Introduction
Trauma stands as a global health crisis, responsible for over 5 million deaths annually and impacting nearly a billion individuals worldwide who require medical intervention for injuries.1 2 To put this into perspective, the global burden of injury surpasses the combined fatalities of malaria, tuberculosis, and HIV/AIDS by a staggering 32%.2 For those between the ages of 5 and 45, trauma is the primary driver of disability-adjusted life-years, highlighting its devastating impact on productive life spans.3 The economic repercussions of premature mortality and long-term disability resulting from injuries are substantial.4 5 Road traffic accidents alone, which constitute less than a third of global injuries, are estimated to cost nations up to 2% of their gross national product.5
A stark reality is that nearly 90% of injury-related deaths occur in low and middle-income countries (LMICs). If LMICs could achieve fatality rates among the injured comparable to those in high-income countries (HICs), approximately 2 million lives could be saved each year.5 6 India, in particular, accounts for over 20% of the world’s trauma deaths,7 where injuries have been recognized as a significant public health challenge.8–11 Alarmingly, a Delphi study focusing on injury-related deaths in India revealed that over half of these deaths were considered preventable.12
Robust trauma research and monitoring are integral components of advanced trauma systems prevalent in HICs.13–16 Despite the overwhelming burden of injury in LMICs, the majority of trauma care research originates from HICs.17 18 This scarcity of injury-related data and research in LMICs poses a significant obstacle to the development of global emergency and trauma care systems.17–19 To bridge this critical gap, the Towards Improved Trauma Care Outcomes (TITCO) data project was initiated in India, aiming to enhance trauma care information systems and systematically gather essential injury data.20
Comparing risk-adjusted trauma mortality rates between HICs and LMICs is crucial for pinpointing specific patient demographics and injury patterns that necessitate targeted interventions. However, the limited availability of detailed patient and injury data from LMICs hinders such comparative analyses. This study endeavors to address this knowledge deficit by identifying independent predictors of trauma mortality and conducting a thorough analysis of differences in demographics, physiology, injury burden, and ultimately, injury mortality between India and the USA. Such a comparative study can illuminate specific areas for improvement in trauma care and guide strategies to reduce trauma-related mortality in India and similar resource-constrained environments.
Methods
This research constitutes a retrospective cohort study comparing injured patients admitted to university hospitals in India (representing an LMIC) and the USA (representing an HIC). Data for this comparison were drawn from the US National Trauma Data Bank (NTDB) and India’s TITCO database.
The Indian data were compiled and overseen by a research consortium encompassing university hospitals across four major metropolitan regions: Apex Trauma Centre of the All-India Institute of Medical Sciences, New Delhi (north central India); Lokmanya Tilak Municipal General Hospital, and King Edward Memorial Hospital, Mumbai (western India); Seth Sukhlal Karnani Memorial Hospital, Kolkata (eastern India); and Rajiv Gandhi Hospital, Chennai (south India). While the Apex Trauma Centre operates as a dedicated trauma center, the other sites are trauma units within university-affiliated teaching hospitals. These facilities function as tertiary care centers, providing services to the public, largely free of charge or with minimal user fees, thereby ensuring access to care for individuals across lower socioeconomic strata. In India, data collection was carried out by project officers through meticulous record reviews and direct observation within the trauma patient reception areas. To ensure comprehensive data capture, these officers were authorized to consult healthcare staff for parameter values not initially documented in patient records.
Data collection periods spanned from January 2013 to December 2015 for the NTDB data, and from July 2013 to December 2015 for the TITCO data. The study included patients aged 18 years and older who sustained blunt or penetrating trauma, were alive upon arrival at the emergency department (ED), and were admitted either directly from the scene or from a transfer facility. Notably, patients with isolated limb injuries and those who were dead on arrival were excluded from the TITCO dataset during data collection and were consequently also excluded from the NTDB dataset.20 Patients who died within the ED were, however, included in the analysis. Exclusion criteria also encompassed patients under 18 years of age, burn injury patients, and patients in the NTDB who did not present to university hospitals.
