INTRODUCTION

Approximately one in five people in the USA speaks a language other than English at home1 with Spanish, Chinese, French, Tagalog, and Vietnamese being the most common ones.2 Among these 58 million people, 25 million report having limited English proficiency (LEP).1 These individuals have difficulty speaking, reading, writing, or understanding English,3 which presents significant obstacles to accessing and receiving care in an English-dominant healthcare system. Due to language barriers, LEP patients have worse healthcare quality and outcomes4,5,6,7 and decreased access to preventive services8, 9 and cancer screening.10,11,12,13 Although federal regulations require healthcare organizations to provide trained interpreters for LEP patients, many hospitals and clinicians are non-compliant with these regulations, leading to persistent inequities in care.14 Language concordance, when a physician is fluent in a patient’s preferred language, offers a way of reducing these disparities.

Health disparities due to language barriers are reduced when care is provided by language-concordant clinicians. For example, studies demonstrate that LEP patients with language-concordant physicians report receiving more education about their care, have fewer unasked questions, and have better medication adherence and fewer emergency room visits.15,16,17 When a clinician and an LEP patient communicate in a language in which only one of them is fluent, this partial language concordance can further obfuscate clinical interactions by leading patients and providers to believe they understand each other, thus contributing directly and indirectly to medical errors and poor outcomes.14, 18 While communicating in a patient’s native language may build rapport, non-fluent clinicians must know when to call an interpreter19 and have an accurate gauge of their own limitations.14, 20 Patient satisfaction with care is higher, adherence to medications is better, and patients have fewer unanswered questions about their care when language-concordant care is provided.15, 16, 21 However, there is also evidence that language concordance is associated with lower cancer screening rates for patients with LEP.17, 22 Given these divergent findings, it is important to determine the overall impact of language concordance in clinical settings.

The purpose of this systematic review was to synthesize the data on the impact of language-concordant care for LEP patients compared to other interventions to improve language access, including professional interpreters and untrained (ad hoc) interpreters. We hypothesized that language-concordant care would be beneficial to LEP patients. Our goal was to create a comprehensive picture of the relationship between language concordance and quality of care for patients with LEP.

METHODS

Literature Search and Data Sources

We searched 5 databases for this systematic review: PubMed (NLM, 1945 to October 2017), PsycINFO (OVID, 1806 to October 2017), Web of Science (Thomson Reuters, 1945 to October 2017), Cochrane Library (Wiley, 1966 to October 2017), and EMBASE (Elsevier, 1966 to October 2017). Two authors (LCD and KM, a research librarian) developed strategies to find relevant articles. Online Appendix 1 provides detailed search strategies for each database. We did not apply limits to date or article language. The Endnote (Thomson Reuters) bibliographic citation management program was used to manage the citations and remove duplicates.

Data Synthesis and Analysis

A systematic title and abstract review was conducted by two authors (KI and LD) using the Population, Intervention, Comparison, and Outcome (PICO) framework.23 Qualitative articles that systematically gathered data on language-concordant interventions were also eligible for inclusion. Articles were eliminated without further review if they did not focus specifically on the impact of language-concordant care and clinical outcomes. Because of heterogeneity in the methods and outcomes, we pooled our results using qualitative rather than quantitative approaches.

Data Abstraction

Two authors abstracted data from the remaining 33 articles independently (LD and KI or DC). We abstracted 16 items from each article: study locations, sample sizes, type of clinical outcomes studied, participants’ ages (including range, mean, and standard deviation if provided), participants’ race and/or ethnicity, languages interpreted, type of patient navigators, study designs, recruitment methods, facility types, comparison groups, and outcomes and results/major findings. One author (LD) reviewed all abstractions and registered any discrepancy between authors. These discrepancies were resolved by consensus. For studies with more than one outcome, we categorized outcomes based on the best result (e.g., if one outcome was better for patients with language-concordant physicians and another showed no difference, we categorized the study as demonstrating better outcomes for language-concordant care).

RESULTS

Study Selection

Our search yielded 8618 citations. After adding 5 articles identified from other sources and removing duplicates and exclusion of articles that were not relevant, 126 articles remained for full-text review. Of these, 92 were excluded yielding 34 articles. During full-text review, one additional article was eliminated because it was a descriptive study of a language concordance intervention for which there were no patient outcomes described.24 A total of 33 articles were included in our qualitative analysis. Figure 1 shows the PRISMA flow diagram.25

Figure 1
figure 1

PRISMA flow diagram.

