The influence of sociodemographic factors on trauma center transport for severely injured older adults

Abstract

Objective

To examine sociodemographic predictors of trauma center (TC) transport of severely injured older adults.

Data Sources

The data source was the Healthcare Cost and Utilization Project, New York Inpatient Database (2014).

Study Design

This study was a secondary analysis of injured older adults. Key sociodemographic variables were age, gender, race/ethnicity, median household income, and primary payer. Confounding variables were injury severity, geographic location, number of chronic conditions, and injury mechanism. The outcome variable was TC transport.

Data Collection/Extraction Methods

The database was filtered on the following criteria: age =/> 55 years, primary diagnosis of injury (International Classification of Diseases, 9th Revision, Clinical Modification [ICD‐9‐CM], 800.0‐957.9, excluding poisoning, late effects, and interfacility transfers), admitted to an acute care hospital in New York.

Principal Findings

Records of 33 696 patients were included. Multivariate logistic regression analysis revealed that all variables were statistically significant predictors of TC transport except primary payer. Predictors of TC transport were as follows: higher injury severity (OR 2.1, CI 1.79‐2.46; 3.39, CI 2.85‐4.05); Asian/Pacific and Hispanic race/ethnicity (OR 2.51, CI 1.92‐3.27; OR 1.1, CI 0.86‐1.42), highest median household income (OR 1.24, CI 1.01‐1.52), high population density (OR 1.32, CI 1.12‐1.55; OR 3.2, CI 2.68‐2.83), and vehicle crashes (OR 3.39, CI 2.79‐4.11). Predictors of non‐TC transport were as follows: older age groups (OR 0.92, CI 0.76‐1.11; OR 0.79, CI 0.64‐0.96; OR 0.77, CI 0.63‐0.95), females (OR 0.65, CI 0.57‐0.74), Black and “other” race (OR 0.75, CI 0.0.56‐1.0; OR 0.96, CI 0.77‐1.20), lower median household income (OR 0.76, CI 0.62‐0.93; OR 0.86, CI 0.71‐1.05), low population density (OR 0.96, CI 0.67‐1.36; OR 0.89, CI 0.53‐1.51), and number of chronic conditions (OR 0.89, CI 0.87‐0.91).

Conclusions

Sociodemographic factors are a source of disparity for access to TCs. Further research is needed to confirm bias and test bias reduction strategies. Comprehensive education and policies are needed to reduce disparities in access to trauma care.

What this study adds.

• Evidence‐based prehospital trauma triage guidelines exist and are periodically updated as new evidence emerges. Prehospital trauma undertriage of older adults has been and continues to be a prevalent and concerning problem despite the existence of triage guidelines.

• This study adds a focus on sociodemographic factors that might bias prehospital trauma triage.

• All of the variables in this study except health insurance significantly predicted trauma transport destinations (trauma center/nontrauma center).

• Variables that predicted an increased likelihood of trauma center transport included male gender, Asian/Pacific Islander and Hispanic race/ethnicity, highest median household income for zip code, geographic location in a region with high population density, injury mechanism motor vehicle crash, and injury severity.

1. INTRODUCTION

Because of the sudden and severe nature of many injuries, emergency responders must act quickly, often with limited information about the patient. Emergency Medical Services (EMS) providers’ triage decision making is guided by a written algorithm, reflective of the national prototype Field Triage Decision Scheme,1 and supplemented by physician communication to determine transport destination. Sadly, an ample body of research has demonstrated that undertriage of severely injured older adults has been persistent.2, 3, 4, 5, 6, 7, 8 Further complicating triage decision making is the knowledge that older patients have higher injury morbidity and mortality9 and frequently present with occult injuries that are undetected at the injury scene.10, 11 Evidence from earlier studies suggests multiple causes of undertriage, including inadequate sensitivity of the algorithms to injury,12 decision making based on EMS provider judgment,8, 13, 14 inadequate training or knowledge among EMS providers,8 age bias,8, 15 and patient preference for a specific NTC hospital.16 Additional factors that influence triage decision making are weather conditions, distance to the nearest TC,17 and available transport resources for travel to a more distant TC.

