Early use of professional interpreters improves trauma outcomes: Results of a single-center retrospective study

Abstract

Patients with limited English proficiency (LEP) experience reduced pain assessment and treatment, less comprehensive physical exams, and fewer explanations of the next steps in care. These disparities persist in hospitals with staffed professional interpreters, raising questions about interpreter access and the impact on outcomes. A retrospective review of 1133 trauma activations at a single center Level 1 Trauma Center in 2021–2022 was conducted. Demographic, injury, and outcome data were drawn from the institutional trauma registry, and patient-preferred language was pulled from EMR data. Early interpreter use was defined as documentation of professional interpreter use within 24 h of arrival. LOS and ICU LOS were compared between language groups using Cox regression, and mortality was compared using Fischer's exact test. 1114 patients had data available on initial injury severity and preferred language. Of the 70 LEP patients, 62 (88.6 %) required an interpreter, and 41 of those (66.1 %) had evidence of professional interpreter use within 24 h of arrival. LEP patients who lacked early interpreter use had longer hospital stays than both English proficient (EP) patients (HR 0.59, p < 0.05) and LEP patients with early interpreter use (HR 0.51, p < 0.05) when stratified by ISS and controlling for GCS and patient age. There is no difference in LOS between LEP trauma patients who used an interpreter and EP patients, suggesting that early use of an interpreter may improve the length of stay in LEP trauma patients.

Graphical abstract

Highlights

• •Despite U.S. law requiring interpreters in healthcare, they are not always made available.
• •Trauma patients without interpreter access have longer hospital stays.
• •With interpreter access, length of stay is equivalent to that of English-speaking patients.

Introduction

Over the last several years, language barriers have garnered increasing attention as determinants of care quality and outcomes in hospitalized patients. Previous studies have shown that patients with limited English proficiency (LEP) experience improved care quality and outcomes with the use of professional interpreters [1].

Despite the documented positive association between professional interpretation and outcomes, many studies continue to document a lack of use of interpreters and numerous disparities in care for LEP patients with traumatic injuries. LEP patients have been found to experience fewer pain assessments and treatment [2,3], and are more likely to have poorly conducted initial care [4,5]. Furthermore, significant differences in initial care provided to trauma patients based on primary language have been observed and may be correlated with poor trauma outcomes. Spanish-speaking patients (SSP) in one study received less comprehensive motor and sensory examinations, as well as less frequent explanations of next steps in care or reassurances from providers, in comparison to their EP counterparts [5]. These disparities in care were observed in hospitals staffed with professional interpreters: one study revealed that LEP patients had fewer pain assessments per day and lower pain scores than English proficient (EP) patients despite the presence of interpreters [2]. These findings suggest that LEP patients do not receive high quality healthcare despite the presence of interpreters, which indicates that staffing professional interpreters may not be sufficient to resolve disparities experienced by patients with limited English proficiency.

The purpose of this study is to evaluate the effect of interpreter utilization on LEP patient trauma outcomes. Our primary analysis examined the relationship between LEP and hospital length of stay (LOS), ICU LOS, and mortality. Our secondary analysis evaluated whether provision of a professional interpreter to LEP patients within the first 24 h of trauma presentation to the hospital impacted mortality, hospital LOS, and ICU LOS. We hypothesized that LEP trauma patients who presented to the hospital and accessed a professional interpreter in the first 24 h would have reduced mortality rates, LOS, and ICU LOS in comparison to LEP patients who did not.