Full size image

Quality Appraisal

We systematically appraised the quality of all articles using the Downs and Black checklist.26,27,28 For this review, the modified Downs and Black was used.29 Online Appendix 2 shows the Downs and Black checklist components for each article. To assess the quality of the language services provided, the authors also abstracted information to describe any language proficiency testing clinicians received, if documented in the study. The quality of the articles overall was very good, with some variation. This was due, in part, to incomplete information from some of the included articles.

Study Characteristics

Of the 33 articles, 4 (6.9%) were randomized controlled trials and the remaining 29 (87.9%) were cohort studies. The 33 studies were grouped by the outcome measure studied, including quality of care (further subdivided into primary care, diabetes care, pain management, cancer, and hospital setting), satisfaction with care/communication, medical understanding, and mental health, as shown in Table 1. The USA was the setting of 94% of the studies (31/33). Seventy-six percent (25/33) focused on interventions for Spanish-speaking patients and 30% (10/33) for speakers of Asian languages.

Table 1 Description of Articles
Full size table

Language Concordance Outcomes

Seventy-six percent (25/33) of the studies demonstrated that at least one of the outcomes assessed was better for patients receiving language-concordant care, while 15% (5/33) of studies demonstrated no difference in outcomes, and 9% (3/33) studies demonstrated worse outcomes in patients receiving language-concordant care (Table 2).

Table 2 Study Outcomes
Full size table

QUALITY OF CARE: PRIMARY CARE

Among the studies looking at quality of care in a primary care setting, 4 demonstrated that language concordance had a positive impact on one or more outcomes studied30, 32, 33, 36 (better patient experience, higher likelihood that patients would receive and agree with counseling on diet and physical activity, and better access to and utilization of primary care providers), and 3 showed no difference31, 35 (the rate of flu vaccines and mammography and reported quality of well-child care), while 2 demonstrated a negative effect of language-concordant care on an outcome17, 22 (lower colorectal screening rates in both studies).

Quality of Care: Diabetes

In the studies examining quality of care for diabetic patients, 3 showed that language concordance had a positive impact on an outcome38, 39, 41 (better glycemic control, significant reduction in LDL, blood pressure, and HbA1C, and higher rates of adherence to CVD medications) and 1 did not show a difference in outcomes37 (in blood pressure, LDL, or HbA1C).

QUALITY OF CARE: PAIN MANAGEMENT

There was 1 study examining quality of care in pain management,42 which found that Spanish fluency and level of experience with Hispanic/Latino patients with pain had an impact on implementation of established pain management practices, while treatment with opioids was more influenced by practical matters and beliefs (e.g., finances and addiction concerns) rather than provider factors.

QUALITY OF CARE: CANCER

In the studies examining quality of care for cancer patients, 1 study showed that the language-concordant intervention was associated with higher rates of radiation therapy following surgery for breast cancer but not for prostate cancer patients44 (where there was no statistically significant difference), and another study showed that language-concordant intervention resulted in higher rates of colorectal cancer screening (more participants returned the fecal occult blood test).

QUALITY OF CARE: HOSPITAL SETTING

In the 3 studies that examined quality of care in the hospital setting, 2 showed a positive impact of language concordance46, 47 (increased satisfaction with care, fewer ED visits upon discharge, better quality transition between the hospital and the outpatient setting) and 1 showed a negative outcome45 (longer ED throughput times).

Satisfaction with Care/Communication

The 8 studies looking at satisfaction with care and communication each showed at least one positive outcome15, 21, 48,49,50,51,52 (higher satisfaction with care, fewer questions about care, better health education, enhanced privacy, enhanced communication, and increased therapeutic alliance).

Medical Understanding

Of the 2 studies evaluating medical understanding, each showed a positive outcome54, 55 (better understanding of diagnosis and higher likelihood of reporting problems understanding a medical situation in medical understanding).

