Characterizing the burden of complications in traumatic brain injury: a retrospective study

Introduction

Traumatic brain injury (TBI) is one of the leading causes of hospitalization, death, or disability worldwide, with between 64–74 million global incidences of TBI annually (1,2). Previously, pediatric TBI alone has contributed more than $1 billion to hospital charges each year (3,4). As the incidence of TBI across all age groups continues to rise, the significant public health concern and financial burden of this “silent epidemic” necessitates the study of factors that may impact outcomes (1,5). Previous research has suggested that complications can increase in-hospital mortality, length of stay, and costs for trauma patients (6-10). Additional research has also attempted to use complication rates as an indicator of trauma care performance (11). However, the influence of complications on TBI outcomes specifically, remains unclear.

Prior research has focused on characterizing the epidemiology of TBI and various factors impacting mortality and injury severity, including associated injuries, blood loss, and mechanism of injury (12-15). To our knowledge, no group has focused on characterizing the burden of complications on TBI outcomes. Therefore, the primary goal of our study was to analyze the burden of twelve distinct complications on discharge disposition and total hospital length of stay (LOS) for isolated TBI patients. We hypothesized that among survivors, the occurrence of a complication would increase the duration of hospital stay and the percentage of TBI patients who require care post-discharge. We focused primarily on patients with isolated TBI to minimize the likelihood that complications were associated with systemic trauma, rather than with the post-injury management of TBI. Nonetheless, we also examined whether trends across isolated and all TBI populations were similar. We present the following article in accordance with the STROBE reporting checklist (available at https://dx.doi.org/10.21037/jeccm-21-53).

Methods

Study population

This study was conducted using data from the 2012–2016 National Trauma Data Bank (NTDB). The NTDB is managed and compiled by the American College of Surgeons (ACS) Committee on Trauma (16). Although more recent years of the NTDB are available, these files do not contain important clinical and facility-related confounding variables, precluding their use in this investigation. The study population consisted of traumatic brain injury patients between ages 0 and 89. TBI patients were identified using the following International Classification of Diseases, 9^th Revision Clinical Modification (ICD-9-CM) and International Classification of Diseases, 10^th Revision Clinical Modification (ICD-10-CM) codes for intracranial injury: 850.0–854.1 (ICD-9-CM) and codes in the S06 category (ICD-10-CM). Isolated TBI was defined as having only ICD-9-CM or ICD-10-CM codes corresponding with TBI. Patients with missing age information were excluded. To identify and remove patients with a low chance of survival, we followed a similar method as Watson et al. (17), eliminating patients who exhibited no signs of life (admission systolic blood pressure =0 and pulse rate =0) and no neurological activity (GCS =3) upon admission to the ED (Figure 1).

Figure 1 Study population exclusion criteria. Data reduction steps using data from the 2012–2016 National Trauma Data Bank. Isolated TBI population subset: n=103,668; Total TBI population: n=968,909.

Demographic and clinical data including gender, age, race, primary payment method, Injury Severity Score derived from hospital-submitted Abbreviated Injury Scores (ISSAIS score), total Glasgow Coma Scale (GCS) score, comorbidities present, injury type (blunt vs. penetrating vs. other/unspecified), mortality, total length of stay (LOS) in the hospital, the occurrence of unplanned intubation, and hospital discharge disposition were isolated for each patient. Only comorbidities present in at least 1% of the patient population were investigated, in order to reduce the risk of model overfitting. These included hypertension requiring medication (33.3%), bleeding disorder (10.0%), congestive heart failure (4.2%), history of stroke or cerebrovascular accident (4.0%), diabetes (14.7%), respiratory disease (7.5%), and history of myocardial infarction (1.8%). Patients with missing GCS or hospital LOS information were removed. To determine the impact of specific complications on TBI outcomes, we identified the occurrence of the following complications based on the native NTDB complications variable: organ space surgical site infection (SSI), superficial SSI, acute respiratory distress syndrome (ARDS), deep vein thrombosis (DVT), myocardial infarction (MI), pneumonia (PNA), pulmonary embolism, stroke or cerebrovascular accident (CVA), cardiac arrest, urinary tract infection (UTI), acute kidney injury (AKI), and sepsis. Only patients who developed a condition after initial presentation were counted as having the complication. For each patient, facility data including hospital teaching status (community or university), region of the hospital, and TBI volume were also obtained from the NTDB.

