BACKGROUND

For patients injured in the out-of-hospital setting, the emergency medical services (EMS) decision on where to transport for definitive care can have lasting consequences. Survival from critical trauma improves when care is provided in a designated trauma center (Haas et al., 2010; Mackersie, 2006). Trauma system-level quality is optimized when noncritical trauma patients are primarily cared for in the most appropriate location of care, which is often a local hospital (MacKenzie et al., 2006; Mackersie, 2006; Sasser et al., 2009). Dating back to 1986, a variety of field triage decision guides have been developed and revised to help determine the most appropriate destination for patients who experience traumatic injury in the out-of-hospital setting (Lerner, Cushman, et al., 2017; McCoy et al., 2013; Sasser et al., 2009). The current 2011 version of Centers for Disease Control and Prevention (CDC) Guidelines for Field Triage of Injured Patients includes four distinct and sequential steps (Sasser et al., 2012):

Step 1: Vital signs and level of consciousness (vital signs & LOC)

Step 2: Anatomy of injury (AOI)

Step 3: Mechanism of injury (MOI) criteria

Step 4: Special considerations.

Importance

The CDC Guidelines for Field Triage of Injured Patients (CDC Screen) drive the destination decision for millions of EMS-transported trauma patients annually (Sasser et al., 2009), yet limited information exists regarding performance and relationship with patient outcomes as a whole. The individual steps have been most heavily studied, suggesting relatively satisfactory performance of the vital signs & LOC criteria (Lerner, Drendel, et al., 2017; Pearson et al., 2012; Warwick et al., 2021). Substitution of the Glasgow Coma Scale (GCS) motor score of 6 or less for the full GCS has been reported to increase simplicity with no loss of predictive fidelity (Kupas et al., 2016), although it has been reported to reduce sensitivity (Brown et al., 2014), especially for older patients (Deeb et al., 2021b).

Overall, assessment of the AOI criteria has shown acceptable predictive value (Isenberg et al., 2011; Lerner et al., 2013). More controversy surrounds MOI criteria, with a growing body of data suggesting that MOI alone is not a reliable predictor of the need for higher-level care (Isenberg et al., 2011; Lerner et al., 2021). And finally, debate also surrounds the elements of the special considerations step (Doumouras et al., 2012); several recent studies have suggested that patient age (Nakamura et al., 2012) and "provider judgment" may represent extremely valuable indicators of the need for trauma center care (Newgard et al., 2012).

A cut point of Injury Severity Scores greater than or equal to 16 has been utilized as a proxy for whether a patient was correctly transported to an American College of Surgeons (ACS) Level I or Level II trauma center (Newgard et al., 2016; Newgard et al., 2017). Limited information exists related to the accuracy of the CDC Screen in predicting patient outcomes, including hospital admission or mortality on a large scale.

Objective

The purpose of this study was to evaluate the relationship between the 2011 CDC Guidelines for Field Triage of Injured Patients and the odds of hospitalization and mortality using a large national EMS dataset.

KEY POINTS

* The current state of out-of-hospital trauma triage relies on Centers for Disease Control and Prevention (CDC) Guidelines for Field Triage of Injured Patients; however, limited information exists regarding performance and its relationship with patient outcomes.

* Patients meeting multiple CDC Guidelines Field Triage of Injured Patients steps were at greater risk of hospitalization and death.

* Vital signs and level of consciousness were associated with the greatest risk of death among those only meeting one step.

* Mechanism of Injury alone appears to have limited predictive power.

METHODS

Study Design and Setting

We retrospectively analyzed linked prehospital and hospital data from January 1, 2019, to December 31, 2019, from the ESO Data Collaborative (Austin, TX) public-use research dataset. ESO is a large EMS electronic health record provider, and the ESO Data Collaborative consists of participating EMS agencies who permit research using their de-identified records. The 2019 research dataset for this investigation contained 8,340,148 EMS encounters from 1,322 EMS agencies throughout the United States who agreed to share their de-identified data for research purposes. A subset of these agencies participated in a bidirectional health data exchange, which incorporated emergency department (ED) and hospital outcome data with the prehospital patient care record.

For this study, all 911 responses for transported patients with a complete prehospital CDC 2011 Guidelines for Field Triage of Injured Patients screen and linked ED dispositions were included. Pediatric patients (<18 years of age) and patients who had cardiac arrests prior to EMS arrival were excluded. The institutional review board at St. David's HealthCare determined that this study was exempt (1632498-1).

Measurements

The primary outcomes were hospital admission and overall mortality. The exposure of interest was the CDC Screen findings. The CDC Guidelines for Field Triage of Injured Patients criteria are grouped into four steps-Step 1: vital signs & LOC, Step 2: AOI, Step 3: MOI, or Step 4: special considerations. Variables were created for patients meeting criteria in none of the steps, a single step, or multiple steps. A secondary descriptive analysis of those meeting special considerations criteria was also performed, describing the patients that met each consideration.

