Sage Journals: Discover world-class research

Abstract

Objective: Many current trauma mortality prediction tools are either too intricate or rely on data not readily available during a trauma patient’s initial evaluation. Moreover, none are tailored to those necessitating urgent or emergent surgery. Our objective was to design a practical, user-friendly scoring tool using immediately available variables, and then compare its efficacy to the widely-known Revised Trauma Score (RTS). Methods: The adult 2017-2021 Trauma Quality Improvement Program (TQIP) database was queried to identify patients ≥18 years old undergoing any urgent/emergent operation (direct from Emergency Department to operating room). Patients were divided into derivation and validation groups. A three-step methodology was used. First, multiple logistic regression models were created to determine risk of death using only variables available upon arrival. Second, the weighted average and relative impact of each independent predictor was used to derive an easily calculated Immediate Operative Trauma Assessment Score (IOTAS). We then validated IOTAS using AUROC and compared it to RTS. Results: From 249 208 patients in the derivation-set, 14 635 (5.9%) died. Age ≥65, Glasgow Coma Scale score <9, hypotension (SBP <90 mmHg), and tachycardia (>120/min) on arrival were identified as independent predictors for mortality. Using these, the IOTAS was structured, offering scores between 0-8. The AUROC for this was .88. A clear escalation in mortality was observed across scores: from 4.4% at score 1 to 60.5% at score 8. For the validation set (250 182 patients; mortality rate 5.8%), the AUROC remained consistent at .87, surpassing RTS’s AUROC of .83. Conclusion: IOTAS is a novel, accurate, and now validated tool that is intuitive and efficient in predicting mortality for trauma patients requiring urgent or emergent surgeries. It outperforms RTS, and thereby may help guide clinicians when determining the best course of action in patient management as well as counseling patients and their families.

Keywords

blunt trauma mortality scoring tool trauma and injury severity

Introduction

In 2019 alone, unintentional injuries accounted for approximately 170 000 deaths in the United States (US).¹ While the development of regional trauma systems has significantly diminished morbidity and mortality linked to trauma,^2–4 challenges of risk prediction remain a significant hurdle. While valuable in assessing injury severity and estimating outcomes, current trauma mortality prediction scores carry notable flaws, particularly in the early phases of trauma care.

Multiple trauma mortality prediction tools have been developed in order to enhance triage and direct resource allocation.^5–10 Unfortunately, none are widely used in clinical practice. In 1989, Champion and colleagues introduced the Revised Trauma Score (RTS), a triage scoring system that incorporates physiologic parameters including the Glasgow Coma Scale (GCS), systolic blood pressure (SBP), and respiratory rate (RR) to predict mortality.⁹ However, calculation of the RTS is too complicated for use in the emergency department (ED) and lacks injury specific data. Other trauma scores such as the Injury Severity Score (ISS) and Trauma and Injury Severity Score (TRISS) require information not readily available upon patient presentation, and none are intended specifically for use in patients requiring urgent or emergent surgery after trauma. Simpler scoring systems, such as the Glasgow Coma Scale, Age, and Systolic Blood Pressure (GAP) score, have been introduced to utilize readily accessible information but lack specificity for patients requiring urgent surgical intervention—a critical need considering that over 10% of all trauma activations necessitate emergency surgery, and represent a far larger proportion of those with significant injuries, at highest risk for death.¹¹

Recognizing this need, this study aimed to develop a novel, easy-to-use and user-friendly Immediate Operative Trauma Assessment Score (IOTAS) tool designed for immediate application in predicting mortality among trauma patients requiring urgent or emergent surgical intervention. In addition, we aimed to compare this against RTS, hypothesizing improved mortality risk prediction accuracy.

Methods

This study was deemed exempt by our institutional review board as it utilizes a national deidentified database. The 2017-2021 Trauma Quality Improvement Program (TQIP) database, a prospectively maintained database including data from over 850 participating centers in the US, was queried to identify patients >18 years old undergoing any urgent or emergent operations, defined by disposition from the ED to the operating room within 6 hours of arrival.