Materials and methods

We conducted a retrospective cohort study of patients at a single Level 1 Trauma Center in a mid-sized city in the southern United States following one or more traumatic injuries within 14 days of initial hospital encounters between 2021 and 2022. The study size of 1133 patients includes all trauma patients in 2021 and 2022. Chart review was performed from October 2023 through January 2024. Patients were selected using the National Trauma Data Standard and ICD-10-CM codes, which standardize data considering injury of external causes [6]. Patients that sustained a traumatic injury (represented by the ICD-10-CM range of S00–S99, T07, T14, T79.A1–A9) and at least one non-superficial injury (outside the ranges of S00, S10, S20, S30, S40, S50, S60, S70, S80, and S90) were included in the study. Patients who had injuries that resulted in death, who were transferred from one acute care hospital to another acute care hospital, and who were admitted for inpatient observation were all included in the study. Patients were excluded if they were admitted for elective surgeries, admitted directly without a trauma activation, had isolated superficial injuries not requiring admission, under the age of 18, or over the age of 64. Demographic factors, method of injury, work-relatedness, trauma severity based on Injury Severity Score (ISS), initial Glasgow Coma Score (GCS), length of stay (LOS), Intensive Care Unit length of stay (ICU LOS), survival probability per the Trauma and Injury Severity Score model, and trauma team activation level were all reviewed. Patient identity was confirmed using MRN number, date of birth, and date of admission. Limited English proficiency was defined as the documentation of a preferred or spoken language other than English in the demographics section of the electronic medical record. Identification of a patient as limited English proficiency initiated the chart review process for interpretation details.

All notes with a time of service within 24 h of arrival were reviewed for evidence of interpreter use. “Early interpreter use” was defined as the documented use of a certified interpreter, either in-person or via virtual interpreter services, within 24 h. If the patient was unresponsive on arrival and did not regain consciousness within the first 24 h, “early interpreter use” was defined as the documented use of a certified interpreter to communicate with friends or family members. “No early interpreter use” was defined as no documentation of certified interpreter use to communicate with the patient, or family and friends if the patient was unconscious, within 24 h. If only a family member, friend, or uncertified staff member was used, this was considered “no early interpreter use.”

Morbidity outcomes included ICU length of stay and hospital length of stay. Mortality was analyzed using death during hospitalization. Each patient was followed through discharge (mean 4 days). Three “language access” groups were created for comparison: English proficient (EP) patients, limited English proficiency (LEP) patients with early interpreter use, and LEP patients with no early interpreter use. Nineteen patients with incomplete outcome data and eight patients who had languages other than English listed in their chart but were recorded as requiring “no interpreter” in the demographic section were excluded from the analysis, as shown in Fig. 1.

Fig. 1.

Inclusion criteria and study groups.

Cox regressions were used to compare hospital and ICU LOS between groups. GCS on arrival and patient age were included as covariates; the model was stratified on ISS in order to meet the proportional hazards condition. These variables were selected to control for any differences in baseline clinical status between groups. Mortality analysis was performed using Fisher's exact test.

In accordance with 45 CFR 46.104(d), this study is exempt from Human Research Subject Regulations. The IRB has also approved the request for a Waiver of HIPAA Authorization, after determining that the waiver of authorization satisfies the criteria set forth in the HIPAA Privacy Rule at 45 CFR 164.512(i)(2). The waiver of authorization was reviewed and approved by the IRB. The IRB has determined that the protected health information necessary to be used and accessed is as outlined in the IRB application.

Results

There were a total of 1133 trauma activations among patients ages 18–64 leading to admission at our Level 1 trauma center in 2021 and 2022. 1114 of these patients had available data on key information related to initial injury severity and preferred language. 70 patients (6.3 %) spoke a language other than English; 63 (5.7 %) were Spanish speakers. Of these, 62 (88.6 %) required an interpreter and were thus classified as “limited English proficiency,” and 41 (66.1 %) had documented evidence of professional interpreter use within 24 h of arrival (Table 1).

Table 1.

Cohort demographics (sex, language, age).

	English proficient (n = 1044)	Limited English proficiency with interpreter (n = 42)	Limited English proficiency without interpreter (n = 20)	p-Value
Age Mean (std)	39.6 (14.5)	36.1 (13.0)	34.0 (10.4)	0.086†,ns
Preferred Language Count (%)				<0.001‡,⁎⁎⁎
English	1044 (100)	0 (0)	0 (0)
Spanish	0 (0)	38 (90.5)	19 (95)
Other	0 (0)	4 (9.5)	1 (5)
Male Count (%)	744 (71.3)	35 (85.4)	19 (90.5)	0.024§,⁎

^§ Pearson's Chi-Square.

^‡ Fisher's Exact Test.

^† Kruskal-Wallis Rank Sum Test.

^⁎ Significant at p < 0.05.

^⁎⁎ Significant at p < 0.01.

^⁎⁎⁎ Significant at p < 0.001.