To investigate the impact of complications on discharge disposition, the native hospital discharge disposition outcomes within the NTDB were grouped as home (discharged home without home services, discharged home under care of organized home service, routine discharge), care facilities (transfer to a long-term care facility, transfer to an intermediate care facility, transfer to a skilled nursing facility, transferred to rehab), or terminal (death or hospice). Patients who left against medical advice, who were transferred to a different hospital for care, or whose outcomes were unknown were excluded. We quantified burden on discharge disposition as the risk difference upon exposure to a complication using the equation for population attributable risk (PAR) for a variable with a single level of exposure (complication vs. no complication). This metric takes into account both the proportion (p) of patients with a complication and relative risk (RR) of that complication for determining a poorer outcome Eq. [1].

PAR=p(RR−1)1+p(RR−1)[1]

LOS burden was calculated as a ratio of number of extra hospital days attributable to a complication divided by the mean LOS.

Statistical analysis

To determine the impact of each complication on isolated TBI outcomes, we performed multivariate regression analysis using the total study population. Analyses controlled for age, sex, race, primary method of payment, region of hospital, hospital teaching status, hospital TBI volume, comorbidities, GCS score, ISS score, and unplanned intubation. The primary outcomes investigated included total hospital LOS and whether the patient was discharged home or to a care facility. Statistical significance was assessed at P<0.05. All analyses were performed using Stata Version 16.1 (StataCorp, College Station, TX, USA).

We also conducted a secondary set of analyses to investigate whether the trends we observed in the isolated TBI population persisted in TBI with polytrauma and across TBI subtypes. Two duplicate sets of analyses were performed using the following TBI populations: isolated TBI and all TBI (including isolated TBI and TBI with polytrauma). For both groups, we separately divided the study population into subcategories based on GCS (severe, moderate, mild) or age (pediatric vs. adult) for analysis. Following earlier research and guidelines (18), patients in the severe, moderate and mild TBI groups were defined as having a total GCS score of 3–8, 9–12 and 13–15, respectively, upon admission to the hospital. Age categories were defined as pediatric (0–18 years) or adult (18–89 years). We controlled for the same variables as in the primary analysis.

Ethics

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). This study was exempt from Institutional Board Review due to the publicly available and de-identified nature of the data in the NTDB. The National Trauma Data Bank is available to researchers in all partner institutions that contribute data to the dataset. Ethical review and approval were not required for this study on human participants in accordance with the local legislation and institutional requirements.

Results

Study population

After applying the exclusion criteria, there were 103,668 isolated TBI admissions to 857 distinct institutions in the 2012–2016 NTDB. 83,735 (80.8%) patients were adults, with a mean age of 49.5 years old [standard deviation (SD) =27.1, range, 1–89]. 19,933 (19.2%) patients constituted the pediatric population. The majority of the total cases were white (72.2%) and male (59.3%), and injuries were most commonly blunt (73.9%). Patients were also more likely to be treated at a university hospital (45.8%) than a community (40.0%) or non-teaching institution (14.2%; Table 1). Within TBI subtypes, there were 9,590 severe TBI patients (mean severe TBI GCS =4.2±1.8), 4,594 moderate TBI patients (mean moderate TBI GCS =10.7±1.1), and 89,484 mild TBI patients (mean mild TBI GCS =14.8±0.5; Table 1). The secondary analysis dataset, which included all types of TBI (non-isolated and isolated) contained 968,909 TBI patients (Table S1).

