A criteria-directed protocol for in-hospital triage of trauma patients

Automated stratification of trauma injury severity across multiple body regions using multi-modal, multi-class machine learning models

OBJECTIVE

The timely stratification of trauma injury severity can enhance the quality of trauma care but it requires intense manual annotation from certified trauma coders. The objective of this study is to develop machine learning models for the stratification of trauma injury severity across various body regions using clinical text and structured electronic health records (EHRs) data.

MATERIALS AND METHODS

Our study utilized clinical documents and structured EHR variables linked with the trauma registry data to create 2 machine learning models with different approaches to representing text. The first one fuses concept unique identifiers (CUIs) extracted from free text with structured EHR variables, while the second one integrates free text with structured EHR variables. Temporal validation was undertaken to ensure the models' temporal generalizability. Additionally, analyses to assess the variable importance were conducted.

RESULTS

Both models demonstrated impressive performance in categorizing leg injuries, achieving high accuracy with macro-F1 scores of over 0.8. Additionally, they showed considerable accuracy, with macro-F1 scores exceeding or near 0.7, in assessing injuries in the areas of the chest and head. We showed in our variable importance analysis that the most important features in the model have strong face validity in determining clinically relevant trauma injuries.

DISCUSSION

The CUI-based model achieves comparable performance, if not higher, compared to the free-text-based model, with reduced complexity. Furthermore, integrating structured EHR data improves performance, particularly when the text modalities are insufficiently indicative.

CONCLUSIONS

Our multi-modal, multiclass models can provide accurate stratification of trauma injury severity and clinically relevant interpretations.

Association between trauma triage and time-to-vaso-occlusive events in patients with sickle cell disease after traumatic injury: a retrospective study

Introduction

Sickle cell disease (SCD) is associated with vaso-occlusive events (VOEs) that can lead to disease complications, including early mortality. Given that similar inflammatory responses characterize VOE and traumatic injury, injured patients with SCD may be vulnerable to acute complications. This study is the first to examine whether traumatic injury is associated with increased severity of future VOEs.

Methods

This cohort study was conducted using electronic health record data from an SCD clinic in Western Pennsylvania; 356 patients with SCD from January 2000 to July 2021 were identified via retrospective chart review. 55 patients were eligible based on continuous medical record data spanning 1 year preinjury and postinjury. Patients were sorted into three treatment groups based on injury management: (1) Neither triage to trauma team activation (TTA) nor inpatient admission (Early Discharge), (2) Triage but no inpatient admission (Triage Only), and (3) Triage and In-patient. Outcomes included time from injury to first VOE, annual VOE counts requiring an emergency department (ED) visit, and ED length of stay (LOS) for the first VOE after injury.

Results

Early Discharge individuals experienced a VOE event within 2.93 days of injury, significantly shorter time to event than Triage and In-patient individuals at 52.375 days and Triage Only individuals at 100.16 days (p=0.0058). No difference in annual VOE counts was noted postinjury across all groups. However, a significant increase in VOE LOS preinjury (16.1 hours) to postinjury (77.4 hours) was noted only for the Triage Only group (p=0.038). Cox regression model showed that shortened time to VOE events was marginally associated with TTA status (p=0.06).

Conclusion

Despite minimal changes in long-term VOE outcomes after injury, traumatic injuries may accelerate the time-to-VOE among the Early Discharge group. Therefore, future research is warranted to analyze whether the absence of postinjury triage assessment and intervention may cause unforeseen physiologic stressors contributing to VOE outcomes.

Level of evidence

Level IV: retrospective case-control study with three negative criteria.

Implementation of a major trauma team. Analysis of activation and care times in patients admitted to the ICU

OBJECTIVE

To analyze the factors associated with the activation of the severe trauma care team (STAT) in patients admitted to the ICU, to measure its impact on care times, and to analyze the groups of patients according to activation and level of anatomical involvement.

DESIGN

Prospective cohort study of severe trauma admitted to the ICU. From June 2017 to May 2019. Risk factors for the activation of the STAT analysed with logistic regression and CART type classification tree.