As complex and problematic as the prehospital triage process is, we must be assured that the process is free of bias and promotes health equity with regard to TC care. Health equity, defined as fairness with regard to providing equal opportunities for each person to attain full health potential,18 should be present in every aspect of health care. However, biased decision making often leads to health inequity, resulting in health disparities.19

As humans, we archive past interactions and experiences with people of various social groups. These interactions and experiences contribute implicitly or explicitly to our thoughts (stereotypes) and feelings (prejudices), and influence our behaviors, including decision making.19 Individuals with explicit biases are aware of their biases; however, implicit biases are unrecognized.20 The stereotypes sometimes fade into our unconscious memory and are activated in a variety of circumstances, including fatigue, feeling overwhelmed, or required to make decisions quickly, often with little information.19 EMS providers routinely encounter such situations, potentially leading to biased trauma transport decisions. Other times, stereotypes remain conscious and knowingly influence our perceptions of others.

Ideally, EMS providers should follow an up‐to‐date algorithm for trauma patient triage, but previous research suggests this is often not the case.13, 14 The current FTDS1 four‐step algorithm allows consideration of patient age‐associated risk factors and EMS provider judgment in the final step to determine transport destination. This step of the algorithm sets the stage for activation of stereotypes, which may lead to biased decision making and limit access to TC care. This study integrates an implicit‐explicit bias decision making model as a factor in EMS judgment in this final step of the FTDS.

Our conceptual model of biased decision making, adapted from Blair et al,20 illustrates how bias might affect EMS provider judgment or patient response in the final step of the FTDS algorithm. In this model, biases and background experiences of EMS providers and patients are interrelated. Their backgrounds and experiences affect their behaviors and judgments to an extent that the EMS provider might decide against TC transport for a severely injured patient or a severely injured patient might refuse TC transport. Importantly, situations that allow for judgment calls by health care providers allow bias to influence decision making.20, 21 In this study, sociodemographic factors, including age, gender, median household income, race/ethnicity, and health insurance status, are examined as they related to TC transport. These variables were added to the model based on published studies and other documents indicating possible biases contributing to health disparities.7, 8, 18, 20, 22, 23, 24 The sociodemographic variables are nonmodifiable, whereas bias and the triage algorithm are modifiable. This model was integrated into the FTDS, step 4, EMS judgment condition. Both implicit and explicit bias related to sociodemographic factors were viewed as factors that might influence EMS provider judgment.

Although the literature is replete with studies reporting emergency care services and patient injury characteristics that contribute to undertriage, less is known about the influence of patient sociodemographic factors and whether they elicit a biased transport decision, leading to disparities in the access to TC care. Previous investigators reported significant age and sex differences among patients triaged to TCs and NTCs, with an abundance of younger men triaged to TCs and older women triaged to NTCs.3, 5, 8, 15, 24, 25 Two studies of EMS undertriage in Maryland revealed an association between age and TC transport. In the earlier study,15 patients age 55 and older who met only physiology or mechanism of injury criteria were less likely to be transported to TCs compared to younger patients. In a subsequent mixed methods study to determine whether age bias was a factor in triage decision errors, EMS providers ranked possible age bias as one of the top three causal factors leading to undertriage.8 Additional factors for undertriage included a lack of understanding about elderly trauma how to use the triage algorithm, and believing that “transport to a TC was not worth it.” After ruling out alternative explanations for undertriage, the investigators concluded that age bias likely resulted in older adults with life‐threatening injuries being 50 percent less likely to be transported to a TC than younger adults.8 In a more recent study of race and social class bias among 248 trauma surgeons,26 74 percent reported implicit race bias that favored White persons and 91 percent reported implicit class bias that favored upper social class persons.