The analysis encompassed patient demographics, injury patterns, physiological parameters upon presentation, and in-hospital trauma mortality data. The Abbreviated Injury Scale (AIS) for each body region and the Injury Severity Score (ISS) were employed for both overall analysis and detailed subanalyses. Statistical comparisons between the two patient groups utilized univariate statistics, including counts, percentages, means with standard deviations (SDs), medians with interquartile ranges (IQRs), t-tests, Wilcoxon rank-sum tests, and Pearson χ2 tests. Multivariate logistic regression models were then constructed to compare risk-adjusted in-hospital mortality and to perform subgroup analyses. In-hospital mortality served as the primary dependent outcome variable, with the patient’s location (India vs. USA) as the principal independent variable. Logistic regression models were adjusted for age, sex, physiological status, and injury severity (ISS). Statistical significance was pre-defined at a p-value of less than 0.001.
All statistical analyses were conducted using Stata V.16 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, Texas: StataCorp).
Results
Patient Characteristics and Unadjusted Outcomes
From an initial pool of nearly 2.6 million trauma patient records, 675,611 from the US NTDB and 11,796 from the India TITCO database met the study’s inclusion criteria, resulting in a total cohort of 687,407 patients (figure 1). Indian patients presented with a significantly younger mean age of 38.7 years (SD 15.8) compared to 48.03 years (SD 20.5) in the USA, and demonstrated a significantly higher male predominance (83.4% vs. 68.9%). In the USA, a significantly larger proportion of patients were admitted due to falls (31.3% vs. 30.4%), penetrating trauma (11.0% vs. 4.8%), and gunshot wounds (5.6% vs. 0%). Conversely, India exhibited a significantly higher proportion of patients presenting with road traffic injuries (49.3% vs. 37.8%), encompassing injuries related to all motor vehicles, motorcycles, bicycles, three-wheeled vehicles, and pedestrians (table 1).
Figure 1. Patient Flow Diagram
Figure 1: Patient flow diagram illustrating the selection process from over 2.6 million patients across the NTDB and TITCO databases to the final study cohort of 687,407 patients. Exclusion criteria were applied in parallel, with numbers representing true counts for each category, not sequential exclusions. NTDB: National Trauma Data Bank, TITCO: Towards Improved Trauma Care Outcomes.
Table 1. Patient Demographics, Physiology, Anatomy, and Outcomes: USA vs. India
| Characteristic | USA (NTDB) | India (TITCO) | P value |
|---|---|---|---|
| Demographics | |||
| Age mean (SD) | 48.0 (20.5) | 38.7 (15.8) | <0.001 |
| Sex (% male) | 68.9 | 83.4 | <0.001 |
| Mechanism of Injury | |||
| Fall (%) | 31.3 | 30.4 | <0.001 |
| Road traffic injury (%) | 37.8 | 49.3 | <0.001 |
| Penetrating injury (%) | 11.0 | 4.8 | <0.001 |
| Firearm injury (%) | 5.6 | 0 | <0.001 |
| Physiology | |||
| RR median (IQR) | 18 (16–20) | 20 (18–22) | <0.001 |
| Respiratory distress (RR 29) (%) | 6.8 | 27.2 | <0.001 |
| SBP median (IQR) | 136 (120–152) | 118 (110–130) | <0.001 |
| Shock (SBP <90) (%) | 4.1 | 6.3 | <0.001 |
| GCS median (IQR) | 15 (14–15) | 15 (8–15) | <0.001 |
| Altered mental status (GCS ≤13) (%) | 16.4 | 44.5 | <0.001 |
| Anatomy | |||
| AIS head ≥3 (%) | 25.7 | 54.7 | <0.001 |
| AIS face ≥3 (%) | 0.2 | 0.1 | 0.150 |
| AIS chest ≥3 (%) | 22.5 | 8.1 | <0.001 |
| AIS abdomen ≥3 (%) | 5.5 | 4.9 | <0.001 |
| AIS extremity ≥3 (%) | 9.7 | 8.2 | <0.001 |
| AIS external ≥3 (%) | 0.03 | 0.25 | <0.001 |
| ISS median (IQR) | 9 (5–17) | 9 (9–14) | <0.001 |
| ISS ≥25 (%) | 11.8 | 7.5 | <0.001 |
| Outcome | |||
| Mortality (%) | 2.8 | 23.2 | <0.001 |
Table 1: Comparison of demographic, physiologic, anatomic, and outcome characteristics between trauma patients in the USA (NTDB) and India (TITCO). AIS: Abbreviated Injury Scale, ISS: Injury Severity Scale, NTDB: National Trauma Data Bank, SBP: Systolic Blood Pressure, TITCO: Towards Improved Trauma Care Outcomes, RR: Respiratory Rate, GCS: Glasgow Coma Scale.
