Abstract
Background:
Trauma field triage matches injured patients to the appropriate level of care. Prior work suggests the Glasgow Coma Scale motor (GCSm) is as accurate as the total GCS (GCSt), and easier to use. However, older patients present with higher GCS for a given injury and as such it is unclear if this substitution is advisable. Our objective was to compare the GCS deficit patterns between geriatric and adult patients presenting with severe TBI, as well as the diagnostic performance of the GCSm versus GCSt within the field triage criteria in these populations.
Materials and Methods:
We conducted a retrospective, observational cohort study of patients ≥16 years in the National Trauma Data Bank 2007-15. GCS deficit patterns were compared between adults (16-65) and geriatric patients (>65). Measures of diagnostic performance of GCSt≤13 versus GCSm≤5 criteria to predict trauma center need (TCN) were compared.
Results:
4,480,185 patients were analyzed (28% geriatric). Geriatric patients more frequently presented with non-motor only deficits than adults (16.4% vs. 12.4%, p<0.001) and these patients demonstrated higher severe TBI (40.3% vs. 36.7%, p<0.001) and craniotomy (5.8% vs. 5.1%, p<0.001) rates. GCSt was more sensitive and accurate in predicting TCN for geriatric patients and had lower rates of undertriage as compared to GCSm.
Conclusions:
Geriatric patients more frequently present with non-motor only deficits after injury and this is associated with severe head injury. Substitution of GCSm for GCSt would exacerbate undertriage in geriatric patients and, thus, the total GCS should be maintained for field triage in geriatric patients.
Keywords: Geriatric, Triage, Glasgow Coma Scale, Traumatic Brain Injury, Prehospital
Introduction
The National Field Triage Guidelines (NFTG) represent the consensus recommendations put forth by the American College of Surgeons Committee on Trauma and the Centers for Disease Control.1 The NFTG are a step-wise algorithm to help prehospital (PH) providers identify patients who require the resources of a trauma center. The steps include physiologic, anatomic, mechanism of injury, and special consideration criteria, structured in descending specificity for severe injury.
Geriatric patients comprise an estimated 25% all trauma patients in the U.S. and this is expected to grow. Despite the substantial number of trauma activations represented, this population encounters persistently high undertriage.2, 3 They present with a disproportionate trauma burden for low-energy mechanisms than comparable mechanisms in younger adult patients.4 This is commonly observed in ground-level falls. Geriatric patients mount a different physiologic response to stress and injury; and frailty, or the lack of physiologic resilience, contributes to adverse outcomes after injury. This population demonstrates a blunted hemodynamic response to blood loss and a unique neurologic response to head injury with delayed, but rapid deterioration.5–7 Given the differences of injury physiology, the most recent NFTG include a separate criterion for a systolic blood pressure (SBP), indicating <110 mmHg may represent shock in older patients.8
The Glasgow Coma Scale (GCS) is well established and has long been used as a criterion in the NFTG. The components of GCS are motor, verbal and eye-opening subscales, with a GCS ≥13 considered positive criterion for transport to the highest level of trauma care. The GCS, however, is often considered complex for the PH setting.9, 10 At the last NFTG revision in 2011, consideration was given to substitution of the GCS motor for GCS total for field triage due to its relative ease of use in the field. Ultimately, this change was not made due to a lack of strong evidence. However, since then, several studies have demonstrated GCS motor (GCSm) is easier to employ in the field, more reliable, and has similar accuracy to the GCS total (GCSt) for field triage as a predictor of traumatic brain injury (TBI) and mortality.9, 11, 12
The NFTG are again under revision and given this interim data, a substitution of the GCSm for GCSt is surely under consideration. Importantly, however, these studies do not specifically evaluate the geriatric population and no further guideline adaptations are recommended for these patients. 2–4, 13, 14 Geriatric patients present with a different injury pattern and higher total GCS than younger adult patients with a TBI of comparable severity.15–17 Therefore, it is important that we evaluate them independently and consider the need for unique neurologic triage criteria. Two retrospective studies evaluated nearly 13,000 patients and sought to determine GCS differences upon presentation in geriatric TBI and the associated outcomes. Kehoe et al15 found that overall geriatric GCS was higher despite equal anatomic severity as measured by AIS for all levels, including for head AIS 5, and that these patients experienced higher mortality. They postulated a potential mechanism via increased accommodation for hematoma and edema volume as a result of parenchymal atrophy that delays elevation in intracerebral pressure that might precede GCS deficits. Susman et al17 also demonstrated a higher overall GCS despite lower ISS in geriatric patients and that even a GCS >13 was associated with an 8-fold increase in adjusted mortality and worsened functional status at discharge for geriatric TBI. Because of this, it is unclear if the GCSm performs accurately in this population to allow for safe substitution in triage protocols for geriatric patients.
Our objective was to compare the GCS deficit patterns between geriatric and younger adult patients presenting with severe TBI, as well as the diagnostic performance of the GCSm versus GCSt within the field triage scheme in these respective populations. We hypothesized geriatric patients demonstrate a unique GCS deficit pattern and the GCS motor criterion would be less accurate than the GCS total in field triage of geriatric patients.
Material and Methods
Study Design & Population
We conducted a retrospective, observational cohort study utilizing the National Trauma Data Bank (NTDB). All subjects age 16 years and older transported from the scene of injury during 2007-2015 were included. Patients who underwent inter-facility transfer, had burn injury, or were dead on arrival were excluded. Demographics, injury characteristics and severity, prehospital and admission vital signs, International Classification of Diseases (ICD) 9th revision diagnosis codes, ICU admission, emergency department (ED) disposition, and hospital disposition were abstracted for each patient. The adult cohort was defined as ages 16 through 65 years. The geriatric cohort was defined as age greater than 65 years.