An examination of racial and ethnic differences in access to TC care in Los Angeles, Chicago, and New York City,27 revealed that certain Black census tracts in New York City had greater odds of being in trauma deserts, compared to White census tracts. In contrast, Hispanic and Latino census tracts in New York City and Los Angeles were less likely to be located in trauma deserts compared to those in Chicago. In a recent study of trauma patients, investigators reported that increased age and minority race predicted undertriage.28 Several additional nontriage studies of Black trauma patients revealed disparities among racial minorities. In a secondary analysis of brain‐injured adults, investigators found that insured and uninsured Black, Asian, and Hispanic patients, compared to insured White patients, had decreased odds of discharge to a rehabilitation facility following acute care.29 Investigators, in another secondary analysis of race and mortality, reported that Black and Asian patients had 50‐100 percent greater odds of death compared to White patients. The differences were most notable among Black patients with moderate injuries and Asian patients with severe injuries.30

In two EMS provider judgment studies of trauma patients, investigators reported that EMS provider judgment was used either in combination with other triage criteria or as the sole criterion to determine triage transport destination. In the earlier study, provider judgment was used in 36 percent of all field trauma activations, including 23 percent of cases in which only provider judgment was considered to decided transport destination.14 In a subsequent study, EMS provider judgment accounted for 40 percent of the decisions to transport triage‐positive patients to TCs, often in conjunction with criteria from earlier steps. However, for 29 percent of these patients, provider judgment was the sole triage criterion used.13

Given the paucity of studies examining sociodemographic factors and access to TC care, prehospital decision making, and the self‐reported bias of one group of EMS providers, the purpose of this study was to examine sociodemographic factors associated with TC transport among injured older adults.

2. METHODS

This study was an analysis of data from the Agency for Healthcare Research and Quality (AHRQ) Healthcare Utilization Project (HCUP), New York State Inpatient Database (SID) 2014 (January 1 through December 31).31 Inclusion criteria were as follows: age 55 years and older, primary diagnosis of injury (International Classification of Diseases, 9th Revision, Clinical Modification [ICD‐9‐CM], 800.0‐957.9, excluding poisoning and late effects), admitted to an acute care hospital in New York. Also excluded were interfacility transfers (patients transferred to a TC from another acute care hospital). Sociodemographic variables of interest included age, gender, race/ethnicity, median household income by state quartile for ZIP code (a proxy for socioeconomic status), and primary payer (health insurance). Race/ethnicity was defined as White, Black, Hispanic, Asian/Pacific Islander, Native American, and other (mixed race). Median household income quartiles 1 through 4 ranked income from the lowest (1) to the highest (4), within each zip code. Primary payer was designated as Medicare, Medicaid, private (commercial) insurance, other government insurance, and uninsured (self‐pay/charity). Other government insurance included Veterans Affairs (VA) Health Care, Tricare, workers’ compensation, and additional government insurance programs. Injury mechanism was categorized as motor vehicle traffic, fall, cutting, drowning, firearm, burns, nature/environmental, struck, and suffocation.

Composite variables were created to designate TC admission (yes/no), mechanism of injury (based on E‐codes), maximum injury severity score 1 [MAIS 1], maximum injury severity score 2 [MAIS 2], maximum injury severity score 3 [MAIS 3], and overall injury severity using the New Injury Severity Score [NISS].

Because the SID 2014 database does not include injury scores, an injury score mapping program32 was used to translate ICD‐9‐CM codes to injury severity scores. Using this mapping program, injury severity was determined for six body regions (head/neck, chest, abdominal and pelvic contents, extremities and pelvic girdle, face, and external) by assigning an Abbreviated Injury Score (AIS) to each ICD‐9‐CM injury diagnosis, 800‐957.32 The AIS for each body region injury was scored from 1 to 6, where 1 = minor injury and 6 = currently untreatable/probably nonsurvivable injury. A total injury severity score, the New Injury Severity Score (NISS),33 was calculated by summing the squares of the three highest AISs (MAIS 1, MAIS 2, MAIS 3) for each patient. An AIS of 6 is automatically recorded as a NISS of 75. AIS and NISS are ordinal scales. NISSs range from 1 (minor injury) to 75 (fatal injury). NISSs of 16 and higher denoted severe, critical, and probable nonsurvivable injuries. This dataset accommodates as many as 25 diagnoses for each patient record; therefore, each diagnosis was queried for an injury, which was then mapped to body region and severity.