Significant disparities were observed in physiological parameters upon presentation. A considerably higher proportion of patients in India exhibited respiratory abnormalities (respiratory rate 29: 27.2% vs. 6.8%), circulatory compromise (systolic blood pressure <90 mmHg: 6.3% vs. 4.1%), and neurological dysfunction (Glasgow Coma Scale score ≤13: 44.5% vs. 16.4%) compared to their US counterparts (table 1). Anatomically, significant differences were noted across various body regions, most prominently in head injuries, where a substantially greater percentage of Indian patients presented with serious or critical head injuries (AIS score ≥3: 54.7% vs. 25.7%). While the median ISS was identical at 9 for both groups, the interquartile ranges differed significantly (Wilcoxon rank-sum test), and a higher proportion of patients in the US cohort sustained severe injuries (ISS >25: 11.8% vs. 7.49%). Crude, unadjusted mortality rates were more than eight times higher in India (23.15% vs. 2.79%) (table 1).
Adjusted Mortality and Subgroup Analyses
Multivariate logistic regression analysis identified age, male sex, physiological abnormalities, and ISS as independent predictors of mortality. After controlling for these factors, the strongest predictor of mortality was treatment location, with patients in India exhibiting an odds ratio (OR) of 13.85 (95% CI 13.05 to 14.69). All covariates included in the model were statistically significant predictors of mortality (p<0.001) (figure 2).
Figure 2. Multivariate Logistic Regression of Mortality Predictors
Figure 2Figure 2: Forest plot illustrating the odds ratios (ORs) and 95% confidence intervals (CIs) from multivariate logistic regression analysis of independent mortality predictors. All variables shown were statistically significant (p<0.001). GCS: Glasgow Coma Scale, ISS: Injury Severity Score, SBP: Systolic Blood Pressure.
Subgroup analyses comparing adjusted mortality between India and the USA were conducted across various demographic, physiological, and anatomical injury characteristics. In every subgroup analyzed, adjusted mortality in India remained significantly higher than in the USA. The odds of mortality in India were elevated for patients younger than 65 years and for those presenting with initially normal physiological parameters. While confidence intervals overlapped in subgroup analysis by sex, no significant difference in mortality odds was found between males and females in India. Among injury mechanisms, road traffic injuries (including pedestrians and vehicle occupants) were associated with the highest mortality odds in India (table 2).
Table 2. Odds of Mortality in India by Patient Subgroups
| Characteristic | OR | 95% CI | P value |
|---|---|---|---|
| Demographics | |||
| Ages 18–39 | 13.74 | (12.61 to 14.96) | <0.001 |
| Ages 40–64 | 15.29 | (13.87 to 16.87) | <0.001 |
| Age >65 | 9.50 | (8.00 to 11.30) | <0.001 |
| Males | 13.76 | (12.90 to 14.69) | <0.001 |
| Females | 13.82 | (11.94 to 15.99) | <0.001 |
| Mechanism of Injury | |||
| Fall | 14.50 | (12.92 to 16.28) | <0.001 |
| Road traffic injury | 18.17 | (16.68 to 19.79) | <0.001 |
| Penetrating injury | 4.95 | (3.31 to 7.39) | <0.001 |
| Physiology | |||
| RR 29 | 8.05 | (7.24 to 8.95) | <0.001 |
| RR ≥10 or ≤29 | 15.65 | (14.57 to 16.82) | <0.001 |
| SBP <90 | 10.17 | (8.79 to 11.77) | <0.001 |
| SBP ≥90 | 14.66 | (13.75 to 15.64) | <0.001 |
| GCS ≤13 | 10.61 | (9.94 to 11.35) | <0.001 |
| GCS >13 | 22.59 | (20.06 to 25.44) | <0.001 |
| Anatomy (AIS Region) | |||
| Head (AIS 1–2) | 57.79 | (44.58 to 74.91) | <0.001 |
| Head (AIS 3–6) | 9.02 | (8.39 to 9.69) | <0.001 |
| Face (AIS 1–2) | 20.02 | (17.37 to 23.09) | <0.001 |
| Face (AIS 3–6) | 2.53 | (0.48 to 13.35) | 0.272 |
| Chest (AIS 1–2) | 20.78 | (16.23 to 26.63) | <0.001 |
| Chest (AIS 3–6) | 10.76 | (8.90 to 13.00) | <0.001 |
| Abdomen (AIS 1–2) | 14.93 | (11.58 to 19.25) | <0.001 |
| Abdomen (AIS 3–6) | 8.06 | (6.35 to 10.22) | <0.001 |
| Extremity (AIS 1–2) | 17.62 | (15.08 to 20.59) | <0.001 |