Prevalence of complications

Pneumonia was one of the most prevalent complications in the isolated TBI population (1.5%), followed by UTI (1.3%; Table 1). Similar trends were also observed in the expanded dataset that included all TBI patients (PNA =3.4%, UTI =1.8%; Table S1) and across all TBI subtypes (Table S2). Over the study period, the yearly incidence of pneumonia decreased from 1.6% to 1.1%. Analysis of prevalence by TBI subtype revealed that severe TBI patients, specifically, experienced higher rates of ARDS (Isolated Severe TBI =2.3%, All Severe TBI =4.1%) and cardiac arrest (Isolated Severe TBI =1.8%, All Severe TBI =4.0%; Table S2) compared to patients in the less severe TBI groups (Table S2). Interestingly, investigation of prevalence by facility demonstrated wide inter-hospital variability in complication rates, particularly for pneumonia (Range: Isolated TBI =0–28.5%; All TBI =0–50.0%), UTI (Range: Isolated TBI =0–25.0%; All TBI =0–19.4%) and MI (Range: Isolated TBI =0–33.3%; All TBI =0–12.5%; Figure S1).

Discharge disposition

Unadjusted comparison of changes in discharge disposition after the occurrence of each complication demonstrated dramatic shifts in the number of patients discharged home versus a care facility for isolated TBI patients (Figure 2A). Presence of a complication was also associated with a greater shift towards poor outcomes when present in the all TBI population compared to the isolated TBI population (Figure 2A,B). Subsequent controlled multivariate analysis in the isolated TBI dataset showed that developing pneumonia had a significant negative impact on discharge outcome [odds ratio (OR) =4.605, 95% CI, 3.858–5.496, P<0.001] and presented the greatest burden on discharge disposition (PAR =0.052; Table 2, Figure 3). DVT (OR =3.117, 2.411–4.029, P<0.001), UTI (OR =3.923, 3.352–4.591, P<0.001), and stroke (OR =6.288, 3.937–10.043) also decreased the likelihood that patients returned home, with burdens of 0.011, 0.037 and 0.015, respectively (Table 2, Figure 3). Sensitivity analysis of the impact of complications on discharge disposition in the population expanded to include non-isolated TBI supported these trends, with similar values for burden on discharge disposition for these complications (PNA =0.070, UTI =0.035, DVT =0.020, Stroke =0.011; Table S3, Figure 3).

Figure 2 Impact of each complication on the proportion of patients discharged home vs. care facility/rehabilitation center. Unadjusted secondary comparison of discharge disposition (home vs. care facility or rehab) based on the presence of each complication. (A) Study population = Isolated TBI. (B) Study population = All TBI (isolated and non-isolated). AKI, acute kidney injury; ARDS, acute respiratory distress syndrome; DVT, deep vein thrombosis; MI, myocardial infarction; SSI, surgical site infection; PNA, pneumonia; UTI, urinary tract infection.

Figure 3 Burden associated with each complication by TBI type. Adjusted PAR associated with the occurrence of a complication in the isolated TBI and all-TBI population. Isolated TBI, light gray bars. All TBI = dark gray bars. AKI, acute kidney injury; ARDS, acute respiratory distress syndrome; DVT, deep vein thrombosis; MI, myocardial infarction; SSI, surgical site infection; PNA, pneumonia; UTI, urinary tract infection.

Hospital LOS

Multivariate analysis of the burden of each complication on total hospital LOS also demonstrated highly significant relationships for isolated TBI patients. In particular, the occurrence of an organ space SSI (+18.9 days, P<0.001) or superficial SSI (+18.4 days, P<0.001; Figure 4) was correlated with a dramatic increase in the duration of in-hospital treatment. Additionally, SSIs were associated with the greatest LOS burden (Organ space SSI: 4.53, Superficial SSI: 4.42; Table 2). Developing pneumonia (+10.8 days), DVT (+7.8 days), and UTI (+7.7 days, P<0.001 for all complications; Figure 4) also resulted in significant extensions in LOS, with corresponding high values of burden (Table 2). Cardiac arrest decreased time spent in the hospital by 1.6 days (Figure 4), likely due to increased in-hospital mortality associated with this event. Secondary subtype analyses further supported these findings, demonstrating the same trends across TBI subtypes (Tables S4,S5, Figure S2).

Figure 4 Adjusted difference in LOS attributable to each complication. All results are reported for an interval of +1 patients after multivariate analysis controlling for demographic, clinical, and facility variables. Error bars show 95% confidence intervals. Differences in LOS for isolated TBI patients (light blue). Differences in LOS including non-isolated TBI (dark blue). AKI, acute kidney injury; ARDS, acute respiratory distress syndrome; DVT, deep vein thrombosis; MI, myocardial infarction; SSI, surgical site infection; PNA, pneumonia; UTI, urinary tract infection.