We evaluated patient demographics, including age, sex (male or female), race and ethnicity, and the community size where the EMS encounter took place. A Rapid Emergency Medicine Score (REMS) was calculated using the first set of documented prehospital vital signs for each patient. The REMS is calculated by assigning point values to categories of mean arterial pressure, pulse rate, respiratory rate, oxygen saturation, GCS, and patient age. The REMS values range from 0 to 26, and higher values are associated with an increased probability of hospitalization and mortality among EMS patients (Alter et al., 2017; Crowe et al., 2022).

Outcome Measures

Hospitalization

The hospitalization outcome compared patients admitted or transferred to another hospital to those discharged alive from the ED, including those who left against medical advice or who left without being seen, versus those who were not discharged from the ED. The evaluation of hospitalization excluded patients who died in the ED.

Mortality

The mortality outcome dichotomized patients who were discharged alive versus patients who died in the ED or inhospital setting. Patients who were transferred to another facility and those who did not have inpatient disposition data available at the end of the study period were excluded from this measure.

Analysis

Descriptive statistics were used to describe baseline characteristics. Continuous variables were presented as medians and interquartile ranges (IQR), and categorical variables were presented as counts and percentages. The statistical significance threshold was set at .05. Univariable odds ratios were calculated to estimate the measure of association between independent variables and the outcome of interest. All analyses were performed using Stata v15.1 (College Station, TX).

RESULTS

Patient Demographics and Event Characteristics

During the study period, n = 86,462 patient records from n = 197 EMS agencies met inclusion criteria (Figure 1). The characteristics of the analysis population are described in Table 1.

Figure 1. Inclusion of patients and outcomes.

Table 1. Analysis Population Characteristics (

CDC Guidelines for Field Triage of Injured Patients Criteria

There were n = 65,967 (76.3%) patients who did not meet criteria in any step, n = 16,443 (19.0%) were positive in a single step, and n = 4,052 (4.7%) were positive in multiple steps (Table 2). Among those meeting a single step, n = 1,571 (9.6%) were positive in Step 1: vital signs & LOC, n = 1,030 (6.3%) in Step: 2 AOI, n = 993 (6.0%) in Step 3: MOI, and n = 12,849 (78.1%) in Step 4: special considerations (Table 3). Patients 55 years and older accounted for over half of those meeting Step 4, and those in this age group overwhelmingly had their race and ethnicity documented as White, not Hispanic, or Latino(a) (see Supplemental Digital Content Table 1, available at: http://links.lww.com/JTN/A80).

Table 2. Number of CDC Trauma Triage Criteria Met

Table 3. Specific Criteria Met

Patient Outcomes

There were n = 64,954 (75.1%) of patients who were discharged from the ED, and n = 19,576 (22.7%) were admitted (Table 1). When compared with injured patients who did not meet any criteria, the odds of hospitalization increased threefold for those in Step 1: vital signs & LOC (odds ratio [OR]: 3.07, 95% confidence interval [CI]: 2.77-3.40), fourfold for those with Step 2: AOI (OR: 3.94, 95% CI: 3.48-4.46), and twofold for those with Step 3: MOI (OR: 2.00, 95% CI: 1.74-2.29), or Step 4: special considerations (OR: 2.46, 95% CI: 2.36-2.56). Odds of hospitalization increased ninefold for those positive in multiple steps (OR: 8.97, 95% CI: 8.37-9.62) (see Supplemental Digital Content Figure 1, available at: http://links.lww.com/JTN/A81).

Overall, mortality data were available for n = 84,473 (97.7%) of patients. Among those, n = 886 (1.0%) died (Table 1). When compared with injured patients who met no CDC Guidelines criteria, the odds of mortality increased 10-fold for those positive in Step 1: vital signs & LOC (OR: 9.58, 95% CI: 7.30-12.56) and twofold for those in Step 2: AOI (OR: 2.34, 95% CI: 1.28-4.29), or Step 4: special considerations steps (OR: 2.10, 95% CI: 1.71-2.60). No difference in odds of mortality was found for those only positive in Step 3: MOI (OR: 0.22, 95% CI: 0.03-1.54). When compared with injured patients who met no CDC criteria, the odds of mortality were increased 23-fold for those positive in multiple steps (OR: 22.7, 95% CI: 19.7-26.8) (see Supplemental Digital Content Figure 2, available at: http://links.lww.com/JTN/A82).