| Extremity (AIS 3–6) | 16.50 | (13.74 to 19.82) | <0.001 |
| External (AIS 1–2) | 15.57 | (14.08 to 17.22) | <0.001 |
| External (AIS 3–6) | 6.26 | (2.14 to 18.31) | 0.001 |
Table 2: Odds of mortality in India compared to the USA, stratified by demographic, physiologic, and anatomic injury characteristics from multivariate logistic regression models. AIS: Abbreviated Injury Scale, GCS: Glasgow Coma Scale, SBP: Systolic Blood Pressure, RR: Respiratory Rate.
Analysis of anatomical injuries, considering injury presence in each region irrespective of injuries in other regions, revealed significantly higher mortality odds in India across all AIS regions, except for severe facial injuries. Furthermore, for mild to moderate injuries (AIS score 1–2) compared to more severe injuries (AIS score 3–6) within each AIS region, mortality odds were significantly elevated in India, with the exception of extremity injuries, where confidence intervals overlapped (table 2). Subgroup analysis based on overall injury burden using ISS demonstrated increased mortality odds in India across all ISS categories, with the most pronounced difference observed in the least severely injured patients (figure 3).
Figure 3. Mortality Odds by Injury Severity Score (ISS) Category
Figure 3: Forest plot representing mortality odds in India from multivariate logistic regression models for each Injury Severity Score (ISS) category, with 95% confidence intervals (CIs). All ISS categories demonstrated statistically significant differences (p<0.001).
Discussion
Over recent decades, LMICs have achieved substantial progress in reducing mortality from communicable diseases. However, the decline in injury-related mortality has not kept pace.21 This disparity, coupled with increasing populations, urbanization, and mechanization, has resulted in a rise in the proportion of deaths attributable to injury.22 Conversely, HICs have witnessed remarkable advancements across the spectrum of trauma care, leading to significant reductions in trauma-related mortality and long-term disability.
While existing studies compare outcomes, primarily mortality, between LMICs and HICs, the vast majority focus on crude mortality rates. Without adjusted mortality analyses that account for variations in patient populations, physiological characteristics, and injury patterns, a clear understanding of the true differences remains elusive. Crucially, the absence of detailed subgroup analyses that identify specific patient populations and injury types contributing most significantly to mortality impedes the development of targeted interventions to effectively reduce trauma-related mortality in LMICs. This study addresses this critical gap by comparing adjusted mortality between an LMIC (India) and an HIC (USA). Furthermore, the detailed subgroup analysis pinpoints specific patient and injury profiles that represent key areas for mortality reduction efforts.
The independent predictors of mortality identified in this study align with well-established factors: chronic physiology (age), acute physiology (abnormal respiratory rate, hypotension defined as systolic blood pressure below 90 mmHg, and altered neurological status indicated by a GCS score of 13 or less), and the extent of anatomical injury (AIS and ISS). After adjusting for these known mortality predictors, the most significant factor influencing mortality was the location of treatment – India. While crude mortality was eight times higher in India, adjusted mortality was over 13 times higher. This divergence between crude and adjusted mortality underscores the importance of risk adjustment for an accurate and nuanced understanding of the problem. The subgroup analyses revealed a particularly striking finding: younger patients, those with less severe injuries, and patients with initially normal physiological parameters exhibited higher odds of mortality in India. These seemingly paradoxical findings suggest that among older, more physiologically compromised, and severely injured patients, mortality may be inevitable regardless of care quality.