Discussion

In this study, we found that pneumonia, UTI, and DVT presented the greatest burden on isolated TBI outcomes (Table 2, Figure 3, Table S5, Figure S2A). SSIs also contributed significantly to total LOS (Figure 4, Table 2). Previous trauma research has established that complications can negatively impact patient outcomes (9,10). TBI-focused studies have also found that associated injuries and mechanism of injury can influence hospital course and mortality (12-15). Due to the paucity of information on the burden of complications for TBI patients specifically, the present study aimed to characterize the impact of 12 complications on TBI outcomes.

Pneumonia was associated with worse TBI outcomes

With some of the highest incidences of all investigated complications (Table 1), pneumonia presented the greatest burden for discharge disposition (Table 2, Figure 3). Unadjusted comparison of the impact of pneumonia on discharge disposition in the isolated vs. all TBI population showed that complications had a greater effect in the dataset that included all TBI patients (Figure 3). These results were supported by controlled analysis, which demonstrated greater odds ratios and PAR values for the inclusive dataset (Table S3, Figure 3). One possible explanation is that complications in the setting of multi-system injuries elicit a cascade effect where one complication leads to another. Alternatively, complications in the polytrauma context may more directly reflect the presence of injury to other organ systems. For example, pulmonary complications such as pneumonia have been shown to trigger hypoxic or hypercapnic respiratory failure in TBI patients, and the management of comorbid TBI and respiratory complications is difficult due to the competing effects of ventilator strategies on intracranial pressure (19,20). Given that pulmonary complications can increase the burden of initial TBI, preventative measures against pneumonia development are crucial, particularly in more severe TBI cases where diminished protective airway reflexes leave patients more susceptible to aspiration (19).

In the isolated TBI population, where polytrauma is unlikely to bias the onset of respiratory complications, the occurrence of pneumonia was associated with a change in LOS of 10.8 additional days (Figure 4) and high LOS burden (Table 2). These findings align with prior research, which showed that ventilator associated pneumonia was associated with longer hospital LOS (21). Furthermore, these effects were consistent in an identical sensitivity analysis that included non-isolated TBI patients (Figure 4, Table S3). Taken together, these data highlight pneumonia prevention as a potential critical node for improvement of TBI care. Analysis of prevalence by facility revealed great inter-institutional variability in pneumonia rates (Figure S1), suggesting this may be a modifiable factor related to local TBI management strategies.

UTI and DVT were correlated with worse TBI outcomes

Multivariate analysis also demonstrated that the development of a UTI or DVT during treatment were associated with poorer discharge disposition, longer LOS, and greater burden (Table 2, Figure 4). Studies investigating the connection between TBI and DVT have found that TBI is independently associated with the development of DVT, regardless of when prophylactic treatment is initiated (22). The safest and most effective timing for chemoprophylaxis after DVT also remains controversial, particularly for patients with intracranial hemorrhage (23). Given our present findings that DVT is associated with worse discharge disposition and longer LOS, further development of measures to mitigate the risk of DVT development may benefit TBI outcomes and decrease the cost associated with treatment of these patients.

Previous research has also highlighted TBI-induced immunosuppression as a potential contributor to high rates of post-trauma infectious complications, particularly the development of UTIs, pneumonia, and intracranial infections (24,25). The mechanisms underlying immunosuppression after TBI that provide a permissive environment for infectious complications remain an area of active research (24). Future efforts that aim to understand the susceptibility of the immune system to TBI-related dysregulation may inform preventative strategies for nosocomial infections.

Complications were associated with longer hospital LOS

Our analyses also demonstrated that with the exception of cardiac arrest, all complications were highly associated with an extended LOS (Table 2, Figure 4). In particular, organ space and superficial SSIs, contributed significantly to time spent in the hospital, increasing the LOS by as much as 450% (Table 2). Evidence from prior critical care research has demonstrated a significant cost associated with an increase in LOS (6-8). Therefore, identifying methods of reducing complication rates may be one strategy to lower costs for both the patient and institution, as well as decrease the burden on physicians and hospital staff. Interestingly, however, when secondary analyses were performed based on TBI subtype, only pneumonia, DVT, UTI, and superficial SSIs remained significant across severe, moderate and mild, as well as pediatric and adult isolated TBI categories (Figure S2A). Although these differences from the total study population results may be influenced by a smaller sample size, the results underscore the burden of these specific complications on TBI outcomes. Further research into the causes of the variation between groups is needed.

