Limited English Proficiency and Clinical Outcomes After Hospital-Based Care in English-Speaking Countries: a Systematic Review

Abstract

Background: Limited English proficiency (LEP) is common among hospitalized patients and may impact care. We synthesized the literature comparing clinical outcomes after in-hospital care for English-proficient(EP) versus LEP patients.

Methods: This systematic review searched PubMed, Embase, and Web of Science from database inception through June 7, 2020, to identify research investigating clinical outcomes in patients receiving hospital-based care (in the emergency department, inpatient ward, surgical/procedural suite, or intensive care unit) that compared patients with LEP to an EP group. We assessed mortality, length of stay (LOS), readmissions/revisits, and complications. Study quality was evaluated using the Newcastle-Ottawa Scale.

Results: Twenty-six studies met eligibility criteria. Study settings and populations were heterogeneous. Determination of primary language varied; a majority of studies (16/26) used patient self-report directly or via hospital records. Of 16 studies examining LEP and all-cause mortality, 13 found no significant association. Of 17 studies measuring LOS, 9 found no difference, 4 found longer LOS, 3 found shorter LOS, and 1 had mixed LOS results among patients with LEP. Several investigations suggested that LOS differences may be mediated at the hospital level. Nine studies evaluated inpatient readmissions. Among patients with LEP, there was evidence for increased readmissions in the setting of chronic medical conditions such as heart failure, but no evidence for increased readmissions among cohorts undergoing surgeries/procedures or with acute medical conditions. Five studies evaluated complications or harm related to a hospitalization, and no differences were found between language groups.

Discussion: The research community lacks a standardized definition of LEP. Most studies did not find an association between English proficiency and mortality or complications. LOS findings were mixed and may be influenced at the hospital level. Differences in readmissions by language were concentrated in chronic medical conditions. Given the paucity of studies examining LEP populations, additional research is imperative.

Prospero registration number: CRD42020143477.

Keywords: communication barriers; inpatients; language; limited English proficiency.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses(PRISMA) flow diagram.

Figure 2

Inpatient mortality estimates for the limited English proficiency (or analogous) group compared to those of the English-proficient group. In the case of three or more groups, estimates are reported for the largest LEP group compared to the EP group. For Anderson and Biswas, estimated OR and 95% CI were calculated from data provided in each manuscript. For Douglas and Basic, OR and 95% CI reported in each manuscript were inverted to display the LEP vs. EP estimate. Studies that did not provide in-hospital mortality data or that reported data stratified into ≥ 5 conditions without an overall analysis were excluded. Appy = appendectomy, Chole = cholecystectomy, EGS = emergency general surgery, LOA = lysis of adhesions.

Figure 3

Readmission rate estimates for the limited English proficiency (or analogous) group compared to that of the English-proficient group. Estimates are displayed for studies recording 30-day readmissions or hazard ratios for readmission. For Biswas and López, ORs were calculated from data provided in the manuscript. For Regalbuto, confidence intervals were estimated using the reported p value and the method described by Altman and Bland. Effect estimates for Biswas, Inagaki, Karliner, and López are ORs, denoted by the square marker. Effect estimates for Regalbuto and Seman are hazard ratios, denoted by the diamond marker. Effect estimates for Rawal are relative risks, denoted by the circle marker. COPD = chronic obstructive pulmonary disease, Hip fx = hip fracture, HF = heart failure, PNA = pneumonia.