The cohort was majority male (n = 806, 72.4 %). The average age was 39.4 years (std dev 14.4 years). 555 (49.8 %) patients were White and 463 (49.8 %) of patients were Black. 88 patients were documented as “Other Race”; of these, it is notable that all 69 patients of Hispanic or Latino ethnicity were documented to be “Other Race” (Table 1).

As documented in Table 2, the most common mechanism of injury was motor vehicle collisions (374, 33.6 %), followed by falls (196, 17.6 %) and gunshot wounds (157, 14.1 %). 67 (6.0 %) injuries were work-related. 428 (38.4 %) of injuries were classified as major trauma by an Injury Severity Score of 16 or greater [7]. 123 (11.0 %) injuries involved a major traumatic brain injury as classified by Glasgow Coma Score of 3–8 [8]. The median length of stay was 4 days (IQR 2–9 days). 481 (43.2 %) patients were admitted to the ICU. 29 (2.6 %) patients died.

Table 2.

Cohort injury characteristics.

	English proficient (n = 1044)	Limited English proficiency with interpreter (n = 42)	Limited English proficiency without interpreter (n = 20)	p-Value
ISS Median (IQR)	13 (10)	13 (8)	21 (20.8)	0.016†,⁎
GCS Count (%)				0.075†,ns
3–8	30 (2.9)	1 (2.4)	1 (5)
9–12	110 (10.5)	5 (11.9)	6 (30)
13–15	904 (86.6)	36 (85.7)	13 (65)
Work-related Count (%)	41 (3.9)	13 (31.0)	10 (50)	<0.001§,⁎⁎⁎
Admitted to ICU Count (%)	446 (42.7)	15 (35.7)	15 (75)	0.01§,⁎⁎
ICU LOS Median (IQR)	3 (5)	3 (1)	4 (8)	Pairwise, see Table 4
LOS Median (IQR)	4 (7)	4 (4)	9.5 (13.8)	Pairwise, see Table 3
Mortality Count (%)	26 (2.5)	1 (2.4)	2 (10)	0.14‡,ns

^§Pearson's Chi-Square.

^‡Fisher's Exact Test.

^†Kruskal-Wallis Rank Sum Test.

^⁎Significant at p < 0.05.

^⁎⁎Significant at p < 0.01.

^⁎⁎⁎Significant at p < 0.001.

Hospital LOS was significantly longer for LEP patients without early access to an interpreter (median 9.5 days) compared to both LEP patients with early access to an interpreter (median 4 days) and EP patients (median 4 days). This is demonstrated by hazard ratios for discharge of 0.51 (CI 0.29–0.90) and 0.59 (CI 0.37–0.95), respectively. The difference was maintained when controlling for age, ISS, and GCS on admission. The LEP with early interpreter access group did not demonstrate a different length of stay versus the EP group (hazard ratio of discharge 1.68, 95 % CI 0.97–2.92). These hospital length of stay findings are summarized in Table 3. There was no significant difference in ICU LOS between groups after adjusting for age, ISS, and GCS, as summarized in Table 4. There was no difference in in-hospital mortality between groups (Table 2).

Table 3.

Hazard ratios of discharge.

	English proficient (n = 1044)			LEP with interpreter (n = 42)			LEP without interpreter (n = 20)
	Median ratio	Raw HR	Adj HR	Median ratio	Raw HR	Adj HR	Median ratio	Raw HR	Adj HR
English proficient					1.27 (0.93–1.73)	1.15 (0.84–1.58)	2.13	0.49⁎⁎⁎ (0.31–0.79)	0.59⁎ (0.37–0.95)
LEP with interpreter	1	0.79 (0.58–1.08)	0.87 (0.63–1.19)				2.13	0.39⁎⁎⁎ (0.22–0.68)	0.51⁎ (0.29–0.90)

^⁎Significant at p < 0.05.

^⁎⁎⁎Significant at p < 0.001.

Table 4.

Hazard ratios of transfer from ICU.