SETTING

Second level hospital ICU.

PATIENTS

Patients admitted consecutively.

INTERVENTIONS

No.

MAIN VARIABLES OF INTEREST

STAT activation. Demographic variables. Injury severity (ISS), intentionality, mechanism, assistance times, evolutionary complications, and mortality.

RESULTS

A total of 188 patients were admitted (46.8% of STAT activation), median age of 52 (37-64) years (activated 47 (27-62) vs not activated 55 (42-67) P = 0.023), males 84.0%. No difference in mortality according to activation. The logistic model finds as factors: care (16.6 (2.1-13.2)) and prehospital intubation (4.2 (1.8-9.8)) and severe lower extremity injury (4.4 (1.6-12.3)). Accidental fall (0.2 (0.1-0.6)) makes activation less likely. The CART model selects the type of trauma mechanism and can separate high and low energy trauma.

CONCLUSIONS

Factors associated with STAT activation were prehospital care, requiring prior intubation, high-energy mechanisms, and severe lower extremity injuries. Shorter care times if activated without influencing mortality. We must improve activation in older patients with low-energy trauma and without prehospital care.

Emergency Severity Index (ESI) algorithm in trauma patients: The impact of age during the pandemic

OBJECTIVE

The ESI algorithm is widely used to triage patients in the emergency room. However, few studies have assessed the reliability of ESI to accurately triage trauma patients. The aim of this study was to compare vital signs, resource utilization, and patient outcomes among trauma patients during the pandemic in 2020 vs. the previous year prior to the pandemic.

METHODS

This retrospective study was conducted over a 24-month period at an urban adult level one trauma center. Demographic and clinical characteristics, resource utilization, and patient outcomes were extracted from the electronic medical records and trauma registry. Trauma patients assigned ESI level 2 were stratified by age (<65 years and ≥ 65 years) and year (2019 vs. 2020) for data analysis.

RESULTS

A total of 3,788 trauma patients were included in the study. Males represented 68.4% (2,591) of patients and the median age was 50 years (IQR: 31, 69). The majority of patients were assigned ESI level 2 (2,162, 57.1%) and had a blunt mechanism of injury (3,122, 82.4%). In 2020, patients <65 years of age utilized less resources compared to 2019 (p < 0.001). Likewise, patients >65 years of age required less lab tests [OR: 0.1, 95% CI: (0.05 - 0.4)], IV fluids [OR: 0.2, 95% CI: (0.2 -0.3)], IV medications [OR: 0.6, 95% CI: (0.4 - 0.7)], and specialty consultations [OR: 0.4, 95% CI: (0.3 -0.5)] compared to 2019 (p < 0.0001). Within 2020, vital signs and resources utilized between younger and elderly patients varied significantly (p < 0.01). Correspondingly, the clinical outcomes between younger and elderly patients within 2020, differed significantly (p < 0.01).

CONCLUSIONS

The COVID-19 pandemic affected the triage of trauma patients. During 2020, patients utilized less resources compared to the previous year. Additionally, younger and elderly patients had different vital signs, resource utilization, and clinical outcomes although both being assigned ESI level 2. Younger trauma patients may have been over-triaged in 2020 due to the COVID-19 pandemic.

Validation of the Swedish Trauma Registry (SweTrau)

PURPOSE

Validation of registries is important to ensure accuracy of data and registry-based research. This is often done by comparisons of the original registry data with other sources, e.g. another registry or a re-registration of data. Founded in 2011, the Swedish Trauma Registry (SweTrau) consists of variables based on international consensus (the Utstein Template of Trauma). This project aimed to perform the first validation of SweTrau.

METHODS

On-site re-registration was performed on randomly selected trauma patients and compared to the registration in SweTrau. Accuracy (exact agreement), correctness (exact agreement plus data within acceptable range), comparability (similarity with other registries), data completeness (1-missing data) and case completeness (1-missing cases) were deemed as either good ([Formula: see text] 85%), adequate (70-84%) or poor (< 70%). Correlation was determined as either excellent ([Formula: see text] 0.8), strong (0.6-0.79), moderate (0.4-0.59) or weak (< 0.4).