2.1. Data analysis

R, version 3.6.1, statistical software34 was used for all analyses, with P < .05 indicating statistical significance. Frequency, proportion, mean, standard deviation, median, and interquartile range (IQR) were calculated for the variables of interest and confounders (Table 1). Based on frequency distributions, several variables were recoded because of very low frequencies in some of the levels. Race/ethnicity was collapsed to include Native Americans in the other category. The primary payer category was collapsed to include charity care in the other category. NISS (injury severity) was collapsed to include unmapped (nonspecified) injuries into the minor category and maximal (unsurvivable) into the severe category. Mechanism of injury was collapsed into three categories: falls, motor vehicle traffic, and other mechanisms.

Table 1.

Sociodemographic characteristics of patients transported directly to trauma centers and nontrauma centers

Variable	Median (IQR)
Age	80 (69‐87)
Injury severity	9 (5‐9)
Number Chronic Conditions	5 (3‐7)

	Frequency (%)
Gender
Females	21 287 (63.2)
Median household income quartile categories
1st (poorest)	5754 (17.9)
2nd	7877 (24.5)
3rd	8668 (27.0)
4th (wealthiest)	9825 (30.6)
Primary payer
Medicare	25 637 (76.1)
Medicaid	1749 (5.2)
Private (commercial insurance)	4982 (14.8)
Self‐pay	459 (1.4)
No charge	(<0.1)*
Other (military, government, etc)	866 (2.6)
Race/ethnicity
White	25 272 (75.0)
Black	2305 (6.8)
Hispanic (ethnicity)	2301 (6.8)
Asian/Pacific Islander	990 (2.9)
Native American	49 (0.1)
Other race (including mixed race)	2779 (8.2)

Abbreviations: IFT, interfacility transfers to TCs; NTC, nontrauma center; TC, trauma center.

^*Frequences less than 10 not reported.

A multivariate logistic regression analysis was performed to predict the odds of TC transport as a function of the sociodemographic variables of age, gender, race/ethnicity, median household income, and primary payer, as well as potential confounders NISS injury severity, number chronic conditions, geographic location, and injury mechanism (Table 2). The number of chronic conditions was treated as a continuous variable in the logistic regression model. All other predictors entered as categorical variables.

Table 2.

Sociodemographic predictors of trauma center transport for severely and critically injured older adults