GCS Deficit Patterns
Given that geriatric patients are thought to present with unique GCS deficits after head trauma, we evaluated the pattern of TBI according to GCS deficits. To accomplish this, all patients with a GCS ≥13 were selected and further categorized into those who had a motor only deficit (Verbal=5, Eye=4, Motor <6), non-motor only (Verbal < 6 OR Eye <4, Motor=6) and those with both motor and non-motor deficits. The first two subsets were of interest in the present study given that patients with both deficit patterns would be captured by either the GCSt or an abnormal GCSm.
Triage Criteria
The presence of physiologic step 1 and anatomic step 2 NFTG criteria were determined for each patient based on prehospital data for step 1 including GCS total ≥13, SBP <90 mmHg, respiratory rate (RR) < 10 or >29 breaths/minute or need for ventilatory support; and anatomic injury classification based on ICD-9 diagnosis codes for step 2 including penetrating injury, flail chest, open skull fracture, ≤2 proximal long bone fractures, pelvic fracture, crush injury, amputation, and paralysis. A GCSm ≥5 was defined as a positive triage criterion to allow comparison to the current GCS ≥13 criterion for this study. For purposes of assessing diagnostic accuracy of the NFTG criteria, patients were assessed triage positive using all eleven physiologic and anatomic criteria as would be applied in the field. This was performed separately, using the current GCS ≥13 with all other step 1 and step 2 criteria, and subsequently using GCSm ≥5 instead of total GCS, again with all other step 1 and step 2 criteria for comparison. We used the other step 1 and step 2 criteria to evaluate the real-world application to the NFTG, where differences between the individual GCSt and GCSm criteria may ultimately not impact overall NFTG performance if patients are identified through other criteria in the algorithm.
Missing Data
To address missing data, multiple imputation was performed for prehospital EMS data including SBP, respiratory rate (RR), and Glasgow Coma Scale total and motor subscales. A fully conditional specification model based on available age, prehospital and admission physiology, and mortality was performed using chained multiple imputations with ten imputations to develop a complete data set. The imputed prehospital vital signs used in the NFTG step 1 criteria utilize numeric threshold values, and thus it was possible for criteria to be both positive and negative in the same patient across imputations. For this reason, the average values for SBP and RR, and the median values for GCSt and GCSm across imputations were used to determine whether NFTG step 1 criteria were present or absent. Between 30% and 35% of observations were missing for the imputed variables (Appendix, Table A.1). Complete and incomplete case analysis was compared to affirm the fidelity of the multiple imputation procedure. The distribution of imputed variables was similar in the imputed and complete data. No differences were observed between complete case analysis and imputed analysis, thus results from the imputed data are presented (Appendix, Table A.2).
Outcomes
The primary outcome measures were severe TBI, defined as head abbreviated injury scale (AIS) ≥3, and need for craniotomy within 24 hours of admission. Our secondary outcome for evaluating NFTG triage diagnostic accuracy was trauma center need (TCN). TCN is a previously published composite of injury severity and indicators of hospital resource utilization.8 It was defined as a composite of any of the following: injury severity score (ISS) >15, ICU admission 24 hours or greater, need for urgent surgery (ED disposition to the operating room), or ED mortality. We used this composite to identify patients most likely to benefit from the specialized care of a trauma center18, 19, and it represents the gold standard against which the accuracy of the field triage criteria are evaluated within this study.
Statistical Analysis
To characterize the pattern of TBI and associated outcomes, we stratified patients based on their GCS deficit (non-motor versus motor-only). We compared severe TBI (head AIS ≥3), the need for craniotomy within 24 hours of admission, and TCN between geriatrics and younger adults. Diagnostic performance of GCSm versus GCSt for both geriatric and younger adult cohorts was tested in their ability to predict TCN and included sensitivity, specificity, positive predictive values (PPV), negative predictive value (NPV), overtriage, undertriage, and accuracy. Receiver operating characteristic area under the curve (AUC) was calculated and compared for the NFTG using GCSm versus GCSt.
Continuous and categorical data are reported as median (interquartile range) and frequency (percentage), respectively. For univariate comparisons, Wilcoxon rank-sum tests were used for non-normally distributed continuous variables and Chi-square tests were used for categorical variables. Absolute standardized difference was used for comparisons of patient demographics and characteristics. The standardized difference represents the difference between groups divided by the pooled standard deviation, making it insensitive to large samples sizes. An absolute standardized difference greater than 0.1 represents significant non-overlap in the distributions of a given variable between the two groups. Reported p-values were two-tailed, and a p-value <0.05 was considered statistically significant. All analyses were performed using Stata version 15 (StataCorps LP, College Station, TX). Given the de-identified dataset, the current investigation was exempted by the Institutional Review Board.
Subgroup Analyses
Several subgroup analyses were performed for underlying injuries or comorbidities that could theoretically alter the accuracy of GCS assessment to detect TBI and predict trauma center need. First, we evaluated the impact of TBI subtype on GCS deficits and outcomes. TBI was classified into contusion, subdural hematoma, epidural hematoma, subarachnoid hemorrhage, and multi-compartment hemorrhage based on ICD-9 diagnosis codes. Next, we evaluated the accuracy of GCS in geriatric patients with a history of stroke or dementia given that these comorbidities may confound the accurate calculation of GCS and its subscales. To this end, we performed a subgroup analysis comparing those with these comorbidities to those without, across all major endpoints. Finally, we also evaluated age subgroups within the geriatric cohort, comprised of age 66-75, age 76-85, and age >85 to determine whether more precise levels of aging within the geriatric cohort impacted the field GCS assessment and accuracy as a triage criterion.