RESULTS

The data in SweTrau had good accuracy (85.8%), correctness (89.7%) and data completeness (88.5%), as well as strong or excellent correlation (87.5%). Case completeness was 44.3%, however, for NISS > 15 case completeness was 100%. Median time to registration was 4.5 months, with 84.2% registered one year after the trauma. The comparability showed an accordance with the Utstein Template of Trauma of almost 90%.

CONCLUSIONS

The validity of SweTrau is good, with high accuracy, correctness, data completeness and correlation. The data are comparable to other trauma registries using the Utstein Template of Trauma; however, timeliness and case completeness are areas of improvement.

Characteristics and outcome of traumatic cardiac arrest at a level 1 trauma centre over 10 years in Sweden

Background

Historically, resuscitation in traumatic cardiac arrest (TCA) has been deemed futile. However, recent literature reports improved but varying survival. Current European guidelines emphasise the addressing of reversible aetiologies in TCA and propose that a resuscitative thoracotomy may be performed within 15 min from last sign of life. To improve clinician understanding of which patients benefit from resuscitative efforts we aimed to describe the characteristics and 30-day survival for traumatic cardiac arrest at a Swedish trauma centre with a particular focus on resuscitative thoracotomy.

Methods

Retrospective cohort study of adult patients (≥ 15 years) with TCA managed at Karolinska University Hospital Solna between 2011 and 2020. Trauma demographics, intra-arrest factors, lab values and procedures were compared between survivors and non-survivors.

Results

Among the 284 included patients the median age was 38 years, 82.2% were male and 60.5% were previously healthy. Blunt trauma was the dominant injury in 64.8% and median Injury Severity Score (ISS) was 38. For patients with a documented arrest rhythm, asystole was recorded in 39.2%, pulseless electric activity in 24.8% and a shockable rhythm in 6.8%. Thirty patients (10.6%) survived to 30 days with a Glasgow Outcome Scale score of 3 (n = 23) or 4 (n = 7). The most common causes of death were haemorrhagic shock (50.0%) and traumatic brain injury (25.5%). Survivors had a lower ISS (P < 0.001), more often had reactive pupils (P < 0.001) and a shockable rhythm (P = 0.04). In the subset of prehospital TCA, survivors less frequently received adrenaline (epinephrine) (P < 0.001) and in lower amounts (P = 0.02). Of patients that underwent resuscitative thoracotomy (n = 101), survivors (n = 12) had a shorter median time from last sign of life to thoracotomy (P = 0.03), however in four of these survivors the time exceeded 15 min.

Conclusion

Survival after TCA is possible. Determining futility in TCA is difficult and this study demonstrates survivors outside of recent guidelines.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13049-022-01039-9.

Trauma triage criteria as predictors of severe injury - a Swedish multicenter cohort study

BACKGROUND

Adequate performance of trauma team activation (TTA) criteria is important in order to accurately triage trauma patients. The Swedish National Trauma Triage Criteria (SNTTC) consists of 29 criteria that trigger either a Trauma Alert, the highest level of TTA, or a Trauma Response. This study aimed to evaluate the SNTTC and its accuracy in predicting a severely injured patient in a multicenter setting.

METHODS

A cohort study in Sweden involving six trauma receiving hospitals. Data was collected from the Swedish Trauma Registry. Some 626 patients were analyzed with regard to the specific criteria used to initiate the TTA, injury severity with New Injury Severity Score (NISS) and emergency interventions. Sensitivity, specificity, positive predictive value (PPV) and positive likelihood ratio (LR+) of the criteria were calculated, as well as undertriage and overtriage.