	English proficient (n = 1044)			LEP with interpreter (n = 42)			LEP without interpreter (n = 20)
	Median ratio	Raw HR	Adj HR	Median ratio	Raw HR	Adj HR	Median ratio	Raw HR	Adj HR
English proficient					1.65 (0.96–2.82)	1.68 (0.97–2.92)	1.33	0.76 (0.43–1.33)	1.01 (0.58–1.76)
LEP with interpreter	1	0.61 (0.35–1.03)	0.59 (0.34–1.03)				1.33	0.46⁎⁎⁎ (0.22–0.98)	0.60⁎ (0.28–1.29)

^⁎Significant at p < 0.05.

^⁎⁎⁎Significant at p < 0.001.

Discussion

This study adds important data to the growing body of literature concerning care of LEP patients. Given that limited English proficiency has been associated with disparities in care [[2], [3], [4], [5],[9], [10], [11], [12]], we hypothesized that EP patients at our academic medical center would have better outcomes than LEP patients. In accordance with the Civil Right Act of 1964 and Executive Order 13166, which requires institutions provided with federal funding to use interpreters in healthcare, our institution offers interpreter services for over 240 languages free of charge to all LEP patients during their medical treatment and hospitalization. These services are available by telephone and through video remote interpretation 24 h a day. We found that while LEP patients with early access to professional interpreter services had the same length of stay (LOS), intensive care unit (ICU) LOS, and mortality as EP patients. However, LEP patients without early access to a professional interpreter had longer ICU and hospital length of stay than EP patients. This difference persisted when controlling for age, Injury Severity Score (ISS), and Glasgow Coma Scale (GCS) on admission. As we hypothesized, early provision of trauma care in a patient's preferred language improves outcomes. Of our LEP cohort, 66.1 % had documented professional interpreter use within 24 h of arrival, indicating significant room for improvement in providing language-concordant trauma care.

The demographic breakdown of our study cohort aligned with several previous studies [1,[9], [10], [11], [12]]. Our LEP cohort was majority young, male, and Hispanic, and the preferred language was Spanish. One prior study followed a mostly male cohort with an average age of 45.13 years [12], while two other studies were comprised primarily of older women [2,11]. These studies were all composed of a majority Latino or Hispanic patients [2,11,12]. Our findings also resembled those of previously studied cohorts in terms of language preference, with Spanish being the most, or one of the most, preferred languages [2,[10], [11], [12]]. We anticipated that our cohort with similar demographics may reveal similar LOS, ICU LOS, and mortality outcomes.

The most common mechanism of injury (MOI) for our cohort was motor vehicle collisions, followed by falls and gunshot wounds. A small percentage were work-related. One previous study found that LEP patients had a greater percentage of work-related injuries compared to EP patients [12], which was true in our cohort (37 % in LEP patients vs. 3.9 % in EP patients, p < 0.001). Our study found that 11 % of injuries involved a major traumatic brain injury (TBI) (GCS < 9), which was similar to studies that documented high median GCS, indicating that a high proportion of traumas present without major TBI [2,11].

While EP patients and LEP patients with early interpreter access had the same length of stay (median ratio 1), LEP patients without early interpreter access had a longer length of stay (median ratio 2.13, p < 0.05) than both groups. The median ICU LOS was longer for LEP patients without an interpreter (4 days) versus LEP patients with an interpreter (3 days); however, this difference did not persist when controlling for ISS, GCS, and patient age. Two larger studies (n = 12,562 [10] and n = 10,375 [12]) observed a lack of statistically significant differences in LOS and ICU LOS between EP and LEP patients and additionally did not observe any differences in 30-day readmission rates. However, Castro et al. (n = 13,104) contraindicated these results noting increased hospital LOS and decreased ICU LOS for LEP patients compared to EP patients [11]. Notably, none of these studies addressed interpreter use among LEP patients. These discrepancies in the literature challenge the effect of preferred language on outcomes and raise the possibility that differences in injury presentation, cohort demographics, location and hospital catchment population, or discharge practices could have yielded these differences [10]. Further research is required to evaluate the source of these differing observations. 2.6 % of our cohort patients died. There was no significant difference in mortality between our patient groups, which was consistent with findings from a study conducted in Georgia [12]. A California study reported slightly increased mortality for LEP patients with a language preference other than Chinese or Spanish compared to EP patients [11]; however, only 8 % of our LEP patients spoke a language other than Spanish, so we were unable to compare patients by preferred language. Further research is needed to evaluate the provision of interpreters for patients preferring languages other than those most commonly observed to ensure that their needs are being equally recognized and addressed.