Variable	Category	N	No. (%) TC transport	Adjusted odds ratio (95% CI)a
Age (P = .045)	55‐64	5216	365 (7)	[reference]
	65‐74	6050	271 (4.5)	0.92 (0.76‐1.11)
	75‐84	9188	275 (3)	0.79 (0.64‐0.96)
	≥85	11 670	284 (2.4)	0.77 (0.63‐0.95)
Gender (P < .0001)	Male	11 661	625 (5.4)	[reference]
	Female	20 463	570 (2.8)	0.65 (0.57‐0.74)
Race (P < .0001)	White	24 637	866 (3.5)	[reference]
	Black	2001	62 (3.1)	0.75 (0.56‐1)
	Hispanic	1958	80 (4.1)	1.1 (0.86‐1.42)
	Asian/Pacific Islander	887	82 (9.2)	2.51 (1.92‐3.27)
	Other (mixed) + Native American	2641	105 (4)	0.96 (0.77‐1.2)
Median household income (P < .0001)	1st (poorest)	5754	210 (3.6)	[reference]
	2nd	7877	227 (2.9)	0.76 (0.62‐0.93)
	3rd	8668	291 (3.4)	0.86 (0.71‐1.05)
	4th (wealthiest)	9825	467 (4.8)	1.24 (1.01‐1.52)
Primary payer (P = .099)	Medicare	24 683	631 (2.6)	[reference]
	Medicaid	1339	69 (5.2)	1.32 (0.98‐1.77)
	Private	4813	428 (8.9)	1.17 (0.96‐1.43)
	Self‐pay	446	22 (4.9)	0.91 (0.57‐1.46)
	Other + no charge	843	45 (5.3)	0.86 (0.6‐1.22)
NISS injury severity (P < .0001)	Nonspecified and minor	24 754	611 (2.5)	[reference]
	Moderate	4115	249 (6.1)	2.1 (1.79‐2.46)
	Severe	1875	217 (11.6)	3.39 (2.85‐4.05)
	Critical	1380	118 (8.6)	2.63 (2.11‐3.26)
Geographic location (P < .0001)	Central counties of large metro areas	14 002	412 (2.9)	[reference]
	Fringe counties of large metro areas	10 648	456 (4.3)	1.32 (1.12‐1.55)
	Counties in metro areas 250k‐999k	3476	242 (7)	3.2 (2.68‐3.83)
	Counties in metro areas 50k‐249k	1505	31 (2.1)	0.99 (0.68‐1.45)
	Micropolitan counties	1781	38 (2.1)	0.96 (0.67‐1.36)
	Not metro or micro (rural)	712	16 (2.2)	0.89 (0.53‐1.51)
Injury type (P < .0001)	Falling	26 632	743 (2.8)	[reference]
	Motor vehicle/traffic	1995	337 (16.9)	3.39 (2.79‐4.11)
	Other	3497	115 (3.3)	0.89 (0.72‐1.1)
Number chronic conditionsb (P < .0001)				0.89 (0.87‐0.91)

Median household income—Median household income for zip code. Payer—Primary payer, other government includes Veterans Affairs (VA) Health Care, Tricare, workers’ compensation, and additional government insurance programs.

^aAdjusted odds ratio, 95% confidence interval (CI), and p‐values are based on a multivariate logistic regression model predicting TC transport, including all predictors in the table. Interfacility transfers excluded. Reference—reference category for odds ratio.

^bIncluded as a continuous variable in the model; all other predictors treated as categorical.

3. RESULTS

A total of 33 696 severely injured patients meeting inclusion and exclusion criteria were admitted to TCs and NTCs directly from the injury scene. Total TC admissions equaled 1211 patients (3.6 percent). Age of the sample ranged from 55 to 114 years, with a mean age of 78.1 (SD = 11.6 years) and a median age of 80 (IQR = 69‐87 years). Injury severity ranged from 1 to 75, with a median score equal to 9 (IQR = 5‐9). Sociodemographic characteristics are depicted in Table 1.

The multivariate logistic regression analysis showed that all variables were statistically significant predictors of TC transport except for primary payer (Table 2). Females were less likely to be transported to TCs. Most racial/ethnic groups except Asian/Pacific Islander revealed a similar tendency for TC transport; Asian/Pacific Islanders were appreciably more likely to be transported to TCs, whereas Blacks were less likely to be transported to TCs.

An increasing tendency for TC transport was observed for median income categories above the lowest; the second lowest was significantly less likely to be transported to TCs relative to the lowest, and the highest category was significantly more likely relative to the lowest. An increasing tendency for TC transport was seen among the first three injury severity categories; all categories beyond the first had a significantly elevated chance of TC transport relative to the first. With respect to geographic location, the second most densely populated category had a higher likelihood of TC transport relative to the densest populated but the most striking feature is that the middle category (counties in metro areas 250 000‐999 000 population) had an appreciably higher likelihood relative to all other categories. The injury mechanism involving motor vehicle crash was statistically significantly more likely to lead to TC transport relative to the other two categories (falling, other). The number of chronic conditions was strongly negatively related to TC transport, and age showed a modest negative association. Primary payer was not statistically significant.

