In the USA, over 25 million people have limited English proficiency.1 Language barriers in healthcare can lead to miscommunications between patients and clinicians that have significant repercussions on patient safety and health outcomes. Patients with limited English proficiency experience high rates of medical errors, avoidable readmissions, and low rates of outpatient follow-up, use of preventive services, and medication adherence.1, 2

Though many hospitals and clinics now provide interpreter services, there is growing evidence that patients should communicate with a language-concordant physician (i.e., a physician fluent in the patient’s native language) for more optimal care.3 Prior research has shown that language concordance is associated with more trusting patient-physician relationships, greater patient satisfaction, and an increased understanding of and adherence to treatment plans, but little is known about the impact on healthcare utilization.4, 5 In this study, we compared primary care, specialist, inpatient, and emergency department (ED) utilization between non-English-speaking patients with and without a language-concordant primary care physician (PCP).


Our study population included all non-English-speaking members enrolled in SCAN Health Plan, a Medicare Advantage health plan in California, between January 1, 2019, and December 31, 2019. A member was designated as a non-English speaker if their spoken or written language was any language other than English. The language-concordant group included members with a PCP fluent in their primary language, while the language-discordant group included members without a PCP fluent in their primary language. Descriptive analyses were used to characterize the groups, and the CMS Hierarchical Condition Categories (HCC) risk score was used as a proxy for health status. For each care setting (primary care, specialty, inpatient, ED), the average number of visits per patient per year was computed and compared between the language-concordant group and language-discordant group. Two-tailed t tests and chi-squared tests were used as appropriate. Statistical significance was defined as p < 0.05. A generalized linear regression model was used to control for age, gender, race, language, and HCC risk score. Statistical analyses were performed using SAS. This study was determined not to be human subject research and not reviewed by the SCAN Institutional Review Board.


A total of 34,600 members met eligibility criteria for non-English speakers. They were predominantly over 65 years of age (94.9%), Hispanic (42.8%), and Spanish-speaking (93.7%), and had a HCC risk score less than 1.0 (53.7%) (Table 1). We found that 64.1% (22,195) of non-English-speaking members had a language-concordant PCP and 35.9% (12,405) did not. On average, patients with a language-concordant PCP had lower total utilization than patients without a language-concordant PCP (12.22 vs. 13.07; p < 0.001). However, patients with a language-concordant PCP had 5.2% more PCP visits (3.04 vs. 2.89; p < 0.001), 9.8% fewer specialist visits (8.73 vs. 9.69; p < 0.001), 19.8% fewer inpatient visits (0.14 vs. 0.17; p < 0.001), and 4.1% fewer ED visits (0.31 vs. 0.32; p = 0.09) compared to patients without a language-concordant PCP (Table 2).

Table 1 Demographic characteristics of non-English-speaking patients enrolled in SCAN Health Plan between January 1, 2019, and December 31, 2019
Table 2 Average number of PCP, specialist, inpatient, and ED visits per patient per year for non-English-speaking patients enrolled in SCAN Health Plan between January 1, 2019, and December 31, 2019


Our results suggest that patient-physician language concordance is associated with greater primary care utilization and lower specialist, inpatient, and ED utilization. It is well established that a sizeable proportion of ED visits can be managed in physician offices and are preventable with timely and appropriate primary care.6 Our findings suggest that patients may be more likely to visit their PCP for preventive and follow-up care if no language barrier exists, resulting in fewer hospitalizations. Thus, matching patients to a language-concordant PCP may improve primary care utilization among non-English-speaking patients. Given the increased costs associated with specialist, inpatient, and ED settings, pairing patients with a language-concordant PCP when possible may lead to significant cost savings.

Limitations include that our study population was limited to members enrolled in a single health plan and that our study period only extended through December 2019. Also, while interpreters are available at most clinical sites, we have no measures of their use. This may confound our results since patients without a language-concordant PCP may still have an interpreter available to them.

There is a clear need to address language barriers in efforts to improve care for minority communities. As healthcare organizations design and implement programs for patients with limited English proficiency, facilitating language concordance between patients and physicians may lead to lower cost and more appropriate use of healthcare resources.