The methodology involved two primary phases: first a development phase of the IOTAS, and then a validation phase. Randomization was performed with the Markov chain approach to produce two equally sized datasets.¹² Both groups included patients undergoing any type of urgent or emergent surgery, with no specific surgical procedure exclusions. The primary outcome assessed was in-hospital mortality. In addition, patient demographic information was collected including age, GCS score, and vitals upon arrival.

In developing the IOTAS, a three-step methodology was employed, which has been previously described.^13–15 First, multiple logistic regression models were created to determine the risk of death using only variables available at the time of patient arrival and within the derivation dataset. These variables were chosen based on a consensus among co-authors and by a review of the literature for known risk factors of mortality in trauma.^16–18 All variables chosen were available upon arrival to the ED, thus allowing us to derive a score consisting of only immediately available variables. Next, variables were then included in a multivariable forward stepwise logistic regression model to identify independent risk factors for mortality using P-value <.05 as the cutoff for statistical significance. The variables that were ultimately selected were age ≥65 years old, GCS <9, tachycardia (HR >120/min) and hypotension (SBP <90 mmHg) upon arrival. The binary nature of all the variables were selected to ensure IOATS was simple and practical to use. The weighted average and relative impact of each independent risk factor was then used to derive the IOTAS. This was done by using the odds ratio and dividing by the lowest common denominator (ie, the lowest odds ratio). This process produced non-integer values which were subsequently rounded to the nearest whole number. To account for error, several iterations were developed for an easy-to-use scoring tool. After each iteration, the receiver operating characteristics (ROC) curve was examined to ensure consistency in the concordance (C) statistic. The C statistic, or area under the receiver operative characteristic (AUROC) curve, measures model success and refers to the ability to discriminate between outcome of interest (death) vs cases without that outcome (survival).¹⁹

Finally, the IOTAS was validated using the validation dataset. The tool’s predictive accuracy was assessed using its C statistic and ability to predict mortality. The IOTAS was then compared to the well-established RTS within the validation set using AUROC. Additional subset analyses of only blunt trauma, only penetrating trauma, alcohol/illicit drug screen-positive, alcohol/illicit drug screen-negative, and a subgroup without traumatic brain injury were performed. These were chosen as some risk prediction tools have been found to perform better for blunt or penetrating trauma and traumatic brain injury can skew risk prediction as well.^6,20–23 All analyses were performed with IBM SPSS Statistics for Windows (Version 29, IBM Corp., Armonk, NY).

Results

A total of 499 390 patients requiring urgent or emergent surgery directly from the ED were identified. Of the 249 208 patients in the derivation set, 14 635 (5.9%) patients died. Most patients were male (70.6%), the most common mechanism of injury was blunt (69.2%), and the most common abdominal and thoracic injuries were to the liver (6.9%) and lung (15.1%), respectively (Table 1).

Table 1.

Mechanisms of Injury and Injuries of Patients Included in the IOTAS Derivation Data Set.

Characteristic	Survived (n = 234 573)	Died (n = 14 635)	P-value
Mechanism
Blunt	161 822 (69.0)	10 648 (72.7)	<.001
Fall	63 931 (27.3)	3452 (23.6)	<.001
Penetrating	68 081 (29.0)	3954 (27)	<.001
Injury
Traumatic brain injury	21 433 (9.1)	7231 (49.4)	<.001
Spine fracture	29 493 (12.6)	4001 (27.3)	<.001
Heart	1867 (.8)	1123 (7.7)	<.001
Rib fracture	29 556 (12.6)	5222 (35.7)	<.001
Lung	31 025 (13.2)	6679 (45.6)	<.001
Diaphragm	5880 (2.5)	1193 (8.2)	<.001
Esophagus	155 (.1)	38 (.3)	<.001
Spleen	12 296 (5.2)	2413 (16.5)	<.001
Liver	13 532 (5.8)	3602 (24.6)	<.001
Pancreas	2619 (1.1)	767 (5.2)	<.001
Stomach	3614 (1.5)	649 (4.4)	<.001
Small intestine	13 797 (5.9)	1972 (13.5)	<.001
Colon	12 484 (5.3)	1704 (11.6)	<.001
Rectum	1679 (.7)	150 (1.0)	<.001
Kidney	6325 (2.7)	1241 (8.5)	<.001
Bladder	2804 (1.2)	375 (2.6)	<.001
Upper extremity fracture	44 505 (19)	1888 (12.9)	<.001
Lower extremity fracture	88 721 (37.8)	3059 (20.9)	<.001

Data are presented as n (%).