Sensitivity Analysis
Given the lower threshold of ICU admission among geriatric patients, we performed a sensitivity analysis altering our definition of TCN using an ICU admission for 72 hours or greater as part of the composite rather than ICU admission 24 hours or greater.
Results
There were 4,480,185 patients who met inclusion criteria, of which 1,258,190 (28%) were geriatric. Geriatric patients were predominantly female with higher overall injury severity, and greater mortality than adult patients (Table 1). Trauma center need was present in the geriatric and adult cohorts in 30.3% and 34.7%, respectively.
Table 1:
Patient Demographics, Prehospital Characteristics, and Outcomes
| Geriatric+ n=1,258,190 (28%) | Adult^ n=3,221,995 (72%) | Absolute Standardized Difference† | |
|---|---|---|---|
| Age | 81 (73, 87) | 37 (25, 51) | 3.576* |
| Sex (male) | 485011 (38.7%) | 2274892 (70.9%) | 0.683* |
| TCN | 381742 (30.3%) | 1117667 (34.7%) | 0.093 |
| ISS | 9 (4, 10) | 6 (4, 13) | 0.059 |
| GCS ≤13 | 96804 (7.7%) | 403812 (12.5%) | 0.173* |
| SBP <90mmHg | 21985 (1.7%) | 111032 (3.4%) | 0.107* |
| RR <10 or >29bpm | 23398 (1.9%) | 128816 (4.0%) | 0.127* |
| Penetrating injury | 14454 (1.1%) | 454697 (14.1%) | 0.504* |
| Flail chest | 6201 (0.5%) | 18475 (0.6%) | 0.011 |
| Open skull fracture | 3877 (0.3%) | 39541 (1.2%) | 0.105* |
| ≥2 long bone fractures | 29094 (2.3%) | 40627 (1.3%) | 0.079 |
| Pelvic fracture | 102696 (8.2%) | 199611 (6.2%) | 0.076 |
| Crush injury | 1420 (0.1%) | 16161 (0.5%) | 0.070 |
| Amputation | 580 (0.05%) | 6920 (0.2%) | 0.047 |
| Paralysis | 1791 (0.1%) | 11020 (0.3%) | 0.041 |
| Mortality | 64652 (5.1%) | 112352 (3.5%) | 0.081 |
Geriatric: >65 years
Adult: 16-65 years
Continuous variables presented as median (IQR); Categorical variables presented as n (%)
TCN, Trauma Center Need; ISS, injury severity score; GCS, Glasgow Coma Scale; SBP, systolic blood pressure; RR, respiratory rate
The absolute standardized difference represents the difference between groups divided by the pooled standard deviation, making it insensitive to large samples sizes unlike hypothesis testing that generate p-values.
Absolute standardized difference >0.1 indicates imbalance between groups.
The pattern of GCS deficits is shown in Figure 1. Overall, geriatric patients presented more frequently with non-motor only deficits than adults (16.7% versus 13.4%, p<0.001). Notably, among patients with severe TBI, geriatrics also presented with non-motor only deficits more frequently than adults (47.6% versus 27.5%, p <0.001). Likewise, among patients that required craniotomy, non-motor only deficits were more frequent in the geriatric cohort as compared to younger adult patients (35.6% versus 15.9%, p<0.001). This demonstrates geriatric patients who sustain a severe head injury present with unique GCS deficit patterns that more commonly manifest with deficits of the verbal or eye-opening scores and not the motor subscale.
Figure 1:
GCS deficit patterns classified by GCS subscales in adult & geriatric trauma patients
In the geriatric cohort, patients with non-motor only deficits demonstrated a significantly higher frequency of severe TBI and need for craniotomy than patients with motor only deficits. Within this same cohort, TCN was also greater, however this did not reach statistical significance (Table 2). Conversely, adult patients with motor only deficits had higher rates of severe TBI and craniotomy than adults that presented with non-motor deficits. A motor only deficit was significantly associated with greater TCN (Table 2). In other words, when geriatric patients suffer a severe TBI, they are more likely to manifest this with a non-motor only GCS deficit.
Table 2:
GCS Deficit & Outcomes in Geriatric and Adult Cohorts
| Non-motor only deficits | Motor only deficits | p-value | |
|---|---|---|---|
| Geriatric* | |||
| Severe TBI | 5484 (40.3%) | 9977 (36.7%) | <0.001 |
| Craniotomy | 814 (5.8%) | 1438 (5.1%) | <0.001 |
| TCN | 7853 (57.8%) | 15459 (56.9%) | 0.092 |
| Adult^ | |||
| Severe TBI | 13411 (27.0%) | 26902 (28.4%) | <0.001 |
| Craniotomy | 2275 (4.4%) | 5605 (5.5%) | <0.001 |
| TCN | 27759 (55.9%) | 60389 (63.7%) | <0.001 |
Geriatric: >65 years
Adult: 16-65 years
Numbers reported are N (%)
TBI, traumatic brain injury; TCN, trauma center need
Regarding TBI subtypes, geriatric patients presented most frequently with SDH (43%), followed by multi-compartment hemorrhage (25%). Multi-compartmental hemorrhage was most common in the adult cohort (32%), followed by subarachnoid hemorrhage (27%; Appendix, Table A.3). In subgroup analysis we demonstrate a higher frequency of non-motor only deficits in geriatric patients for every TBI subtype as compared to adults (Table 3). This, again, affirms the unique GCS deficit manifestation in geriatric patients who sustain a severe head injury, irrespective of TBI subtype.