We evaluated a limited number of outcome variables: length of stay, ICU length of stay, and in-hospital mortality. It is possible that analysis of other common trauma outcomes (readmission, mortality at one year, discharge to acute rehabilitation facility, disability at discharge, disability at one year) may reveal disparities not evident in this study. Because the primary competing risk, death, occurred relatively rarely (2.6 %) in our population, it is unlikely that we overestimated the hazard ratio of discharge in the LEP without interpreter population via use of the Cox model over the Fine-Gray model. However, because the mortality rate was higher (10 % vs. 2.5 %) in this population, correction of theoretical overestimation would simply decrease the HR further, revealing an even more significant disparity.

This study was performed as a retrospective chart review, only limited information on both predictor and outcome variables was available. Information on preferred language and interpreter requirement was gathered through the demographic section of the electronic medical record, which is typically filled out by administrative or healthcare personnel and may not reflect true patient preference. Furthermore, there is no way to ensure that all use of professional interpreters is recorded in the patient chart. As such, it is possible that some patients classified in our study as “no interpreter access” in fact accessed utilized interpretation services without any documentation in the chart. There is also no way to determine the quality of patient interactions with interpreters. The presence of an interpreter does not ensure patient comprehension or positive interactions. This study clearly demonstrates that LEP patients with access to professional interpreters early in their trauma hospitalization do not experience the same disparities as their counterparts without interpretation access. LEP patients are at higher risk for inadequate physical exams [5], improper pain management [2,3], and failures in discharge prescribing [3]. We show that provision of an interpreter may mediate the effects of these disparities on trauma outcomes. One third of our LEP trauma patients did not access an interpreter within 24 h of arrival, a strong reminder to remain diligent and prioritize effective communication with all patients.

As studies of interpretation services in trauma have thus far examined single-center, retrospective data, further investigations are required to increase the generalizability of findings. While a randomized control trial would shine light on the impact of interpreter use in trauma activation outcomes, withholding interpreter services from a control group is outside the scope of ethical practice. Meta-analysis of existing data or multi-center retrospective studies, however, could prove useful in augmenting the literature. Additional directions for future research should include evaluation of the quality of interactions between professional interpreters and LEP trauma patients. Investigation of communication and structural barriers experienced by interpreters and front-line providers working with LEP patients in critical condition would be insightful, as would evaluation of whether communication perceived as empathetic and effective results in different outcomes.

Conclusion

Only two-thirds of patients with LEP were documented as utilizing interpreter services early in their hospital stay, evidence that future research is required to improve implementation of interpreters within the trauma workflow. The current literature is limited regarding solutions to consistent interpreter implementation in critical care environments. A single pediatric emergency department study found that providers will more often use professional interpreters when a video format is offered as opposed to telephone only [13]; however, given that our institution already provides universal access to video interpreter services, the video format does not eliminate all barriers to early interpreter use. Our findings demonstrate that LEP trauma patients with early access to an interpreter experience improved length of stay, eliminating the disparity noted between EP patients and LEP patients without early access to an interpreter.

Ethics approval

In accordance with 45 CFR 46.104(d), this study is exempt from Human Research Subject Regulations. The IRB has also approved the request for a Waiver of HIPAA Authorization, after determining that the waiver of authorization satisfies the criteria set forth in the HIPAA Privacy Rule at 45 CFR 164.512(i)(2). The waiver of authorization was reviewed and approved by the IRB. The IRB has determined that the protected health information necessary to be used and accessed is as outlined in the IRB application.

CRediT authorship contribution statement

Sydney C. Bertram: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft. F. Riley Nichols: Data curation, Investigation, Visualization, Writing – original draft. Lauren E. Cox: Investigation, Writing – original draft. Deepak K. Ozhathil: Supervision, Writing – review & editing. Mike M. Mallah: Conceptualization, Supervision, Writing – review & editing.

Declaration of competing interest

The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article. Data analysis was supported in part by the National Center for Advancing Translational Sciences of the National Institutes of Health under Grant Number UL1 TR001450. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.