Analysis of complication prevalence by institution demonstrated wide inter-hospital variability across all complications, but particularly for pneumonia, UTI, and MI. Differences in hospital status (i.e., university vs. community) may explain these variations. however, because our analysis accounted for center specific variables such as teaching status and TBI volume, these disparities were unlikely to have significantly affected the present results. Regardless, this inter-institution variability in prevalence suggests an area for optimization of TBI outcomes. Another possible explanation for differing complication rates is hospital-specific strategies for managing TBI and related complications. Thus, standardization of treatment practices and systematic training may benefit TBI care through minimization of complications. Future research on the best standards for managing each complication in the context of TBI is necessary.

Limitations

Certain limitations of this study that may restrict the generalizability of the results. The NTDB contains data voluntarily submitted by hospitals and may not be a fully complete representation of true trauma burden and care. However, this database was the best resource for this study because it is the largest trauma registry in the United States. Data on pupillary response, ICP, and volume of intracranial hemorrhage for each patient was not available, so confounding due to differences in case severity may still exist. As with any database, the potential for recording inaccuracies is also present. However, the NTDB implements several quality assurance measures to reduce the frequency of these inaccuracies (26). Because only initial admission data are included in the NTDB, the short-term outcomes assessed here may not precisely reflect the ultimate functional outcomes of the investigated TBI subpopulations. Finally, the data presented here represent associations between complications and outcomes. Information on timing of complication onset was also not available. Future research will be necessary to identify causal effects and impacts of complication timing on outcomes.

Conclusions

Across TBI subtypes, pneumonia, UTIs, and DVT complications presented the greatest burden on isolated TBI care. The occurrence of SSIs also significantly impacted a patient’s likelihood of returning home and increased LOS. These trends suggest targets for optimization of TBI management.

Acknowledgments

An abstract for this manuscript was accepted to the American Association of Neurological Surgeons 2021 Annual Meeting.

Funding: None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://dx.doi.org/10.21037/jeccm-21-53