Table 3:
GCS Deficit & Outcomes in Geriatric and Adult Cohorts stratified by TBI Type
| Contusion | SAH | SDH | EDH | MCH | p-value | |
|---|---|---|---|---|---|---|
| Geriatric* | n=2617 | n=7216 | n=11826 | n=282 | n=14950 | |
| Non-Motor Only | 369 (14.1%) | 1212 (16.8%) | 1647 (13.9%) | 38 (13.5%) | 1906 (12.7%) | <0.001 |
| Motor Only | 771 (29.5%) | 2225 (30.8%) | 3309 (28.0%) | 81 (28.7%) | 3889 (26.0%) | |
| Both | 1477 (56.4%) | 3779 (52.4%) | 6870 (58.1%) | 163 (57.8%) | 9155 (61.2%) | |
| Adult^ | n=15713 | n=25121 | n=19506 | n=2074 | n=48497 | |
| Non-Motor Only | 1384 (8.8%) | 2733 (10.9%) | 2052 (10.5) | 275 (13.3%) | 3606 (7.4%) | <0.001 |
| Motor Only | 3136 (20.0%) | 3136 (20.0%) | 3959 (20.3%) | 548 (26.4%) | 8659 (17.9%) | |
| Both | 11193 (71.2%) | 17212 (68.5%) | 13495 (69.2%) | 1251 (60.3%) | 36232 (74.7%) | |
Geriatric: >65 years
Adult: 16-65 years
Numbers reported are N (%)
TBI, traumatic brain injury; SAH, subarachnoid hemorrhage; SDH, subdural hematoma; EDH, epidural hematoma; MCH, multicompartment hemorrhage
We then evaluated the diagnostic performance of the NFTG to predict TCN and, as expected, substitution of GCSm ≤5 for GCS ≤13 resulted in lower sensitivity and negative predictive value and higher specificity in both adult and geriatric cohorts. Importantly, this analysis was inclusive of all step 1 and 2 physiologic and anatomic criteria to best approximate how the algorithm would be applied in the field and to ensure that any changes in diagnostic performance were only attributable to the alteration of the GCS classification—GCSm vs GCSt. Accuracy and AUC were identical using GCSt or GCSm in the adult cohort (p=0.06); however, accuracy and AUC were slightly higher for GCSt in the geriatric cohort (p<0.01; Table 4). When utilizing the GCSt for geriatric patients, there were 8,177 less false negatives, resulting in a 2.1% reduction in undertriage. In contrast, GCSt overtriaged 5,194 (0.6%) more geriatric patients than GCSm. Thus, there would be a net 1.5% benefit of employing GCSt over GCSm in geriatric patients.
Table 4:
Diagnostic Performance: NFTG Predicting TCN using GCS Total versus GCS Motor
| Sensitivity | Specificity | PPV | NPV | Accuracy | Undertriage | Overtriage | AUC | p-value | |
|---|---|---|---|---|---|---|---|---|---|
| Geriatric* | |||||||||
| GCSt ≤13 | 33.1% | 83.4% | 46.5% | 74.1% | 68.1% | 66.9% | 16.6% | 0.583 | <0.01 |
| GCSm ≤5 | 31.0% | 84.0% | 45.7% | 73.6% | 67.9% | 69.0% | 16.0% | 0.575 | |
| Adult^ | |||||||||
| GCSt ≤13 | 54.7% | 75.2% | 54.0% | 75.8% | 68.0% | 45.3% | 24.8% | 0.644 | 0.06 |
| GCSm ≤5 | 53.4% | 76.2% | 53.9% | 75.1% | 68.0% | 47.6% | 23.8% | 0.643 | |
Geriatric: >65 years,
Adult: 16-65 years
Results represent the overall diagnostic performance for Step 1 and Step 2 criteria, varied by GCS classification: GCS total versus GCS motor.
NFTG, national field triage guidelines; TCN, trauma center need, GCS, Glasgow Coma Scale; PPV, positive predictive value; NPV, negative predictive value
In the geriatric cohort, 8.8% of patients had a prior diagnosis of CVA or dementia. The distribution of non-motor only and motor only deficits was similar among geriatric patients with or without CVA or dementia (Appendix, Table A.4). Severe TBI was slightly more common in those without CVA or dementia, while craniotomy was similar in geriatric patients with or without CVA or dementia. TCN and mortality were higher among geriatric patients with CVA or dementia (Appendix, Table A.5).
When stratifying geriatric patients by age subgroups, we see that ISS is similar and, not surprisingly, mortality increases with age (Appendix, Table A.6). The pattern of GCS deficits in geriatric patients shows a significant increase in non-motor only deficits across the age subgroups (p<0.01, Appendix, Table A.7). Non-motor only deficits are also greater associated with higher rates of severe TBI, craniotomy, and TCN as age increases (Table 5).