Age ≥65 years old, GCS <9, hypotension (SBP <90 mmHg), and tachycardia (HR >120/min) on arrival were identified as independent predictors for mortality (Table 2). Using these variables, IOTAS was structured, offering scores between 0 and 8. A clear escalation in mortality was observed across scores from 4.4% to 26.5% and then 63.2% at scores of 1, 3, and 6, respectively (Figure 1). The AUROC was .88 within the derivation set (Table 3, Figure 2). There were no significant differences between the derivation and validation sets in regard to mechanisms of injury, vitals, and injury patterns (P < .05).

Table 2.

Development of the Immediate Operative Trauma Assessment Score (IOTAS) to Determine Risk of Death in Urgent/Emergently Operated Trauma Patients.

Variable	Points
Glasgow Coma Scale <9	3
Heart rate >120 beats per minute	2
Systolic blood pressure <90 mmHg	2
Age≥65 years	1
Total score	8
AUROC	.875
95% CI for AUROC	.871-.878

IOTAS Score = (age65 * 1) + (GCS <9 * 3) + (hypotensive * 2) + (tachycardic * 2).

Figure 1.

Immediate Operative Trauma Assessment Score (IOTAS) mortality rate.

Table 3.

AUROC for the IOTAS in Derivation and Validation Sets and the RTS.

	AUROC	Standard error	95% CI	P-value
Derivation set	.875	.002	.871-.878	<.001
Validation set	.873	.002	.869-.876	<.001
Revised Trauma Score	.832	.003	.828-.837	<.001

Figure 2.

AUROC for the IOTAS in derivation set (AUROC = .88).

There were 250 182 patients included in the validation set, with an in-hospital mortality rate of 5.8%. The AUROC for the validation set remained consistent at .87, surpassing the RTS’s AUROC of .83 (Table 3, Figures 3–4). The IOTAS continued to perform well in subset analyses of patients presenting with only penetrating mechanisms of injury (AUROC = .87) and only blunt mechanisms (AUROC = .87), as well as a subgroup without traumatic brain injury (AUROC = .87), substance-positive (AUROC = .85), and substance-negative (AUROC = .84) (Table 4, Figures 5 –7).

Figure 3.

Area under the receiver operator characteristics curve for the IOTAS in validation set (AUROC = .87).

Figure 4.

Area under the receiver operator characteristics curve for the revised trauma score (RTS) (AUROC = .83).

Table 4.

AUROC for the IOTAS in Subset Analyses.

	AUROC	Standard error	95% CI	P-value
Only penetrating mechanisms	.872	.004	.865-.879	<.001
Only blunt mechanisms	.874	.002	.870-.877	<.001
Without TBI	.866	.003	.861-.871	<.001
Substance positive	.854	.003	.848-.861	<.001
Substance negative	.841	.005	.831-.850	<.001

TBI, traumatic brain injury, substance: illicit-drugs or alcohol.

Figure 5.

Area under the receiver operator characteristics curve for the IOTAS in subgroup analysis without traumatic brain injury (AUROC = .87).

Figure 6.

Area under the receiver operator characteristics curve for the IOTAS subgroup analysis with only penetrating injuries (AUROC = .87).

Figure 7.

Area under the receiver operator characteristics curve for the IOTAS subgroup analysis with only blunt injuries (AUROC = .87).

Discussion

Definitive management of the severely injured trauma patient is time sensitive. Existing trauma mortality prediction tools often present challenges: they are either too complex for immediate use or require data not available during the initial assessment of a trauma patient. Moreover, none of these tools are specifically designed for patients in need of urgent or emergent surgical intervention. The aim of this study was to design a practical mortality prediction tool (IOTAS) using immediately available variables, and then compare its efficiency to the widely-known RTS. We demonstrated that IOTAS outperforms the RTS, and thereby may help guide clinicians in practice when counseling trauma patients and their families and can be utilized to compare outcomes between trauma centers.