Data Sharing Statement: Available at https://dx.doi.org/10.21037/jeccm-21-53

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at https://dx.doi.org/10.21037/jeccm-21-53). Ms. ARK reports serving as a Director for Cress Health (unpaid) and owns shares of NeuroPace, outside the submitted work. Dr. WFA reports grants from Vivonics Inc., grants from Biogen Inc., outside the submitted work. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). This study was exempt from Institutional Board Review due to the publicly available and de-identified nature of the data in the NTDB. Ethical review and approval were not required for this study on human participants in accordance with the local legislation and institutional requirements.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Dewan MC, Rattani A, Gupta S, et al. Estimating the global incidence of traumatic brain injury. J Neurosurg 2018; Epub ahead of print. [Crossref] [PubMed]
2. Araki T, Yokota H, Morita A. Pediatric Traumatic Brain Injury: Characteristic Features, Diagnosis, and Management. Neurol Med Chir (Tokyo) 2017;57:82-93. [Crossref] [PubMed]
3. Faul M, Xu L, Wald M, et al. Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations and Deaths 2002-2006. Atlanta (GA): Center for Disease Control and Prevention, National Center for Injury Prevention and Control, 2010.
4. Schneier AJ, Shields BJ, Hostetler SG, et al. Incidence of pediatric traumatic brain injury and associated hospital resource utilization in the United States. Pediatrics 2006;118:483-92. [Crossref] [PubMed]
5. Rusnak M. Traumatic brain injury: Giving voice to a silent epidemic. Nat Rev Neurol 2013;9:186-7. [Crossref] [PubMed]
6. O'Keefe GE, Maier RV, Diehr P, et al. The complications of trauma and their associated costs in a level I trauma center. Arch Surg 1997;132:920-4; discussion 925. [Crossref] [PubMed]
7. de Jongh MA, Bosma E, Leenen LP, et al. Increased consumption of hospital resources due to complications: an assessment of costs in a level I trauma center. J Trauma 2011;71:E102-9. [Crossref] [PubMed]
8. Hemmila MR, Jakubus JL, Maggio PM, et al. Real money: complications and hospital costs in trauma patients. Surgery 2008;144:307-16. [Crossref] [PubMed]
9. Ingraham AM, Xiong W, Hemmila MR, et al. The attributable mortality and length of stay of trauma-related complications: a matched cohort study. Ann Surg 2010;252:358-62. [Crossref] [PubMed]
10. Shafi S, Barnes S, Nicewander D, et al. Health care reform at trauma centers--mortality, complications, and length of stay. J Trauma 2010;69:1367-71. [Crossref] [PubMed]
11. Moore L, Stelfox HT, Turgeon AF. Complication rates as a trauma care performance indicator: a systematic review. Crit Care 2012;16:R195. [Crossref] [PubMed]
12. Lawrence T, Helmy A, Bouamra O, et al. Traumatic brain injury in England and Wales: prospective audit of epidemiology, complications and standardised mortality. BMJ Open 2016;6:e012197 [Crossref] [PubMed]
13. Siegel JH. The effect of associated injuries, blood loss, and oxygen debt on death and disability in blunt traumatic brain injury: the need for early physiologic predictors of severity. J Neurotrauma 1995;12:579-90. [Crossref] [PubMed]
14. Wagner RH, Cifu DX, Keyser-Marcus L. Functional outcome of individuals with traumatic brain injury and lower extremity deep venous thrombosis. J Head Trauma Rehabil 1999;14:558-66. [Crossref] [PubMed]
15. Ho CH, Liang FW, Wang JJ, et al. Impact of grouping complications on mortality in traumatic brain injury: A nationwide population-based study. PLoS One 2018;13:e0190683 [Crossref] [PubMed]
16. American C. About NTDB. American College of Surgeons, Chicago, IL. 2019. Available online: https://www.facs.org/quality-programs/trauma/tqp/center-programs/ntdb/about
17. Watson J, Slaiby J, Garcia Toca M, et al. A 14-year experience with blunt thoracic aortic injury. J Vasc Surg 2013;58:380-5. [Crossref] [PubMed]
18. Kortbeek JB, Al Turki SA, Ali J, et al. Advanced trauma life support, 8th edition, the evidence for change. J Trauma 2008;64:1638-50.
19. Racca F, Vianello A, Mongini T, et al. Practical approach to respiratory emergencies in neurological diseases. Neurol Sci 2020;41:497-508. [Crossref] [PubMed]
20. Nyquist P, Stevens RD, Mirski MA. Neurologic injury and mechanical ventilation. Neurocrit Care 2008;9:400-8. [Crossref] [PubMed]
21. Li Y, Liu C, Xiao W, et al. Incidence, Risk Factors, and Outcomes of Ventilator-Associated Pneumonia in Traumatic Brain Injury: A Meta-analysis. Neurocrit Care 2020;32:272-85. [Crossref] [PubMed]
22. Reiff DA, Haricharan RN, Bullington NM, et al. Traumatic brain injury is associated with the development of deep vein thrombosis independent of pharmacological prophylaxis. J Trauma 2009;66:1436-40. [Crossref] [PubMed]
23. Abdel-Aziz H, Dunham CM, Malik RJ, et al. Timing for deep vein thrombosis chemoprophylaxis in traumatic brain injury: an evidence-based review. Crit Care 2015;19:96. [Crossref] [PubMed]
24. Sharma R, Shultz SR, Robinson MJ, et al. Infections after a traumatic brain injury: The complex interplay between the immune and neurological systems. Brain Behav Immun 2019;79:63-74. [Crossref] [PubMed]
25. Busl KM. Nosocomial Infections in the Neurointensive Care Unit. Neurol Clin 2017;35:785-807. [Crossref] [PubMed]
26. Newgard CD, Fildes JJ, Wu L, et al. Methodology and analytic rationale for the American College of Surgeons Trauma Quality Improvement Program. J Am Coll Surg 2013;216:147-57. [Crossref] [PubMed]