Table 5:
GCS Deficit & Outcomes in Geriatric Age Subgroups
| Non-motor only deficits | Motor only deficits | p-value | |
|---|---|---|---|
| 66-75 years | |||
| Severe TBI | 1919 (41.2%) | 3326 (39.5%) | 0.051 |
| Craniotomy | 267 (5.7%) | 457 (5.4%) | 0.460 |
| TCN | 2904 (62.4%) | 5450 (64.7%) | 0.009 |
| 76-85 years | |||
| Severe TBI | 2084 (42.6%) | 3646 (38.5%) | <0.001 |
| Craniotomy | 243 (5.0%) | 452 (4.8%) | 0.600 |
| TCN | 2915 (59.6%) | 5571 (58.8%) | 0.340 |
| 86+ years | |||
| Severe TBI | 1481 (36.6%) | 3005 (32.4%) | <0.001 |
| Craniotomy | 98 (2.4%) | 167 (1.8%) | 0.019 |
| TCN | 2034 (50.3%) | 4438 (47.9%) | 0.012 |
Numbers reported are N (%)
TBI, traumatic brain injury; TCN, trauma center need
When evaluating use of ICU admission 72 hours or greater in our composite definition of TCN, we saw modest overall improvements in the NFTG criteria to predict TCN, but similar relationships in adult and geriatric patients regarding GCSt and GCSm (Appendix, Table A.8)
Discussion
Our results demonstrate geriatric patients who suffer severe head trauma or require urgent craniotomy more commonly present with non-motor only deficits compared to younger adults. This observation held true across TBI subtypes, and among patients with potential confounding and concomitant comorbidities and/or injuries. This problem is exacerbated as age increases within the geriatric cohort. This highlights our concern with revising the triage guidelines to the exclusive use of GCSm for step 1. The geriatric patients frequently present with non-motor only deficits, and by definition would be undertriaged if GCSm was employed as the new triage criterion. This would worsen an already suboptimal criterion and further jeopardize the vulnerable geriatric population.
Among geriatric patients, use of the GCSt reduced undertriage with minimal increase in overtriage, resulting in a net benefit in triage accuracy. This benefit in triage accuracy was observed after taking into account all other step 1 and step 2 criteria and is important considering greater value is placed on the reduction of undertriage from a trauma systems perspective. Although the differences in performance characteristics are small, the considered alteration to national guidelines could translate to large absolute numbers of patients undertriaged across the country in a population that already experiences unacceptably high undertriage.
Trauma triage criteria are established to ensure that severely injured patients are transported to the appropriate level of trauma center. At the time of the last NFTG revision in 2011, the expert panel discussed adding the GCS motor as a substitution for the GCS total due to its comparative simplicity of use for PH providers. Ultimately, no changes were recommended as the panel cited a “lack of confirmatory evidence, long standing use of the GCS, and its familiarity among current EMS practitioners”.1 However, since then, multiple studies have demonstrated a comparable discriminatory performance of the GCSm vs. GCSt for trauma patients irrespective of age.9, 11, 20–23 Brown et al9 compared the use of GCS ≤13 versus GCSm ≤5 as positive triage criteria in the ability to predict TCN in trauma patients of all ages. GCSm had similar discrimination as GCSt and was more strongly associated with and better calibrated to predict TCN. Ross and colleagues investigated the role of GCSm to predict TBI.12 The authors concluded that a GCSm ≤5 was equivalent to GCSt for identifying patients at risk for death, those requiring craniotomy, and those with TBI, and that it was more reliably recorded in the field and thus endorsed it as a preferential criterion over GCSt. Importantly, neither groups independently investigated the geriatric population. The NFTG are currently under evaluation for further revision 24; this additional published evidence may support the substitution of the GCSm for the GCSt in the Step 1 physiologic criteria. Pennsylvania has already substituted a GCSm ≤5 into step 1 criteria in the state-wide EMS protocol.25 Our data suggest this may be detrimental to geriatric patients.
Geriatric patients are often considered to be low risk in the field by existing criteria yet suffer substantial morbidity and mortality. This mis-categorization of injury severity is partially attributable to their differential response to trauma that does not objectively trigger the traditional triage criteria.2, 5–8 Our work suggests that a decision to substitute GCS motor for GCS total for geriatric patients would exacerbate this mis-categorization. The reasons for the differential GCS deficit manifestations in geriatric patients are not well understood. However, we speculate that it is multifactorial and that the non-motor predominance may provide a reflection of the generally poorer functional status of geriatric patients. Additionally, there is an inherent difficulty of interpreting potential deficits given the lack of known functional/physiologic baseline of commonly comorbid geriatric patients. Lastly, there may be a comparative delay in motor manifestations that may be due to greater accommodation of intracranial volume for bleeding due to brain atrophy with aging. Benjamin et al 2 examined older blunt trauma patients in the NTDB who did not meet the criteria for the highest level of trauma team activation and therefore were not triaged to a trauma center. They demonstrated an increased early mortality among this cohort. Those suffering a TBI had nearly three times the likelihood of mortality on adjusted analysis. Remarkably, these patients had increased mortality despite having vital signs and GCS that did not trigger NFTG criteria. This phenomenon has been corroborated in the literature and may indicate the need for different geriatric triage criteria altogether.3, 5, 7, 8 This is supported by current estimates of undertriage in older patients that are in excess of 50%.2, 3, 6, 14 Notably, in the present study, NFTG criteria have significantly lower performance in the geriatric patients compared to adults. Because of this, we also acknowledge that geriatric triage likely requires an independent set of criteria. Further research is necessary to evaluate potential alternative triage criteria, including alternatives to GCS to improve triage performance in this vulnerable population. Werman et al 26 developed a statewide geriatric-specific triage protocol in Ohio. Their data reinforced the observation that geriatric patients present with a higher overall GCS regardless of injury severity. They noted an inflection of increased mortality among geriatric patients with GCS ≤14 compared to the existing GCS ≤13 criterion; however, they did not examine GCSm as a potential triage criterion.