4. DISCUSSION

This study focused on whether or not sociodemographic variables predicted TC transport in an attempt to understand whether bias influenced these decisions. The FTDS1 provides a systematic process for identifying patients who should be transported to TCs. Moving sequentially through the four steps of the algorithm, failure at each step to identify the need for TC transport leads to the next step. Finally, at the last step, EMS providers are reminded of special considerations for older adults that might warrant TC transport. Also, included in this last step is the allowance for EMS provider judgment. None of the sociodemographic variables are considered in steps 1, 2, and 3 of the FTDS. However, Step 4 includes two cautionary statements regarding older adults that are pertinent to this study: risk of injury or death increases after age 55 and low impact mechanisms (ground‐level falls) might result in severe injury. Two problems arise with the algorithm; first, in focus group settings, EMS providers in three studies admitted to using their own judgment solely or in combination with one or more steps of the algorithm.8, 13, 14 The second problem is that not all EMS providers recognize the existence of an algorithm or know how to use it correctly.8, 13

Our results revealed that females and older patients are less likely to be transported to TCs, which is consistent with previous findings.3, 24, 25, 28 However, some of our findings are remarkable and not easily explained. As the number of chronic conditions increased, the likelihood of TC transport decreased, although the effect of age was moderate. Previous studies have noted a link between the increasing prevalence of chronic disease and increased frailty and mortality,35 but these do not explain transport decisions to NTCs. Certain chronic diseases have been associated with the likelihood of TC transport; however, no studies were found that associated the total number of chronic diseases with TC transport.28, 36 Nonetheless, the association of the number of chronic diseases and triage transport decisions is not easily explained.

With regard to race/ethnicity, a larger proportion of Asian/Pacific Islander patients had more severe and critical injuries than all other racial/ethnic groups, which explains their higher likelihood for TC transport. However, the proportion of Black and Hispanic patients with severe and critical injuries was the same, so the disparity in TC transport between these groups is puzzling.

Patients in each quartile of median household income had similar proportions of severe and critical injuries, so the finding that the lowest and highest income quartiles were most likely to be transported to TCs is not easily understood. Patients in cities of 250 000‐999 000 people had the highest odds of transport to TCs, followed by those located in fringe counties of the largest metro areas (eg, New York City suburbs). Large urban areas are likely to have TCs nearby, compared to less populated areas. At the time, these data were collected; of the 38 regional and area state‐designated TCs, 16 were located in New York City. Therefore, it is difficult to explain the increased likelihood of TC transport in areas of less population densely compared to those with even higher population density, particularly because the proportions of severe injuries were similar.

The same proportions of patients who fell and those involved in motor vehicle crashes had severe and critical injuries. However, motor vehicle crashes are generally higher energy‐transfer mechanisms compared to falls, which might have biased EMS decision making in favor of TC transport for patients involved in crashes. All other mechanisms of injury accounted for less than one percent of all severe and critical injuries, so injury severity explains why these patients were less likely to be transported to TCs.

Some of our findings suggest the possibility of age, gender, racial/ethnic, and socioeconomic biases. With regard to injury care, health equity means providing equal opportunity to access TC care when needed. Our data revealed that because many patients whose injuries warranted TC transport but did not receive it, there were disparities in access to TC care for certain groups of patients that were not explained by injury severity. However, it remains unclear whether those disparities were due to EMS provider biases, patient refusal of such care, or other factors.

The sociodemographic factors we examined are nonmodifiable. However, biases, geographic distance to a TC, and the triage algorithm are potentially modifiable and should be addressed by stakeholders at all levels through education and policy. The National Standards for Culturally and Linguistically Appropriate Services in Health and Healthcare37 established comprehensive guidelines to improve health care quality and advance health equity. However, the guidelines are not prescriptive. To advance implementation of these standards, experts recommended multiple strategies that enable health care providers to recognize and reduce bias. These strategies should be considered by all EMS educational programs, agencies, and certification and licensing boards.

