

During the study period, we found that “being hit by a blunt object” increased as a mechanism of injury, together with other assaults (knife-inflicted trauma and GSW)– in contrast to the findings of a study from the USA between 2005 and 2014 [21] but consistent with a Danish study between 2010 and 2019 [12]. We also found that traffic injuries such as MCC and pedestrian accidents decreased, in line with other studies [12, 14, 21]. These changes could be interpreted as a shift towards violence being a more prevalent cause of injury while traffic related trauma decreases, indicating areas to focus on when discussing trauma prevention. Of note, we could not see a significant change in fall injuries, contrary to other studies that have reported a trend of higher incidence of falls [12, 14, 21, 22]. One possible explanation to this could be that this study only examines severe trauma, not the whole trauma population.

Penetrating trauma

The proportion of penetrating trauma increased with more than 50% during the study period (12.4–19.6%, p < 0.001), but the mortality remained at just below 6% with some variations over the years. The increase in penetrating trauma in Sweden is supported by other studies [23, 24], and one has to remember that patients that are pronounced dead at scene are not included in SweTrau. This can lead to false low mortality numbers, as showed in a Swedish study [23] where death due to violence was increased but not in-hospital mortality for penetrating trauma.

Strengths, limitations and generalisability

The major strengths of our study are the substantial size of the population and the extensive time period examined, as well as the inclusion of two out of in total seven university hospitals in Sweden. Using two different statistical methods (JoinPoint regression [25] and Chi-Squared test for trend) also strengthens the validity of the results. JoinPoint is a common way to describe trends in incidence [21, 26] and includes the advantage of a measurement (AAPC) that is easy to comprehend and compare. The additional analysis of estimated incidence of severe trauma in the primary catchment area displayed no significant trend during the study, reinforcing the validity of the percentage rates for the different outcomes. The retrospective nature of this paper carries the usual limitations of a registry based study, which makes more in-depth analyses of individual cases difficult. Also, any international comparisons of the results should be done with the Swedish context of our study in mind. A possible bias is the excluded patients; however, these were scarce (about 2%) and the potential impact should therefore be minimal. Finally, since patients that die on the scene of trauma are not included in SweTrau, the mortality rate needs to be interpreted considering this, and our study therefore only reflects the in-hospital mortality.

Implications and future perspectives

This article has highlighted a very worrisome increase in penetrating trauma and assaults, concerning not only the healthcare sector but the society as a whole. The fact that on-scene mortality is not recorded only underlines the need of further studies combining these parameters as well as mapping other relevant epidemiological aspects. The study finding of lower rate of ICU utilization during the study period, with maintained low mortality in this group of trauma patients, raises the possibility that a lower care level may be appropriate in selected trauma cases, which could improve resource utilisation and access to ICU. Further studies comparing patients treated at a dedicated trauma ward or intermediate ward with patients treated at ICU are needed.