It is, however, important to recognize that sufficient evidence is not yet available to dictate a fundamental change in triage criteria for these patients and as such, the GCS will likely remain in the NFTG physiologic criteria for the near future. The revision currently under consideration: substitution of GCSm for GCSt could prove costly for geriatric patients. We acknowledge that undertriage rates are unacceptably high in this population, evidenced both in our study and prior work. However, we should not publish a revision in the current guidelines that would make a bad situation worse for geriatric patients. Even small percentage improvements in undertriage are worth considering given the high frequency of TBI in geriatric patients across the nation which could result in improved outcomes for thousands of these patients. At minimum, with consideration of our findings, any revisionary substitution to use the GCSm should include, in the special consideration criteria, an alternate use of total GCS for geriatric patients. This would be similar to the current SBP criterion of 110; and avoid higher undertriage, suboptimal care, and adverse outcomes in these patients.
Our findings have several implications. Of note is the disparate GCS deficit presentation of adult and geriatric patients after TBI. Geriatrics more frequently present with a normal motor exam with either verbal and/or eye-opening deficits. This is contrasted from younger adult patients who more frequently display a motor deficit after severe TBI. It is critical that we optimize our guidelines to reflect these differences and better equip PH providers to accurately triage injured geriatric patients. Additionally, these data demonstrate prehospital and in-hospital providers should maintain a high index of suspicion for severe head injury in geriatric patients without motor deficits on exam. While we found similar and improved performance of GCSt relative to GCSm in geriatric patients with and without a history of dementia or stroke, the overall accuracy of GCS was lower in patients that did carry these comorbidities. Thus, a different scale to measure neurologic function may be more appropriate in the geriatric trauma patient for field triage and requires ongoing investigation.
This study has important limitations. This is a retrospective design with pre-determined variables available for analysis. Within the NTDB, PH provider triage rationale is not documented. Thus, we are limited by the retrospective use of ICD coding to estimate field use of step 2 anatomic criteria to predict TCN which does not necessarily reflect the diagnostic capabilities of a scene responder. Additionally, there is limited data collected on mechanism of injury (step 3) or special consideration factors (step 4). Without this granular data, steps 3 and 4 could not be evaluated as part of the NFTG in this study. Intuitively, ICD coding will identify a greater prevalence of injuries than scene providers and the inability to include steps 3 or 4 decreases the sensitivity of our analyses to detect TCN and increased our calculated rates of undertriage. This limitation is consistent with prior work utilizing this database that demonstrated a higher degree of undertriage when steps 1 and 2 are evaluated in isolation.27 Thus, this important limitation leads to the overall poorer performance of the NFTG in general with low sensitivity and AUC, which is why we restricted the diagnostic performance analysis as an exploratory secondary outcome. Notably, step 1 (physiologic) and step 2 (anatomic) criteria provide indications for transport to the highest level of care within the defined trauma system, while step 3 and step 4 criteria do not necessitate triage to the highest level trauma center in the system.1 Thus, despite this limitation, we identify those patients requiring the greatest extent of trauma system resources. In addition, we define TBI by AIS and it is important to acknowledge that AIS coding has undergone revisions during the study period (2008, 2015); however, the NTDB provides codes mapped to the 1990 version for consistency over time, which we utilized. ICD-9 diagnosis codes also underwent annual changes; however, changes within the injury diagnosis codes generally added specificity and would not change the broader injury classifications we have used them for in this study.
Inherent to the NTDB, there is the potential for selection bias as data submission had been voluntary; however, it is now a requirement of level 1 and 2 centers, making it more nationally representative. Lastly, there is substantial missing data, primarily in the prehospital physiologic data. However, multiple imputation was utilized to mitigate this and has previously been validated as best practice for dealing with missing data in the NTDB, especially for step 1 critieria.28–30 We believe the use of the imputed dataset allows for a more cogent analysis in which selection bias of patients with missing data is avoided as there were differences between patients with missing and complete data, particularly for those who did meet GCSt and GCSm triage criteria, when present. The validity of our findings with the imputed dataset is supported by the similar imputed PH datapoints and outcomes across sensitivity analyses, thereby engendering confidence in our results. Our findings are clinically impactful, however, given the aforementioned limitations, the issues we raise on geriatric triage would be most accurately answered in a prospective trial.
Conclusions
This is the first study to demonstrate that geriatric patients more frequently present with non-motor only deficits upon GCS assessment in the field, and these non-motor deficits are associated with severe TBI and need for urgent craniotomy in a national sample. While substituting GCSm ≤5 for the current GCS ≤13 in the context of the NFTG may be desirable among younger adult trauma patients, this change would exacerbate undertriage in geriatric patients. Thus, the total GCS should be maintained for field triage in geriatric patients, and ongoing study is warranted to optimize field triage criteria in our older trauma population.
Supplementary Material
Table A.1: Summary of EMS variables in Complete Cases versus Multiple Imputed Cases
Table A.2: Patient Demographics and Prehospital Characteristics, by Data Completeness
Table A.3: TBI Subtypes in Geriatric and Adult Cohorts
Table A.4: GCS Deficits in Geriatric Cohort with prehospital GCS≤13 and single deficit, stratified by CVA/ Dementia
Table A.5: Presence of CVA/ Dementia & Outcomes in Geriatric Cohort
Table A.6: Patient Demographics, Prehospital Characteristics and Outcomes in Geriatric Cohort, Stratified by Age Subgroups
Table A.7: GCS Deficits in Geriatric Cohort with prehospital GCS ≤13, stratified by Age Subgroups
Table A.8: Diagnostic Performance: NFTG Predicting TCN using GCS Total versus GCS Motor—Sensitivity Analysis in Geriatric Patients by altered TCN composite
Acknowledgments:
Committee on Trauma, American College of Surgeons. NTDB Admission Years 2007-2015. Chicago, IL. The content reproduced from the NTDB remains the full and exclusive copyrighted property of the American College of Surgeons. The American College of Surgeons is not responsible for any claims arising from works based on the original data, text, tables, or figures.