A comprehensive investigation into the precise causes of mortality in both cohorts is beyond the scope of this study. However, a prior study utilizing the same TITCO dataset employed a consensus-based Delphi review to assess preventable deaths and opportunities for improvement. This earlier study concluded that nearly 50% of deaths were considered preventable, and identified broad areas for improvement including: appropriate management of head injuries (23.3%); timely resuscitation and hemorrhage control (16.8%); effective airway management (14.3%); development and adherence to treatment protocols (12.7%); reducing prehospital delays (10.3%); and preventing ventilator-associated complications (5.1%).12
These findings corroborate the risk-adjusted analysis of the current study, which demonstrates that the highest mortality odds in India are observed in younger patients with less physiological derangement and lower injury burden. Both the current risk-adjusted trauma mortality study and the previous Delphi review on preventable deaths suggest that relatively simple, low-cost interventions focused on providing standard, timely trauma care and implementing protocolized treatment pathways – rather than relying on technologically advanced and expensive interventions – are likely to yield the most substantial reductions in trauma-related mortality in settings like India.
Organized trauma care systems demonstrably save lives,23–25 yet despite the immense human cost of injury and the availability of evidence-based, low-cost interventions, national and global health agendas have not prioritized trauma care adequately.26 27 Efforts to improve trauma care systems can be effectively guided by essential trauma care guidelines outlined by the WHO.13 28 While a detailed discussion of specific interventions is beyond the scope of this study, focused interventions with potentially high impact in this context include:
Like all research, this study has certain limitations. Firstly, it exclusively examines trauma patients presenting to university medical centers, which may limit the generalizability of findings to rural, community hospitals, or non-academic medical centers. Secondly, the retrospective design inherently introduces the potential for confounding. Thirdly, the study only includes trauma patients who were alive upon hospital arrival, thus excluding deaths occurring at the scene, during transport, or at transfer facilities, and does not account for spatiotemporal factors such as travel distances and times. Finally, adjusted mortality based on injury severity relies on the accurate determination of ISS, which in turn depends on the extent of imaging utilized to detect and document even minor injuries. ISS may underestimate the true injury burden in resource-limited settings where imaging is less frequently employed, or in cases where patients do not survive long enough to undergo comprehensive imaging. Further research comparing diagnostic and therapeutic interventions is warranted to more definitively identify and correlate specific differences in care processes with observed outcome disparities.
Conclusion
Despite these acknowledged limitations, this study provides compelling evidence that, even after adjusting for demographic differences, physiological abnormalities, and injury severity, trauma-related mortality remains significantly higher in India compared to the USA. Notably, when compared to trauma patients in the USA, the risk-adjusted odds of mortality are most pronounced in younger patients, those with normal physiological parameters upon arrival, and individuals with mild to moderate injuries. While these findings are concerning, they also offer a hopeful perspective, suggesting that implementing relatively straightforward, low-cost interventions focused on delivering standard, timely trauma care and establishing protocolized treatment pathways could lead to substantial improvements in injury-related mortality in India. Future research should prioritize evaluating differences in trauma diagnostic and therapeutic interventions to pinpoint the essential components of effective trauma systems that contribute most significantly to the observed outcome disparities between India and the USA.
Acknowledgments
We extend our gratitude to David Amato for his invaluable assistance in data cleaning and merging the datasets.
Footnotes
Contributors: NR and MM were instrumental in designing the data collection tools and overseeing data collection in India. SA and LB developed the statistical analysis plan and were responsible for data cleaning and analysis. SA and AM drafted and revised the manuscript. SA, LB, MM, NR, and AM contributed to the final edits. SA and AM take full responsibility for the study’s conduct, had full access to the data, and controlled the decision to publish.
Funding: The authors have declared no specific funding was received from any public, commercial, or not-for-profit funding agency for this research.
Competing interests: None declared.
Provenance and peer review: Not commissioned; externally peer reviewed.
Data Availability Statement
TITCO data are publicly available in an open-access repository. NTDB data are available upon reasonable request.
Ethics Statements
Patient consent for publication
Not applicable.
Ethics approval
Data in both the NTDB and TITCO databases were de-identified, and this study was granted exemption from human subjects review by the University of Vermont Institutional Review Board (STUDY00001272).
References
[References from original article would be listed here]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
TITCO data are publicly available in an open-access repository. NTDB data are available upon reasonable request.