RESULTS

All 29 criteria of SNTTC had a sensitivity > 80% for identifying a severely injured patient. The 16 Trauma Alert Criteria had a lower sensitivity of 62.6% but higher LR+ (3.5 vs all criteria 1.4), specificity (82.3 vs 39.1%) and PPV (55.4 vs 37.6%) and the highest accuracy (AUC 0.724). When using only the six physiological criteria, sensitivity (44.8%) and accuracy (AUC 0.690) decreased while LR+ (6.7), specificity (93.3%) and PPV (70.2%) improved.

CONCLUSION

SNTTC is efficient in identifying severely injured patients. The current set of criteria exhibits the best sensitivity compared to other examined combinations and no additional criterion was found to improve the protocol enough to promote a change.

Telemedicine communication using instant messaging technology improves pre-hospital triage in high mountain train disasters

Background:

Differences between pre-hospital triage by an emergency medical technician and Simple Triage and Rapid Treatment triage (START) by emergency staffs often affect manpower management and aggravate the chaos condition of emergency room.

Objectives:

Under the assistance of instant messaging, the authors aimed to identify ways of improving triage differences between emergency medical technician triage grading and Simple Triage and Rapid Treatment triage grading by emergency staffs.

Methods:

Recorded photographs of all patients were reviewed by a smartphone. We categorized patients according to three triage conditions: group 1, accident scene on-site or instantaneous Simple Triage and Rapid Treatment triage by the emergency medical technician; group 2, triage under Simple Triage and Rapid Treatment grading by emergency staffs; group 3, re-triage with START grading using recorded photographs, Glasgow Coma Scale, and vital signs when these patients were arrived in emergency room. The Wilcoxon Signed-Rank test, Spearman rank correlations, and Kruskal–Wallis test are employed to test differences among the groups. We used risk estimates with odds ratios and the chi-square test to statistically analyze the differences in triage grading.

Results:

Statistical analysis found conflicting results among Wilcoxon Signed-Rank test, Spearman rank correlations, and Kruskal–Wallis test. The difference in triage grading between groups 2 and 1 was greater than that between groups 2 and 3 (odds ratio, 6.473; 95% confidence interval, 1.693–24.470; p-value < 0.05).

Conclusion:

Transferred photographs combined with Glasgow Coma Scale and vital signs can help us to understand the real situations of patients. With instant messaging applications, it is possible to make more precise pre-hospital or instantaneous triage.

Predictive value of the shock index (SI) compared to the age-adjusted pediatric shock index (SIPA) for identifying children that needed the highest-level trauma activation based on the presence of consensus criteria

OBJECTIVES

In previous studies, SIPA was shown to be better than the SI in identifying children who have an elevated ISS, required transfusion, or were at a high risk of death. No comparison has been made to the consensus-based criteria that identify patients requiring the highest-level trauma activation. The objective of this study was to determine if the SIPA was more accurate than the SI in identifying children with increased need for trauma team activation as defined by the criterion standard definition, and secondly the sensitivity and specificity of the SI and SIPA.

METHODS

Retrospective review of prospectively collected trauma based data. Children aged 1-17 years admitted to a pediatric level 1 trauma center between 1/1/16 and 12/31/17 and met the prehospital criteria for level 1 or 2 trauma activation were included. We evaluated the ability of SI > 0.9 at ED presentation and elevated SIPA to predict need for trauma activation based on consensus criteria. SIPA cutoffs were > 1.22 (age 4-6), >1.0 (age 7-12), and > 0.9 (age 13-17).

RESULTS

Among 3378 children, 1486 (44%) had an elevated SI and 590 (18%) had an elevated SIPA. There were 160 (5%) patients who met at least one consensus criterion. Broadly, sensitivity and specificity analyses reveal poor sensitivity for both SI and SIPA (59.4% versus 43.1% respectively) measures but a moderate specificity for SIPA (83.8%). Both SI and SIPA have a poor PPV (6.4% versus 11.7%) but high NPV (96.6% versus 96.7%). Overall, SIPA has higher accuracy than SI in predicting consensus criteria 82% versus 57%).