Emergency Medical Services provider education must include initial and ongoing content that integrates cultural competence skill training with an understanding of the social psychology that activates stereotypes and assessments of bias. Social psychologists19 developed a series of workshops designed to recondition individuals to respond differently to cues that trigger bias. In their final workshop, participants learn important strategies for preventing implicit bias that include the following: pursuing egalitarian goals, identifying common identities, counterstereotyping, and perspective taking.19 Consideration should be given to screening for implicit bias, using a valid measure, such as the implicit association test (IAT).38 The IAT helps individuals recognize their own implicit biases. Through ongoing participation in bias recognition, mitigation, and prevention activities, EMS providers should develop an awareness and control of their own biases, understanding how these biases might affect their decision making. Although strategies were identified to reduce implicit biases, none were found to address explicit biases. It is possible that the same strategies that have been used to decrease implicit biases would be effective in reducing explicit biases; however, these must be tested.

The ability of EMS providers to correctly use triage algorithms should be addressed through education. Engaging EMS students and practitioners in interactive learning, including simulation, case studies, and supervised use of the triage algorithm with diverse age groups and with various environmental conditions, should be ongoing until mastery is attained.

Geographic distribution of TCs should be addressed on national, state, and trauma network levels through policy development involving all stakeholders. Examination of the distribution of trauma resources is necessary to Identify trauma deserts, with plans to redistribute trauma resources to these underserved areas. The federal and state governments should provide funding for additional trauma resources and reallocation of existing resources.

The FTDS is an evidence‐based guideline developed by the American College of Surgeons Committee on Trauma. Since its inception, the FTDS has undergone several revisions as more evidence becomes available. The national workgroup that oversees revisions should reconsider the usefulness of the current FTDS1 with older adults. Investigators in Ohio and Florida recognized limitations in the triage algorithm and developed statewide revised algorithms. The Ohio criteria, developed for a geriatric population, changed the parameters for triage decision making in the Stage 1 criterion relative to the Glasgow Coma Scale score.39, 40, 41 A new algorithm was developed in Florida, based on research by McKenzie et al42 This algorithm identified specific age‐related requirements for trauma alerts. For patients age 55 years and older, the algorithm stipulates that patients with an AIS 3 or greater injury (except an isolated hip fracture) should be treated at a TC. Although they later determined that this algorithm was not effective in identifying which patients should be admitted to TCs, nonetheless, they took an important step to address undertriage by evaluating a new strategy, albeit unsuccessful.

Results of this study should be interpreted with caution. Although these findings reflect the diverse population, geography, health care resources, and triage protocols in New York state, they may differ in other geographic locations with different populations, geography, resources, and protocols.

It is imperative that the trauma community moves forward to eliminate disparities in TC access. More research funding is needed to address biases in prehospital care so that trauma access equity is realized for every patient. It is possible that implicit and explicit age, race, and socioeconomic biases influenced triage transport decisions in this study; however, the evidence is not conclusive. Additional research is needed to confirm biases and test the effectiveness of interventions to decrease bias and resulting health disparities. The trauma community must advocate for policy and education to promote health equity and eliminate health disparities with regard to access to TC care.

AUTHOR CONTRIBUTIONS

Linda Scheetz conceived and designed the study, acquired and prepared the data, conducted the initial data analysis, interpreted the data, drafted the original manuscript and drafted the revised manuscript. John Orazem revised the data analysis plan, conducted the revised data analysis, interpreted the data, and contributed to the revision of the manuscript. The authors received institutional support from Lehman College, and used personal funds to secure the needed data sets.

Supporting information

 

Scheetz LJ, Orazem JP. The influence of sociodemographic factors on trauma center transport for severely injured older adults. Health Serv Res. 2020;55:411–418. 10.1111/1475-6773.13270