Multiple trauma scoring tools have been developed and used. The RTS remains one of the most frequently referenced tools and incorporates physiologic parameters including GCS, SBP, and RR. It is not only pivotal in assessing the severity of traumatic injuries but also plays a crucial role in quality control outcomes for trauma centers and provides a “common language” for research studies, facilitating the comparison of trauma populations across different clinical outcome studies. The RTS also comprises the content of the TRISS,⁹ which has been shown to strongly predict probability of survival because it incorporates mechanism of injury, anatomic and physiological variables.²³ However, calculation of either the RTS or TRISS is too cumbersome to be performed in the ED and require a complete knowledge of the patient’s injuries, limiting their utility in the early phases of trauma care. Further limitations of the RTS include the poor interrater reliability, particularly when used in the emergency setting.²⁴ RR, a component of the RTS, is less reliable than other variables due to influences of patient age, mechanism of injury and mechanical ventilation.¹⁰ Validation studies of the RTS have also recognized RR as having poor predictive value for mortality compared to GCS and SBP.²⁵ Moreover, increasing age, a crucial factor in mortality prediction, is incorporated in IOTAS but absent in the RTS.^26,27 To overcome these limitations, the IOTAS employs a binary scoring system that relies solely on readily available variables, while also integrating factors known to significantly influence mortality risk. Furthermore, IOTAS is derived from a cohort of trauma patients undergoing urgent or emergent surgery, enhancing its applicability in high-acuity trauma care, where rapid and accurate decision-making is paramount.

The GAP score, introduced in 2011, represents another evolution in trauma mortality scoring tools.¹⁰ Derived and validated from data in the Japanese Trauma Data Bank (2004-2009), the GAP score is an adaptation of the earlier Mechanism, Glasgow Coma Scale, Age, and Systolic Blood Pressure (MGAP) score derived by Sartorius and colleagues.²⁸ The GAP score strategically omits the ‘mechanism’ element to enhance its generalizability, particularly relevant given the predominance of blunt injuries over penetrating injuries in their study cohort.¹⁰ While the GAP score efficiently predicts short-term mortality and mirrors the RTS in terms of trauma severity prediction, the IOTAS introduces several enhancements. The IOTAS not only incorporates all the variables of the GAP score but also introduces heart rate (HR), a vital parameter for assessing hypovolemic shock. Isolated vital signs have previously been shown to be unreliable in the assessment of hypovolemic shock; however, the Shock Index (ratio of HR to SBP) is a well-studied and validated physiologic indicator of hypovolemic shock.²⁹ In the development of IOTAS, tachycardia (HR >120/min) on arrival was identified as independent predictor for mortality and was therefore incorporated into the final version of the scoring tool. And finally, its design is characterized by a binary, yes-or-no approach for each component, simplifying its use in high-pressure emergency settings.

The incorporation of GCS into trauma scoring systems is widely accepted due to its proven significance as a strong predictor of in-hospital mortality.^30,31 However, an inherent limitation within the GAP score arises from the weighted value assigned to GCS in categorizing trauma severity and mortality risk. Notably, patients with a normal GCS (GCS 15) are automatically excluded from the “severe” (>50% risk of death) classification according to the GAP score criteria.¹⁰ In addition, substance abuse can impair neurological function, leading to alterations in consciousness levels and cognitive abilities, potentially influencing GCS scores on presentation. These discrepancies in GCS assessments due to substance abuse can introduce variability and inaccuracies into trauma scoring tools that incorporate GCS as a predictor, potentially effecting the precision of mortality risk predictions in clinical settings. To overcome these constraints, the IOTAS integrates a more comprehensive set of variables and continued to perform well in subset analyses, demonstrating reliability in patients screening positive for illicit-drugs or alcohol, patients screening negative for illicit-drugs or alcohol, and in patients without traumatic brain injury. This not only enhances the reliability of the IOTAS but also allows its use in a wide range of trauma patients requiring urgent or emergent operation. Furthermore, unlike other scoring systems that rely on incorporation of specific injury mechanisms, the IOTAS is able to maintain efficacy across both penetrating (AUROC = .87) and blunt mechanisms (AUROC = .87) without explicitly incorporating the mechanism into the scoring system. Thus, it can be applied to urban or rural trauma and potentially used to compare differing types of trauma centers (eg, predominantly penetrating or blunt volume).

