a b s t r a c t

Objectives: To determine if a physician-patient language barrier impacts the diagnostic accuracy of pulmonary embolism (PE) evaluation.

Methods: A retrospective chart review, conducted between June 2015 and December 2016, of a consecutive sam- ple of diagnostic computed tomography pulmonary angiogram studies performed on adult patients. Pos- itive and negative CTPA scans were further categorized by patient language and the positive diagnostic yield was determined for each language group. A post collection sub-analysis was performed to determine the yield when interpreter services were identified as necessary.

Results: The yield for English speaking patients was 10.24% (92/898, 95% CI 8.39% to 12.36%), similar to the yield

in Spanish speaking patients of 9.40% (25/266, 95% CI 6.31% to 13.37%, P = 0.69). This contrasted with the yield in patients who identified as bilingual, which was significantly lower at 1.41% (1/71, 95% CI 0.07% to 6.75%) com- pared to both English-(P b 0.02) and Spanish-only speakers (P b 0.03). The yield for non-English speaking pa- tients who requested an interpreter was 7.37% (14/190, 95% CI 4.26% to 11.77%) versus 3.23% (2/62, 95% CI 0.54% to 10.25%, P = 0.25) in those who did not.

Conclusions: The diagnostic yield in English- and Spanish-only speaking patients was similar, however, the yield in those that self-identified as bilingual was significantly lower. In patient groups in which a language barrier existed and an interpreter was not utilized, there was a trend toward a lower diagnostic yield. This suggests an increased propensity to order diagnostic imaging when potential communication barriers exist.

(C) 2017

Introduction

Pulmonary embolism (PE) is an important cause of morbidity and mortality, with N 300,000 cases diagnosed annually in the United States alone [1]. However, the diagnostic evaluation of PE, specifically the utilization of advanced imaging, has significant disadvantages, nota- bly ionizing radiation and intravenous contrast exposure [1]. Despite this, approximately 1.5% of all ED visits include computed tomography (CT) evaluation for PE [2]. With the risks associated with advanced im- aging and increasing health care costs, efforts have developed to im- prove imaging efficiency [2]. In 2014, the American College of Emergency Physicians (ACEP) recommended the reduction of avoidable imaging in low-risk patients as an initiative in the choosing wisely cam- paign [3].

* Corresponding author at: Department of Emergency Medicine, Maricopa Integrated Health System, 2601 E. Roosevelt St., Phoenix, AZ 85008, USA.

E-mail address: [email protected] (J.R. Stowell).

PE risk stratification relies on effective physician-patient communi- cation to identify individuals likely to benefit from imaging studies. Ac- cording to the 2012 US Census, approximately 57 million (20%) people five years and older living in the United States retain limited English proficiency (LEP). Of these, 24.5 million (43%) speak English ‘less than very well’, an increase from 6.6 million in 1990 [4]. Due to LEP, millions of US residents are limited in their ability to communicate with their physicians. The potential impact of language barrier with reference to risk stratification of LEP patients in the diagnostic evaluation of PE has yet to be studied, an area of critical need. This study seeks to describe the efficiency of advanced imaging in the evaluation of PE when a physician-patient language barrier exists.

Methods

The study protocol was approved by the Institutional Review Board and performed as a retrospective chart review conducted in a safety-net hospital system with twenty-four hour onsite, in-person English- Spanish interpreter services. The electronic medical record (EMR)

https://doi.org/10.1016/j.ajem.2017.12.056

0735-6757/(C) 2017

678 J.R. Stowell et al. / American Journal of Emergency Medicine 36 (2018) 677–679

utilized for this study was Epic Systems (Verona, Wisconsin). Patient Language preference and request for interpreter services were obtained during the patient intake process by the emergency department’s regis- tration staff and recorded in the EMR. This information was available to the clinical team via the EMR patient track board prior to patient evalu- ation. Utilization of interpreter services during the clinical evaluation was at the discretion of the provider team.

Consecutive diagnostic computed tomography pulmonary angio- gram (CTPA) imaging studies of the chest performed on adult patients via standard PE protocol between June 2015 and December 2016 were reviewed. Positive results included any diagnosis of PE, acute or chronic, by the attending radiologist. Both positive and negative CTPA scans were further categorized by patient language preference (English, Span- ish, or Spanish/English). Data extraction was performed jointly amongst the study authors per the study protocol. Two individual study authors extracted all data separately, with discrepancies adjudicated by a third. The study authors were not blinded to the study’s objectives during the data extraction process.

