Higher hospital level does not improve 30-day survival after road traffic accidents
(2025) In Scientific Reports 15(1).- Abstract
Globally, road traffic accidents (RTAs) remain a major cause of death, particularly among individuals aged 15–30 years. While Sweden has been at the forefront of traffic safety through the Vision Zero initiative, in-hospital management remains crucial in determining RTA outcomes. Drawing on North American evidence suggesting improved survival at trauma centres, the Swedish healthcare system has increasingly emphasised trauma centralisation. However, comprehensive national data from Sweden are scarce. Given the country’s unique demographic and geographic characteristics, including vast sparsely populated areas, direct comparisons with other Western systems are challenging. We analysed the epidemiology and risk factors for 30-day... (More)
Globally, road traffic accidents (RTAs) remain a major cause of death, particularly among individuals aged 15–30 years. While Sweden has been at the forefront of traffic safety through the Vision Zero initiative, in-hospital management remains crucial in determining RTA outcomes. Drawing on North American evidence suggesting improved survival at trauma centres, the Swedish healthcare system has increasingly emphasised trauma centralisation. However, comprehensive national data from Sweden are scarce. Given the country’s unique demographic and geographic characteristics, including vast sparsely populated areas, direct comparisons with other Western systems are challenging. We analysed the epidemiology and risk factors for 30-day mortality among 95,954 RTA-related hospital admissions in Sweden between 2008 and 2021. Predictors included the ICD-based Injury Severity Score (ICISS), age, sex, Charlson Comorbidity Index (CCI), year of event, and hospital level. Mortality risk was modelled using explainable artificial intelligence (XAI) via Extreme Gradient Boosting (XGBoost) with SHapley Additive exPlanations (SHAP), alongside conventional multivariable logistic regression for comparison. The most influential predictors of 30-day mortality, in descending order, were ICISS, age, CCI, event year, hospital level, and sex. A clear trend toward centralisation was observed, with Level 1 hospitals admitting the most severely injured patients. However, after risk adjustment, the hospital level was not independently associated with 30-day mortality. The XAI model outperformed logistic regression in both discrimination and calibration, confirming these findings. This study represents a comprehensive national analysis of in-hospital outcomes following RTAs in Europe. ICISS, age, sex, and comorbidity influenced mortality risk, while overall survival improved over time. The assumption that trauma centralisation confers a universal survival advantage does not appear to hold in the Swedish context. These findings underscore the need to re-evaluate trauma system design under Scandinavian conditions—ensuring that timely access to hospital care is not compromised by centralisation.
(Less)
- author
- Ydenius, Viktor ; Djerf, Sebastian LU ; Fredrikson, Mats ; Larsen, Robert ; Sjöberg, Folke and Frigyesi, Attila LU
- organization
- publishing date
- 2025-12
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Scientific Reports
- volume
- 15
- issue
- 1
- article number
- 41164
- publisher
- Nature Publishing Group
- external identifiers
-
- scopus:105022611036
- pmid:41272162
- ISSN
- 2045-2322
- DOI
- 10.1038/s41598-025-26519-7
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © The Author(s) 2025.
- id
- eebf9c01-9add-4b45-b85b-9a32faf01ff4
- date added to LUP
- 2026-01-16 14:17:17
- date last changed
- 2026-02-27 18:27:02
@article{eebf9c01-9add-4b45-b85b-9a32faf01ff4,
abstract = {{<p>Globally, road traffic accidents (RTAs) remain a major cause of death, particularly among individuals aged 15–30 years. While Sweden has been at the forefront of traffic safety through the Vision Zero initiative, in-hospital management remains crucial in determining RTA outcomes. Drawing on North American evidence suggesting improved survival at trauma centres, the Swedish healthcare system has increasingly emphasised trauma centralisation. However, comprehensive national data from Sweden are scarce. Given the country’s unique demographic and geographic characteristics, including vast sparsely populated areas, direct comparisons with other Western systems are challenging. We analysed the epidemiology and risk factors for 30-day mortality among 95,954 RTA-related hospital admissions in Sweden between 2008 and 2021. Predictors included the ICD-based Injury Severity Score (ICISS), age, sex, Charlson Comorbidity Index (CCI), year of event, and hospital level. Mortality risk was modelled using explainable artificial intelligence (XAI) via Extreme Gradient Boosting (XGBoost) with SHapley Additive exPlanations (SHAP), alongside conventional multivariable logistic regression for comparison. The most influential predictors of 30-day mortality, in descending order, were ICISS, age, CCI, event year, hospital level, and sex. A clear trend toward centralisation was observed, with Level 1 hospitals admitting the most severely injured patients. However, after risk adjustment, the hospital level was not independently associated with 30-day mortality. The XAI model outperformed logistic regression in both discrimination and calibration, confirming these findings. This study represents a comprehensive national analysis of in-hospital outcomes following RTAs in Europe. ICISS, age, sex, and comorbidity influenced mortality risk, while overall survival improved over time. The assumption that trauma centralisation confers a universal survival advantage does not appear to hold in the Swedish context. These findings underscore the need to re-evaluate trauma system design under Scandinavian conditions—ensuring that timely access to hospital care is not compromised by centralisation.</p>}},
author = {{Ydenius, Viktor and Djerf, Sebastian and Fredrikson, Mats and Larsen, Robert and Sjöberg, Folke and Frigyesi, Attila}},
issn = {{2045-2322}},
language = {{eng}},
number = {{1}},
publisher = {{Nature Publishing Group}},
series = {{Scientific Reports}},
title = {{Higher hospital level does not improve 30-day survival after road traffic accidents}},
url = {{http://dx.doi.org/10.1038/s41598-025-26519-7}},
doi = {{10.1038/s41598-025-26519-7}},
volume = {{15}},
year = {{2025}},
}