This retrospective database study has several limitations inherent to large dataset studies including the potential for misclassification, missing data, reporting bias, and coding errors. It is possible that there are other variables more predictive of mortality risk that are not present within the TQIP database or not evaluated in this study. Additionally, we were unable to calculate mortality risk in IOTAS scores greater than 6 due to sample size limitations. We suspect these rates would continue to increase at higher IOTAS scores; however, further studies with larger sample sizes (possibly with additional years of TQIP data) would be required to verify this assumption. Furthermore, the study does not account for the dynamic nature of trauma care, where patient conditions can rapidly evolve, potentially affecting the applicability of the scoring system in real-time decision-making. Despite these limitations, the IOTAS is derived from and validated using a large national dataset thereby demonstrating generalizability and provides a novel mortality scoring system using only immediately available variables to predict mortality risk in a less studied but high-risk cohort of those undergoing urgent/emergent operations.

Conclusion

In summary, the IOTAS is a validated, practical, and reliable tool to accurately predict mortality in trauma patients requiring urgent or emergent surgical procedures. The IOTAS demonstrates superior performance over the RTS, and effectiveness across various patient subgroups, including those without traumatic brain injury, those screening positive for illicit-drugs/alcohol, and those with either penetrating or blunt trauma mechanisms. Additionally, since this tool is used to predict mortality in those requiring urgent or emergent operations, IOTAS is likely more applicable to critically ill trauma patients. The versatility of IOTAS underlines its potential; however, future multicenter validation studies are needed prior to widespread adoption.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Heron

. Deaths: leading causes for 2019. Natl Vital Stat Rep. 2021;70(9):1-114.

Brown

Rosengart

Kahn

, et al. Impact of volume change over time on trauma mortality in the United States. Ann Surg. 2017;266(1):173-178.

Lansink

Leenen

. Do designated trauma systems improve outcome? Curr Opin Crit Care. 2007;13(6):686-690.

MacKenzie

Rivara

Jurkovich

, et al. A national evaluation of the effect of trauma-center care on mortality. N Engl J Med. 2006;354(4):366-378.

Santos

Kuza

Luo

, et al. Comparison of national surgical quality improvement program surgical risk calculator and trauma and injury severity score risk assessment tools in predicting outcomes in high-risk operative trauma patients. Am Surg. 2023;89(10):4038-4044.

Stopenski

Kuza

Luo

, et al. Comparison of national surgical quality improvement Program surgical risk calculator, trauma and injury severity score, and American society of anesthesiologists physical status to predict operative trauma mortality in elderly patients. J Trauma Acute Care Surg. 2022;92(3):481-488.

Kuza

Matsushima

Mack

, et al. The role of the American Society of anesthesiologists physical status classification in predicting trauma mortality and outcomes. Am J Surg. 2019;218(6):1143-1151.

Yeates

Nahmias

Gabriel

, et al. A prospective multicenter comparison of trauma and Injury Severity Score, American society of anesthesiologists physical status, and national surgical quality improvement Program calculator’s ability to predict operative trauma outcomes. Anesth Analg 2023.

Champion

Sacco

Copes

Gann

Gennarelli

Flanagan

. A revision of the trauma score. J Trauma. 1989;29(5):623-629.

10.

Kondo

Abe

Kohshi

Tokuda

Cook

Kukita

. Revised trauma scoring system to predict in-hospital mortality in the emergency department: Glasgow Coma Scale, Age, and Systolic Blood Pressure score. Crit Care. 2011;15(4):R191.

11.

Chang

. National Trauma Data Bank Annual Report. American College of Surgeons. https://www.facs.org/media/ez1hpdcu/ntdb-annual-report-2016.pdf (2016).

12.

Gionis

AMH

Mielik¨ainen

Tsaparas

. Assessing data mining results via swap randomization. ACM Trans Knowl Discov Data 2007;167-176.

13.

Stopenski

Grigorian

Inaba

, et al. Prehospital variables alone can predict mortality after blunt trauma: a novel scoring tool. Am Surg. 2021;87(10):1638-1643.