The positive diagnostic yield was determined for each group, calcu- lated as the percentage of CTPAs that were positive for PE. The diagnos- tic yield of CTPA for PE calculation presupposes patient risk stratification was performed by the clinical team in real-time and met the threshold for imaging acquisition. A patient with a higher risk of PE is presumably both more likely to receive a CTPA and more likely to have a PE, there- fore, CTPA yield would not be affected by pretest risk factors. A sub- analysis was performed to determine the positive diagnostic yield when interpreter services were identified as necessary, where this in- formation was available. Results were evaluated using chi-squared anal- ysis with mid-p level confidence intervals. A priori power analysis was performed to detect a yield difference of 6% between language groups.

Results

Overall, there were 121/1256 positive CTPA scans, resulting in a yield of 9.63% (95% CI 8.09% to 11.36%). 21 CTPA scans (1.67%) with missing or “Other” language results were removed prior to analysis. The yield for English speaking patients was 10.24% (92/898, 95% CI 8.39% to 12.36%), similar to the yield in Spanish speaking patients of 9.40% (25/266, 95% CI 6.31% to 13.37%, P = 0.69). This contrasted with the yield in patients who identified as bilingual (Spanish/English), which was significantly lower at 1.41% (1/71, 95% CI 0.07% to 6.75%)

compared to both English (P b 0.02) and Spanish-only speakers (P b

0.03) (Fig. 1).

We then examined need for an interpreter to understand the discrepancy in yield between the monolingual versus bilingual groups. In the study population, data for whether an interpreter was requested was available in 901/1256 (71.8%) of the patient en- counters. The yield for non-English speaking patients who request- ed an interpreter was 7.37% (14/190, 95% CI 4.26% to 11.77%) versus 3.23% (2/62, 95% CI 0.54% to 10.25%, P = 0.25) in those who did not. These results do not reach statistical significance; however, this sub-analysis was not sufficiently powered to detect a difference.

Discussion

The diagnostic yield of CTPA for PE in study patients who self- identified as Spanish-speaking only was statistically similar to English- speaking patients and suggests language barrier did not significantly prevent physicians from appropriately risk stratifying patients for PE. Study findings may be due to the benefits of language translation services.

Communication, incorporating verbal conversation, nonverbal cues, and cultural context, forms the foundation of the physician-patient rela- tionship [5]. For LEP patients, inadequate communication is a formida- ble obstacle to obtaining health care and a significant predictor of resource utilization disparity [6]. LEP patients undergo increased test- ing, including advanced imaging, in specific clinical presentations, such as undifferentiated abdominal pain, as compared to English- speaking counterparts [7]. Despite this, LEP is associated with worsened outcomes in multiple disease processes [8]. These findings indicate LEP patients may receive less effective health care resource allocation as compared to English-speaking patients [6].

Utilization of professional interpreters is essential for accurate com- munication, and ultimately effective Health care delivery when a lan- guage barrier exists [9]. While professional interpreters are often available, physicians may alternatively choose to utilize their own lan- guage skills, or rely on surrogate interpreters such as patient family members or friends, which can result in communication gaps [10]. Amongst the study population, the diagnostic yield in those who self- identified as bilingual was significantly lower when compared to the English-speaking only group. Limited interpreter utilization in this group, due to a perceived shared language, may account for this finding.

Fig. 1. Proportion yields for the diagnosis of pulmonary embolism on Computed tomography imaging for each language group.

J.R. Stowell et al. / American Journal of Emergency Medicine 36 (2018) 677–679 679

Indeed, in patient groups in which a language barrier existed and an in- terpreter was not utilized, there was a trend toward a lower diagnostic yield. This suggests potential limitations in algorithmic PE risk stratifica- tion, with an increased propensity to obtain diagnostic imaging studies, when a communication barrier is present and professional interpreters are not employed.