DISCUSSION

In this study of over 86,000 EMS encounters for injured patients, screening positive in one or more steps of the 2011 CDC Trauma Triage Criteria, were associated with increased odds of hospitalization and mortality, although the predictive power of each individual step was not the same. Meeting either the AOI or vital signs & LOC was more predictive of hospitalization, while meeting the vital signs & LOC was more predictive of mortality. Meeting multiple criteria was highly predictive of both hospitalization and mortality.

Positive vital signs & LOC criteria (Step 1) was associated with a significantly higher rate of mortality as well as the second highest odds of hospital admission. Comparable odds ratios have been reported for mortality in the ED (OR: 15.1) and hospital (OR: 2.4) for patients who meet the vital signs & LOC criteria (Warwick et al., 2021). Similar findings have been reported for patients with traumatic brain injury (Pearson et al., 2012). Some authors have questioned whether all elements of Step 1 (respiratory rate, systolic blood pressure, and GCS) are necessary (Brown et al., 2014; Newgard et al., 2010). A large, multisite retrospective study concluded that eliminating any element would reduce the tool's sensitivity and fail to screen some high-risk patients who required high-level trauma care (Newgard et al., 2010). Proposed revisions to the 2011 Trauma Triage Guidelines have replaced a GCS of less than 14 with the motor GCS score of less than 6 (Newgard et al., 2022).

Positive AOI criteria (Step 2) were found to have the highest odds of admission and the second greatest odds for mortality. These findings compare favorably with other studies that reported on the relationship between Step 2 and Trauma Center Need, a research composite score comprised of an Injury Severity Score greater than 15, intensive care unit admission 24 hours or longer, need for urgent surgery based on ED disposition directly to surgery, or ED mortality (Deeb et al., 2021a; Lerner et al., 2013; Willenbring et al., 2016). Although several authors have voiced concern that prehospital assessment cannot accurately identify some of the anatomic conditions included in this step, data from this and other studies appear to validate the predictive strength of this step (Lerner et al., 2013).

It has been shown that although International Classification of Diseases-Tenth Revision (ICD-10) identified nearly double the number of patients who met the screening criteria, EMS assessment had higher sensitivity for the actual need for trauma center services (91% vs. 73%), higher positive predictive value (65 vs. 53%), and a lower false-positive rate. Unfortunately, EMS assessment also had a higher false-negative rate (69% vs. 42%). This false-negative rate may offer an opportunity to improve exact AOI criteria (Deeb et al., 2021a). Proposed revisions from the American College of Surgery have altered anatomic language from diagnostic to "suspected," as well as describing chest injuries and severe bleeding in terms more aligned with EMS assessments (Newgard et al., 2022).

Positive MOI criteria (Step 3) increased the odds of hospital admission, but no statistically significant difference in the odds of mortality was found for patients who were only positive on Step 3 compared to those not meeting criteria in any step. Although originally heralded as an accurate stand-alone EMS triage criteria (Lowe et al., 1986), other authors have reported on the inability of EMS providers to accurately assess all elements of MOI (Holst et al., 2016; Lerner et al., 2011). Further, research suggests that MOI does not accurately predict the need for trauma center care for adults (Isenberg et al., 2011) or children (Lerner et al., 2021), although death of another occupant, fall distance, and extrication time appear more predictive than other MOI elements (Lerner et al., 2011).

As a result, some systems have removed trauma center activation for patients who are only positive in Step 3: MOI (Stuke et al., 2013), and proposed revisions in the trauma triage guidelines have added the need for prolonged extrication to auto crash criteria (Newgard et al., 2022). The fact that 99.9% of MOI only patients in this study survived (only one patient death) supports proposed changes to the trauma triage criteria that recommend transport to a locally available (not necessarily highest level) trauma center for patients whose only risk is MOI (Newgard et al., 2022). This may have significant impact on patients and EMS providers in rural areas.

Special considerations (Step 4) also demonstrated relationships with hospital admission and mortality comparable to other studies. As previously described (Newgard et al., 2012), Step 4 was positive in a large majority (78%) of the patients who met CDC Guidelines criteria in this study. Age 55 years and more is one of the elements of Step 4 and has been described (Nakamura et al., 2012) as a driver of the high utilization of the special considerations step. While this age group comprised 53% of the study sample, they made up 85% of those who were positive for Step 4. However, this result must be tempered by the fact that 70% of study patients 55 years and older were not reported positive for any step.

It appears that age may not have been the sole contributor to a positive Step 4 score. Multiple authors have described the undertriage of older trauma patients (Doumouras et al., 2012; Morris et al., 2021; Newgard et al., 2016) and a failure to utilize the age criteria in Step 4 may contribute to this challenge. It has been shown that mandatory use of the age criteria improved sensitivity for older patients but also resulted in lower specificity (Nakamura et al., 2012). The age criteria were also the most common element documented for those who met Step 4, in contrast to other work in which EMS provider judgment-an element of the Step 4 criteria-has been reported to be the most common element documented for patients who meet special considerations criteria, and the most common element documented (48%) of any in the field triage screening process (Jones et al., 2016; Newgard et al., 2012).