Funding:
This work was supported by the National Institutes of Health [Grant: 2T32GM008516-26, 2020].
Support: National Institutes of Health [Grant: 2T32GM008516-26, 2020]
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Disclosure of Interest:
The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.
Research Data:
The data that support the findings of this study are available from the American College of Surgeons at: https://www.facs.org/quality-programs/trauma/tqp/center-programs/ntdb/datasets Restrictions apply to the availability of this data, which was used under license for this study.
Declarations of interest: none
This paper was presented at the 78th Annual Meeting of The American Association for the Surgery of Trauma and Clinical Congress of Acute Care Surgery, September 2019, Dallas, Texas.
References
- 1.Sasser SM, Hunt RC, Faul M, Sugerman D, Pearson WS, Dulski T, Wald MM, Jurkovich GJ, Newgard CD, Lerner EB, et al. Guidelines for field triage of injured patients: recommendations of the National Expert Panel on Field Triage, 2011. MMWR Recomm Rep. 2012;61:1–20. [PubMed] [Google Scholar]
- 2.Benjamin ER, Khor D, Cho J, Biswas S, Inaba K, Demetriades D. The Age of Undertriage: Current Trauma Triage Criteria Underestimate The Role of Age and Comorbidities in Early Mortality. J Emerg Med. 2018;55:278–287. [DOI] [PubMed] [Google Scholar]
- 3.Chang DC, Bass RR, Cornwell EE, Mackenzie EJ. Undertriage of elderly trauma patients to state-designated trauma centers. Arch Surg. 2008;143:776–781; discussion 782. [DOI] [PubMed] [Google Scholar]
- 4.Velmahos GC, Jindal A, Chan LS, Murray JA, Vassiliu P, Berne TV, Asensio J, Demetriades D. “Insignificant” mechanism of injury: not to be taken lightly. J Am Coll Surg. 2001;192:147–152. [DOI] [PubMed] [Google Scholar]
- 5.Heffernan DS, Thakkar RK, Monaghan SF, Ravindran R, Adams CA Jr., Kozloff MS, Gregg SC, Connolly MD, Machan JT, Cioffi WG. Normal presenting vital signs are unreliable in geriatric blunt trauma victims. J Trauma. 2010;69:813–820. [DOI] [PubMed] [Google Scholar]
- 6.Kodadek LM, Selvarajah S, Velopulos CG, Haut ER, Haider AH. Undertriage of older trauma patients: is this a national phenomenon? J Surg Res. 2015;199:220–229. [DOI] [PubMed] [Google Scholar]
- 7.Martin JT, Alkhoury F, O’Connor JA, Kyriakides TC, Bonadies JA. ‘Normal’ vital signs belie occult hypoperfusion in geriatric trauma patients. Am Surg. 2010;76:65–69. [PubMed] [Google Scholar]
- 8.Brown JB, Gestring ML, Forsythe RM, Stassen NA, Billiar TR, Peitzman AB, Sperry JL. Systolic blood pressure criteria in the National Trauma Triage Protocol for geriatric trauma: 110 is the new 90. J Trauma Acute Care Surg. 2015;78:352–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Brown JB, Forsythe RM, Stassen NA, Peitzman AB, Billiar TR, Sperry JL, Gestring ML. Evidence-based improvement of the National Trauma Triage Protocol: The Glasgow Coma Scale versus Glasgow Coma Scale motor subscale. J Trauma Acute Care Surg. 2014;77:95–102; discussion 101–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Wasserman EB, Shah MN, Jones CM, Cushman JT, Caterino JM, Bazarian JJ, Gillespie SM, Cheng JD, Dozier A. Identification of a neurologic scale that optimizes EMS detection of older adult traumatic brain injury patients who require transport to a trauma center. Prehosp Emerg Care. 2015;19:202–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Healey C, Osler TM, Rogers FB, Healey MA, Glance LG, Kilgo PD, Shackford SR, Meredith JW. Improving the Glasgow Coma Scale score: motor score alone is a better predictor. J Trauma. 2003;54:671–678; discussion 678-680. [DOI] [PubMed] [Google Scholar]
- 12.Ross SE, Leipold C, Terregino C, O’Malley KF. Efficacy of the motor component of the Glasgow Coma Scale in trauma triage. J Trauma. 1998;45:42–44. [DOI] [PubMed] [Google Scholar]
- 13.Demetriades D, Sava J, Alo K, Newton E, Velmahos GC, Murray JA, Belzberg H, Asensio JA, Berne TV. Old age as a criterion for trauma team activation. J Trauma. 2001;51:754–756; discussion 756-757. [DOI] [PubMed] [Google Scholar]
- 14.Lehmann R, Beekley A, Casey L, Salim A, Martin M. The impact of advanced age on trauma triage decisions and outcomes: a statewide analysis. Am JSurg. 2009;197:571–574; discussion 574-575. [DOI] [PubMed] [Google Scholar]
- 15.Kehoe A, Rennie S, Smith JE. Glasgow Coma Scale is unreliable for the prediction of severe head injury in elderly trauma patients. Emerg Med J. 2015;32:613–615. [DOI] [PubMed] [Google Scholar]
- 16.Mosenthal AC, Livingston DH, Lavery RF, Knudson MM, Lee S, Morabito D, Manley GT, Nathens A, Jurkovich G, Hoyt DB, et al. The effect of age on functional outcome in mild traumatic brain injury: 6-month report of a prospective multicenter trial. J Trauma. 2004;56:1042–1048. [DOI] [PubMed] [Google Scholar]