Mental Health

Among the 3 studies focusing on mental health, 2 showed a positive outcome24, 57 (higher rates of adequate treatment and lower rates of pathological symptoms identified), with 1 showing no difference in the perception of mental health needs and discussion of these needs with a provider.56

DISCUSSION

Overall, the results of this review support the notion that language-concordant care is associated with better outcomes for LEP patients. Of the 33 studies included, the majority demonstrated a positive impact of language-concordant care on at least one of the major outcomes studied. The positive findings identified included patient-reported measures, such as patient satisfaction and understanding of diagnosis,48, 49, 51 and objective measures, including glycemic control and blood pressure for diabetic patients.38,39,40 These results align with our initial hypothesis that language-concordant care leads to better outcomes for LEP patients but there remain contrasting findings that may indicate some unintended consequences of language concordance.

The positive effects of language concordance are widespread, though there are some areas where receiving care from a language-concordant provider may be particularly advantageous. The majority of studies related to patient satisfaction, utilization of and access to care, and self-perceived knowledge of diagnosis demonstrated that language-concordant care had a positive impact. Additionally, the majority of publications in which process measures (blood pressure, HbA1C, and LDL) were studied showed that there is an association between language-concordant care and better process measure values for LEP patients. Taken together, these findings indicate that language-concordant care is associated with satisfaction and empowerment among patients, potentially improving their relationships with their long-term providers and producing positive health outcomes.

There were 3 studies in which language concordance was associated with negative results for LEP patients.17, 22, 45 The findings of lower colorectal cancer screening rates and longer ED throughput times for patients receiving language-concordant care may be a product of study design, or could reflect unanticipated impacts of language-concordant care that need to be taken into account by researchers and policy makers. In the context of the studies on language-concordant care in the hospital setting, which indicates higher quality of care, better patient satisfaction, and fewer subsequent ED visits for patients seen in the ED by language-concordant providers, the longer ED throughput times could be a result of providers taking more time with patients and providing better quality of care, and thus, the longer ED throughput times should not necessarily be seen as a negative outcome. The lower rates of colorectal screening that 2 studies demonstrated17, 22 may be the result of better elucidation of the risks of colorectal cancer screening by language-concordant providers, which could lead some patients to opt out of the procedure based on an improved understanding. It may also reflect a patient–provider dynamic where LEP patients have more autonomy in making healthcare decisions. It is also possible that the providers considered to be language concordant in these studies were not truly fluent in their patients’ preferred languages. There was no assessment of the clinicians’ non-English language abilities in any of the studies demonstrating lower rates of colorectal cancer screening. These findings may reflect partial language concordance and poor quality of communication around colorectal cancer screening. The studies demonstrating longer ED throughput times and lower rates of colorectal cancer screening warrant future research to understand the underlying reasons for the findings before it is concluded that they are indicative of inferior patient care.

Although the findings of this review are consistent with previous research in demonstrating the overall positive impact of language concordance, there are limitations that must be addressed. The quality of these studies, as appraised by the Downs and Black Checklist, was generally high, which lends validity to the findings of this review. However, very few of the included studies assessed the fluency of the language-concordant providers, a key variable impacting the relevance of each study’s findings. Without this information, we are left without knowing whether the quality of communication is comparable to that of an English-speaking patient–provider dyad (which often served as the comparison group). This relates to a larger issue in the landscape of patient–provider language concordance, which is that there is currently no standardized assessment of clinician language proficiency that is consistently used by healthcare facilities. If such an assessment were used and its results were reported in future studies, this data would further help in appraising the quality of the research. Another limitation of this review is the fact that most of the studies included were performed in or near major cities in the USA and in clinics or hospitals affiliated with academic medical centers, limiting the generalizability of the findings to other countries and healthcare settings.

CONCLUSIONS

Given the rapidly growing LEP population in the USA, efforts to improve quality of and access to care will be ineffective if they do not attempt to address the barriers faced by LEP patients. The findings of this review indicate that, in the vast majority of situations, language-concordant care improves outcomes. Almost all of the studies included were of good quality, but none provided a standardized assessment of provider language skills. In order to systematically evaluate the impact of truly language-concordant care on outcomes and draw meaningful conclusions, future studies must include an assessment of clinician language proficiency and longitudinal observational studies. Language-concordant care offers an important way for hospitals to meet the unique needs of their LEP patients.