CONCLUSION

SIPA is more accurate than the SI in identifying children who meet a consensus criterion defining the need for highest-level trauma activation. The low PPV and sensitivity suggest that SIPA alone, while somewhat less likely to lead to overtriage than SI is not ideal for ruling in the need for level one resources as defined by the consensus criteria. Prognosis study, retrospective.

LEVEL OF EVIDENCE

Level II.

The transfer of clinical prediction models for early trauma care had uncertain effects on mistriage

OBJECTIVES

This study aimed to assess how transfers of clinical prediction models for early trauma care between different care contexts within a single health system affected mistriage rates.

STUDY DESIGN AND SETTING

Patients aged 15 years or older, registered between 2011 and 2016 in the Swedish national trauma registry, SweTrau, were included. Three data set groups were created: high- and low-volume centers, metropolitan and nonmetropolitan centers, and multicenters and single centers. Clinical prediction models were developed using logistic regression in each data set group and transferred between data sets within groups. Model performance was evaluated using mistriage rate, undertriage rate, and overtriage rate. Multiple imputation using chained equations was used to handle missing data. Model performance was reported as medians with 95% confidence intervals (CIs).

RESULTS

A total of 26,965 patients were included. Changes in mistriage rates after transfer ranged from -0.25 (95% CI -0.21 to 0.04) to 0.29 (95% CI 0.13-0.39). Both overtriage and undertriage rates were affected.

CONCLUSIONS

Transferring clinical prediction models for early trauma care is associated with substantial uncertainty in regards to the effect on model performance. Depending on the care context, model transfer led to either increased or decreased mistriage. Overtriage was more affected by model transfer than undertriage.

ISS alone, is not sufficient to correctly assign patients post hoc to trauma team requirement

Purpose

An injury severity score (ISS) ≥ 16 alone, is commonly used post hoc to define the correct activation of a trauma team. However, abnormal vital functions and the requirement of life-saving procedures may also have a role in defining trauma team requirement post hoc. The aim of this study was to describe their prevalence and mortality in severely injured patients and to estimate their potential additional value in the definition of trauma team requirement as compared to the definition based on ISS alone.

Methods

Retrospective analysis of a trauma registry including patients with trauma team activation from the years 2009 until 2015, who were 16 years of age or older and were brought to the trauma center directly from the scene. Patients were divided into a group with an ISS ≥ 16 vs. ISS < 16. For analysis a predefined list of abnormal vital functions and life-saving interventions was used.

Results

58,723 patients were included in the study (N = 32,653 with ISS ≥ 16; N = 26,070 with ISS < 16). From the total number of patients that required life-saving procedures or presented with abnormal vital functions 29.1% were found in the ISS < 16 group. From the ISS < 16 group, 36.7% of patients required life-saving procedures or presented with abnormal vital signs. The mortality of those was 8.1%.

Conclusions

Defining the true requirement of trauma team activation post hoc by using ISS ≥ 16 alone does miss a considerable number of subjects who require life-saving interventions or present with abnormal vital functions. Therefore, life-saving interventions and abnormal vital functions should be included in the definitions for trauma team requirement. Further studies have to evaluate, which life-saving procedures and abnormal vital functions are most relevant.

Efforts to Improve Appropriate Trauma Activation for Patients Transferred to a Level 1 Trauma Center From Outside Hospitals

Trauma centers monitor under- and overtriage rates to comply with American College of Surgeons Committee on Trauma verification requirements. Efforts to maintain acceptable rates are often undertaken as part of quality assurance. The purpose of this project was to improve the institutional undertriage rate by focusing on appropriately triaging patients transferred from outside hospitals (OSHs). Trauma physicians received education and pocket cards outlining injury severity score (ISS) calculation to aid in prospectively estimating ISS for patients transferred from OSHs, and activate the trauma response expected for that score. Under- and overtriage rates before and after the intervention were compared. The postintervention period saw a significant decrease in overall overtriage rate, with simultaneous trend toward lower overall undertriage rate, attributable to the significant reduction in undertriage rate of patients transferred from OSHs. Prospectively estimating ISS to assist in determining trauma activation level shows promise in managing appropriate patient triage. However, questions arose regarding the necessity for full trauma activation for transferred patients, regardless of ISS. It may be necessary to reconsider how patients transferred from OSHs are evaluated. Full trauma activation can be a financial and resource burden, and should not be taken lightly.