Although the somewhat subjective nature of the EMS provider judgment has caused concern among some authors, it has also been heralded as an early and important step in studies that look at the actual uptake and execution of the 2011 CDC Guidelines for Field Triage of Injured Patients, and an essential element of the "provider gestalt" that drives actual EMS trauma triage (Jones et al., 2016; Newgard et al., 2012). In fact, some data suggest that EMS provider judgment has trauma center need sensitivity that is superior to vital signs or MOI (Lavoie et al., 2010) and is highly sensitive to the need for trauma center care even when physiologic abnormalities and overt anatomic injuries are not present (Mulholland et al., 2008). Some authors have expressed concern about the potential role of bias in the provider judgment determination (Mulholland et al., 2005).

In this study, there appears to be a disproportionate use of special considerations for patients whose race and ethnicity were documented as White, not Hispanic, or Latino(a) (81% of all Step 4 patients) compared with their distribution in the study population (65%). However, those positive in Step 4 for provider judgment in this study mirrored the study population regarding race and ethnicity. The disproportionate number of those documented as White, not Hispanic, or Latino(a) appears to be driven by the 55 and above age category and those meeting multiple special considerations criteria (see Supplemental Digital Content Table 1, available at: http://links.lww.com/JTN/A80). This disparity warrants further study to determine whether all field triage criteria are applied equitably to injured patients regardless of race or ethnicity.

In this study, the most striking predictor of hospital admission and mortality was the presence of two or more positive steps, with nearly 10% of patients in this group dying in the ED (3.3%) or during their hospital stay (6.5%). Although relationships between EMS findings that are similar to individual criteria of the 201 CDC Guidelines (i.e., prehospital index, high velocity impact, and EMT judgment) and trauma center need have been reported (Lavoie et al., 2010; Mulholland et al., 2008), we believe this finding relating two or more positive steps from the 2011 CDC Guidelines for Field Triage of Injured Patients and hospital outcomes has not been reported previously and has the potential to guide revisions of trauma triage protocols, especially related to destination and transport decisions for patients in remote regions. As has been suggested by others (Lyng et al., 2021), the high mortality rate of patients with two or more positive steps may suggest a need-for patients in remote locations-for early stabilization by air medical or nondesignated hospital teams prior to long-distance transport to a designated trauma center. More study is necessary.

Overall, this study supports using the 2011 CDC Guidelines for Field Triage of Injured Patients as a tool to support destination decisions for trauma patients, although some adjustments may result in stronger predictions. Step 2: AOI may benefit from revisions to reduce the EMS false-positive rate. Step 3: MOI appears to have limited predictive power and may be stronger with elimination of MOI elements that have not demonstrated a strong relationship to hospital admission and mortality. Further study may determine the optimal and equitable use of age consideration from Step 4, optimizing sensitivity and specificity.

Limitations

This was a large retrospective evaluation of linked prehospital patient care records obtained from a convenience sample of EMS agencies who voluntarily contributed de-identified data for research. Thus, these data were originally recorded for clinical care purposes and not as part of a study. Our study population was also overwhelmingly from urban communities. These limitations may impact generalizability of the study findings.

The decision to complete the 2011 CDC Guidelines for Field Triage of Injured Patients screening tool was based on EMS system protocols and EMS provider discretion. Not all injured patients requiring EMS transport had a completed form. It is possible that providers used the CDC Guidelines as the basis for patient transport but did not document this in the patient care report; conversely, the screen may not have been completed for patients with perceived minor injuries. Therefore, the study population may not include the totality of injured patients transported by EMS based on the 2011 CDC Guidelines for Field Triage of Injured Patients.

Our inclusion criteria centered on EMS patients with linked hospital data. A small percentage (<3%) of patients did not have mortality data available due to being transferred to another facility or simply missing data. While unlikely, it is possible that outcomes for this small percentage of patients with missing mortality data were different from those reported in this analysis.

CONCLUSIONS

Patients meeting multiple CDC Guidelines for Field Triage of Injured Patients steps were at greater risk of hospitalization and death. Among those positive in only one step, AOI, vital signs, and LOC criteria were associated with a greater risk of hospitalization. In comparison, vital signs and LOC criteria were associated with a greater risk of death. Collectively, these findings may help inform EMS destination decisions and the creation of trauma center-specific activation criteria for patients meeting select CDC Guidelines steps.

References