14.

Kazempoor

Nahmias

Clark

, et al. Scoring tool to predict need for early video-assisted thoracoscopic surgery (VATS) after pediatric trauma. World J Surg. 2023;47(11):2925-2931.

15.

Santos

Grigorian

Kuza

, et al. Development and validation of a renal replacement after trauma scoring tool. J Am Coll Surg. 2023;237(1):79-86.

16.

Haider

Chang

Haut

Cornwell

Efron

. Mechanism of injury predicts patient mortality and impairment after blunt trauma. J Surg Res. 2009;153(1):138-142.

17.

Battle

Carter

Newey

Giamello

Melchio

Hutchings

. Risk factors that predict mortality in patients with blunt chest wall trauma: an updated systematic review and meta-analysis. Emerg Med J. 2023;40(5):369-378.

18.

Champion

Copes

Sacco

, et al. The Major trauma outcome study: establishing national norms for trauma care. J Trauma. 1990;30(11):1356-1365.

19.

Hanley

McNeil

. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143(1):29-36.

20.

Smith

Goldberg

Gaughan

Seamon

. A comparison of injury severity score and new injury severity score after penetrating trauma: a prospective analysis. J Trauma Acute Care Surg. 2015;79(2):269-274.

21.

Millham

LaMorte

. Factors associated with mortality in trauma: re-evaluation of the TRISS method using the National Trauma Data Bank. J Trauma. 2004;56(5):1090-1096.

22.

Camarano

Ratliff

Korst

Hrushka

Jupiter

Predicting in-hospital mortality after traumatic brain injury: external validation of CRASH-basic and IMPACT-core in the national trauma data bank. Injury. 2021;52(2):147-153.

23.

Chun

Zhang

Becnel

, et al. New injury severity score and trauma injury severity score are superior in predicting trauma mortality. J Trauma Acute Care Surg. 2022;92(3):528-534.

24.

Raux

Thicoïpé

Wiel

, et al. Comparison of respiratory rate and peripheral oxygen saturation to assess severity in trauma patients. Intensive Care Med. 2006;32(3):405-412.

25.

Moore

Lavoie

LeSage

, et al. Statistical validation of the revised trauma score. J Trauma. 2006;60(2):305-311.

26.

Adams

Cotton

McGuire

, et al. Unique pattern of complications in elderly trauma patients at a Level I trauma center. J Trauma Acute Care Surg. 2012;72(1):112-118.

27.

Hashmi

Ibrahim-Zada

Rhee

, et al. Predictors of mortality in geriatric trauma patients: a systematic review and meta-analysis. J Trauma Acute Care Surg. 2014;76(3):894-901.

28.

Sartorius

Le Manach

David

, et al. Mechanism, glasgow coma scale, age, and arterial pressure (MGAP): a new simple prehospital triage score to predict mortality in trauma patients. Crit Care Med. 2010;38(3):831-837.

29.

Mutschler

Nienaber

Münzberg

, et al. The shock index revisited - a fast guide to transfusion requirement? a retrospective analysis on 21,853 patients derived from the TraumaRegister DGU. Crit Care. 2013;17(4):R172.

30.

Nik

Sheikh Andalibi

Ehsaei

Zarifian

Ghayoor Karimiani

Bahadoorkhan

. The efficacy of Glasgow Coma Scale (GCS) Score and acute physiology and chronic health evaluation (APACHE) II for predicting hospital mortality of ICU patients with acute traumatic brain injury. Bull Emerg Trauma. 2018;6(2):141-145.

31.

Emami

Czorlich

Fritzsche

, et al. Impact of Glasgow Coma Scale score and pupil parameters on mortality rate and outcome in pediatric and adult severe traumatic brain injury: a retrospective, multicenter cohort study. J Neurosurg. 2017;126(3):760-767.

Immediate Operative Trauma Assessment Score: A Simple and Reliable Predictor of Mortality in Trauma Patients Undergoing Urgent/Emergent Surgery

Abstract

Keywords

Introduction

Methods

Results

Discussion

Conclusion

Footnotes

Declaration of Conflicting Interests

Funding

References