Impact of gender on post- traumatic intensive care and outcomes

Background

Several reports indicate gender disparities in health care provision. There is a well-documented male patient dominance in intensive care unit (ICU) admittance. It is not established if this difference reflects medical needs or is influenced by other factors. The aim of the current study was to investigate if patient gender influences the pattern of ICU admittance in a cohort of trauma patients.

Methods

Data from patients admitted to an urban trauma centre over a 10-year interval were linked to regional and national health registries to obtain data on demographics, co-comorbidities, trauma-related variables, ICU-admittance patterns and mortality. The association between gender and ICU-admission were explored using logistic regression analysis. The association between gender and short- and long-term mortality were explored using Cox regression models.

Results

In this study cohort of approximately 14,000 trauma patients, men had a higher probability of being admitted to the ICU after initial trauma resuscitation. The difference was limited to patients with less severe injuries (ISS < 15). No differences were noted in short-term survival, whereas men had a higher long-term mortality.

Conclusions

In this retrospective cohort study we found a difference between men and women in post trauma ICU admittance patterns, restricted to less injured patients, where men had a higher probability of ICU admittance. Whether this is a true gender bias or an effect of other factors not analysed in this study remains unknown. This finding warrants further studies.

Experiences of suffering multiple trauma: A qualitative study

OBJECTIVES

In an effort to strengthen health care professionals' ability to anticipate and address multiple trauma patients' needs, this study aims to explore the experience of suffering from multiple trauma.

DESIGN

This is a qualitative descriptive study. Nine interviews were analysed using content analysis.

SETTING

The study included patients who had been registered in the Swedish Intensive Care registry [SIR] due to suffering multiple trauma.

FINDINGS

The analysis revealed one theme, A detour in life, based on three sub-themes: (a) Feeling lost and not knowing what to expect, (b) Striving to get life back on track and (c) Dealing with 'dead ends' during rehabilitation. The theme showed that those who suffered multiple trauma did not know what to expect of their recovery and they expressed experiencing a lack of understanding and guidance from healthcare professionals. As it was important to focus on the present and find ways to move on in life, they sought for other ways to find direction in matters of rehabilitation and care.

CONCLUSIONS

A shared understanding is essential in order to define a person's needs. By setting short-term goals and improving documentation, healthcare professionals across the trauma recovery continuum could more easily gain insight of their patients' needs and address them with supportive guidance.

A Consensus-Based Criterion Standard for the Requirement of a Trauma Team

BACKGROUND

Trauma team activation (TTA) represents a considerable expenditure of trauma centre resources. It is mainly triggered by field triage criteria. The overall quality of the criteria may be evaluated based on the rate of over- and undertriage. However, there is no gold standard that defines which adult patients truly require a trauma team. The objective of this study was to develop consensus-based criteria defining the necessity for a trauma team.

METHODS

A consensus group was formed by trauma specialists experienced in emergency and trauma care with a specific interest in field triage and having previously participated in guideline development. A literature search was conducted to identify criteria that have already been used or suggested. The initial list of criteria was discussed in two Delphi round and two consensus conferences. The entire process of discussion and voting was highly standardized and extensively documented, resulting in a final list of criteria.

RESULTS

Initially 95 criteria were identified. This was subsequently reduced to 20 final criteria to appropriately indicate the requirement for attendance of a trauma team. The criteria address aspects related to injury severity, admission to an intensive care unit, death within 24 h, need for specified invasive procedures, need for surgical and/or interventional radiological procedures, and abnormal vital signs within a defined time period.

CONCLUSIONS

The selected criteria may be applied as a tool for research and quality control concerning TTA. However, future studies are necessary to further evaluate for possible redundancy in criteria that may allow for further reduction in criteria.