During study protocol development and analysis, several potential limitations were identified. Patient language preference and request for interpreter services were identified and recorded by registration ser- vices, rather than self-selected by the patient, which may have misidentified language barriers. While the study population is briefly described, general patient demographics were not obtained, which may limit generalizability. Indications for image acquisition were not obtained. Pretest probability was not ascertained, but we have no rea- son to believe native speakers of one language have a different risk pro- file for PE than those who speak another language. Furthermore, any set of patients more likely to have PE would also be more likely to receive imaging, and thus yield should remain similar. Although CTPA studies are primarily utilized for PE evaluation, imaging obtained for alternative etiologies may have impacted the diagnostic yield calculation. Diagno- ses of chronic PE by CTPA were included in the study analysis as the study authors were unable to assess whether the clinical team was aware the patient’s symptoms were due to chronic PE, rather than acute, at the time of image acquisition. Finally, interpreter utilization data were not obtained. Sub-analysis of non-English speaking patients relied on whether interpreter services were requested at the time of intake, which make have impacted the diagnostic yield calculation. Despite these limitations, the study authors feel that the study protocol and data analysis meet the Study objectives.

The diagnostic yield in English- and Spanish-only speaking pa- tients was similar, however, the yield in those that self-identified as bilingual was significantly lower. Limited interpreter utilization in this group may account for this finding. Further studies are warranted to understand the effect of communication barriers on diagnostic accuracy in PE.

Acknowledgment

The study authors would like to acknowledge Heather Jordan, RN, BSN² for assistance with data collection.

Author contribution

JRS and MA conceived and designed the study. JRS, MA, and ATR su- pervised and assisted in data collection. JRS, MA, ATR, LF and MSS con- tributed to the data review and analysis. JRS drafted the manuscript, and all authors contributed substantially to its revision. JRS takes re- sponsibility for the paper as a whole.

Financial disclosure and conflicts of interest

JRS reports no conflict of interest. LF reports no conflict of interest. MSS reports no conflict of interest. ATR reports no conflict of interest.

MA has a career development grant from the Emergency Medicine Foundation for his basic science research in traumatic brain injury.

References

1. Perelas A, Dimou A, Saenz A, et al. CT pulmonary angiography utilization in the emergency department. Am J Med Qual 2015;30(6):571-7.
2. Venkatesh AK, Kline JA, Courtney DM, et al. Evaluation of pulmonary embolism in the emergency department and consistency with a national quality measure. Arch Intern Med 2012;172(13):1028-32.
3. American Board of Internal Medicine Foundation. Choosing wisely. Ten Things Phy- sicians and Patients Should Question. Washington, DC: American College of Emer- gency Physicians; 2012 [press release]. [Accessed June 25, 2017 at] http://www. choosingwisely.org/clinician-lists/acep-ct-pulmonary-angiography-in-ed-patients/.
4. US Bureau of the Census. Statistical Abstract of the US 2012 Census. 131st ed. Wash- ington, DC: US Bureau of the Census; 2012(Accessed June 25, 2017 at https://www. census.gov/library/publications/2011/compendia/statab/131ed.html).
5. Jacobs E, Chen AH, Karliner LS, Agger-Gupta N, Mutha S. The need for more research on language barriers in health care: a proposed research agenda. Milbank Q 2006; 84(1):111-33.
6. Njeru JW, St Sauver JL, Jacobson DJ, et al. Emergency department and inpatient health care utilization among patients who require interpreter services. BMC Health Serv Res 2015;15(1):214.
7. Waxman MA, Levitt MA. Are diagnostic testing and admission rates higher in non- English-speaking versus English-speaking patients in the emergency department? Ann Emerg Med 2000;36(5):456-61.
8. Murray MD, Tu W, Wu J, Morrow D, Smith F, Brater DC. Factors associated with ex- acerbation of heart failure include treatment adherence and health literacy skills. Clin Pharmacol Ther 2009;85(6):651-8.
9. Karliner LS, Perez-Stable EJ, Gregorich SE. Convenient access to professional inter- preters in the hospital decreases readmission rates and estimated hospital expendi- tures for patients with limited English proficiency. Med Care 2017;55(3):199-206.
10. Lindholm M, Hargraves JL, Ferguson WJ, Reed G. Professional language interpreta- tion and inpatient length of stay and readmission rates. J Gen Intern Med 2012; 27(10):1294-9.