- 17.Susman M, DiRusso SM, Sullivan T, Risucci D, Nealon P, Cuff S, Haider A, Benzil D. Traumatic brain injury in the elderly: increased mortality and worse functional outcome at discharge despite lower injury severity. J Trauma. 2002;53:219–223; discussion 223-214. [DOI] [PubMed] [Google Scholar]
- 18.Lerner EB, Willenbring BD, Pirrallo RG, Brasel KJ, Cady CE, Colella MR, Cooper A, Cushman JT, Gourlay DM, Jurkovich GJ, et al. A consensus-based criterion standard for trauma center need. J Trauma Acute Care Surg. 2014;76:1157–1163. [DOI] [PubMed] [Google Scholar]
- 19.Newgard CD, Hedges JR, Diggs B, Mullins RJ. Establishing the need for trauma center care: anatomic injury or resource use? Prehosp Emerg Care. 2008;12:451–458. [DOI] [PubMed] [Google Scholar]
- 20.Acker SN, Ross JT, Partrick DA, Nadlonek NA, Bronsert M, Bensard DD. Glasgow motor scale alone is equivalent to Glasgow Coma Scale at identifying children at risk for serious traumatic brain injury. J Trauma Acute Care Surg. 2014;77:304–309. [DOI] [PubMed] [Google Scholar]
- 21.Beskind DL, Stolz U, Gross A, Earp R, Mitchelson J, Judkins D, Bowlby P, Guillen-Rodriguez JM. A comparison of the prehospital motor component of the Glasgow coma scale (mGCS) to the prehospital total GCS (tGCS) as a prehospital risk adjustment measure for trauma patients. Prehosp Emerg Care. 2014;18:68–75. [DOI] [PubMed] [Google Scholar]
- 22.Chou R, Totten AM, Carney N, Dandy S, Fu R, Grusing S, Pappas M, Wasson N, Newgard CD. Predictive Utility of the Total Glasgow Coma Scale Versus the Motor Component of the Glasgow Coma Scale for Identification of Patients With Serious Traumatic Injuries. Ann Emerg Med. 2017;70:143–157 e146. [DOI] [PubMed] [Google Scholar]
- 23.Kupas DF, Melnychuk EM, Young AJ. Glasgow Coma Scale Motor Component (“Patient Does Not Follow Commands”) Performs Similarly to Total Glasgow Coma Scale in Predicting Severe Injury in Trauma Patients. Ann Emerg Med. 2016;68:744–750 e743. [DOI] [PubMed] [Google Scholar]
- 24.Totten AM, Cheney TP, O’Neil ME, Newgard CD, Daya M, Fu R, Wasson N, Hart EL, Chour R. Physiologic Predictors of Severe Injury: Systematic Review. Comparative Effectiveness Review No. 205. (Prepared by the Pacific Northwest Evidence-based Practice Center under Contract No. 290-2015-00009-I.) AHRQ Publication No. 18-EHC008-EF. Rockville, MD: Agency for Healthcare Research and Quality; April. 2018. [PubMed] [Google Scholar]
- 25.Pennsylvania Department of Health Bureau of Emgergency Medical Services. Pennsylvania Statewide Basic Life Support Protocols. Available at: https://www.health.pa.gov/topics/Documents/EMS/2020%20PA%20BLS%20Protocols.pdf Accessed: Sept 8, 2020.
- 26.Werman HA, Erskine T, Caterino J, Riebe JF, Valasek T, Members of the Trauma Committee of the State of Ohio EMSB. Development of statewide geriatric patients trauma triage criteria. Prehosp Disaster Med. 2011. ;26:170–179. [DOI] [PubMed] [Google Scholar]
- 27.Brown JB, Stassen NA, Bankey PE, Sangosanya AT, Cheng JD, Gestring ML. Mechanism of injury and special consideration criteria still matter: an evaluation of the National Trauma Triage Protocol. J Trauma. 2011;70:38–44; discussion 44-35. [DOI] [PubMed] [Google Scholar]
- 28.Moore L, Hanley JA, Turgeon AF, Lavoie A, Emond M. A multiple imputation model for imputing missing physiologic data in the national trauma data bank. J Am Coll Surg. 2009;209:572–579. [DOI] [PubMed] [Google Scholar]
- 29.Haider AH, Saleem T, Leow JJ, Villegas CV, Kisat M, Schneider EB, Haut ER, Stevens KA, Cornwell EE, 3rd, MacKenzie EJ, et al. Influence of the National Trauma Data Bank on the study of trauma outcomes: is it time to set research best practices to further enhance its impact? J Am Coll Surg. 2012;214:756–768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Moore L, Hanley JA, Lavoie A, Turgeon A. Evaluating the validity of multiple imputation for missing physiological data in the national trauma data bank. J Emerg Trauma Shock. 2009;2:73–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table A.1: Summary of EMS variables in Complete Cases versus Multiple Imputed Cases
Table A.2: Patient Demographics and Prehospital Characteristics, by Data Completeness
Table A.3: TBI Subtypes in Geriatric and Adult Cohorts
Table A.4: GCS Deficits in Geriatric Cohort with prehospital GCS≤13 and single deficit, stratified by CVA/ Dementia
Table A.5: Presence of CVA/ Dementia & Outcomes in Geriatric Cohort
Table A.6: Patient Demographics, Prehospital Characteristics and Outcomes in Geriatric Cohort, Stratified by Age Subgroups
Table A.7: GCS Deficits in Geriatric Cohort with prehospital GCS ≤13, stratified by Age Subgroups
Table A.8: Diagnostic Performance: NFTG Predicting TCN using GCS Total versus GCS Motor—Sensitivity Analysis in Geriatric Patients by altered TCN composite