Association of Patient-to-Primary Care Physician Language Concordance with Outcomes.

Background

Patient-clinician language concordance impacts healthcare experiences and outcomes. The association of patient-primary care physician (PCP) language concordance with resource use is understudied.

Methods

We conducted a retrospective study of adults with a PCP visit from January 2020 through March 2023 at UHealth. Our primary exposure was patient-PCP language concordance. Our primary outcome was emergency department (ED) visit within 1 year after PCP encounter. We used multivariable regression modeling, with sensitivity and exploratory analyses to investigate the interplay of language concordance, preferred language (English/Spanish), and ethnicity.

Results

We included 22,093 patients across 48 PCPs (29 [60.4%] with Spanish language ability). 21,034 (95.2%) were in language concordant patient-PCP dyads and 3,155 (15.0%) preferred Spanish; language concordant patients were less commonly Hispanic (45.8% vs 94.4%, SMD 1.29). Language concordance was associated with lower odds of 1-year ED visit (unadjusted rate: 6.7% vs 9.3%, SMD 0.09, P = 0.002; adjusted OR [95% CI]: 0.77 [0.61, 0.97]). Sensitivity analyses using a more nuanced exposure (English-English concordance = reference; Spanish-Spanish concordance, OR [95% CI]: 1.43 [1.23, 1.66]; discordance, OR [95% CI]: 1.43 [1.13, 1.80]) and, separately, restricting the cohort to Spanish-speaking patients (language concordance, OR [95% CI]: 0.99 [0.76, 1.29]) as well as an exploratory analysis adding preferred language to the primary model (language concordance, OR [95% CI]: 1.00 [0.78, 1.29]) suggested the association of language concordance was fully confounded by preferred language. Ethnicity had less of an effect.

Conclusions

Patient-PCP language concordance is associated with a reduced odds of 1-year ED visit. But as this association was not observed within the subcohort of only Spanish-speaking patients, it appears to be fully confounded by patient preferred language.