Abstract
This study discusses observed associations between limited English proficiency, revisits, and readmissions to 2 Toronto hospitals.
Patients with limited English proficiency (LEP) in predominantly Anglophone settings face barriers to safe and high-quality health care.1,2 We examined whether emergency department (ED) visits or readmissions differed between English-proficient (EP) and LEP patients discharged with acute and chronic conditions.
Methods
This retrospective cohort study included all patients discharged with 2 acute conditions (pneumonia and hip fracture) and exacerbations of 2 chronic conditions (chronic obstructive pulmonary disease [COPD] and heart failure) from 2 academic hospitals in Toronto, Ontario, Canada, between January 1, 2008, and March 31, 2016.
Data were collected from hospital and administrative databases, using International Classification of Diseases, 10th Revision codes to identify patients. Patients with a non-English preferred language listed in the electronic medical record were considered to have LEP. We excluded individuals younger than 18 years, with no recorded postal code or language, or with hip fracture prior to age 45 years.
We studied 30-day ED visits and readmissions 30 or 90 days after discharge using multivariable regression models with log-link binomial generalized linear models to determine relative risks (RRs) for each outcome, adjusting for age, sex, Charlson comorbidity index, income, fiscal year, and hospital. Statistical significance was set at a 2-tailed α =.05. Analyses were completed using SAS version 9.2 (SAS Institute Inc).
The University Health Network’s Research Ethics Board approved the study and granted an informed consent waiver.
Results
Compared with EP patients (n = 7545), individuals with LEP (n = 2336) were older, were more likely to be women, and had lower income and more comorbidities (Table 1). Of the 9881 patients included (1721 with COPD; 2608, pneumonia; 4213, heart failure; and 1339, hip fracture), 14.7% (n=1449) had a 30-day ED visit, 12.5% (n=1240) had a 30-day readmission, and 22.0% (n=2169) had a 90-day readmission (Table 2).
Table 1. Baseline Characteristics (N = 9881).
| Characteristics | No. (%) of Patients | P Value | |
|---|---|---|---|
| Limited English Proficient (n = 2336) | English Proficient (n = 7545) | ||
| Age, mean (SD), y | 80.2 (10.0) | 70.8 (16.7) | <.001 |
| Sex | |||
| Women | 1226 (52.5) | 3641 (48.3) | <.001 |
| Men | 1110 (47.5) | 3904 (51.7) | |
| Hospital | |||
| Toronto General Hospital | 629 (26.9) | 4197 (55.6) | <.001 |
| Toronto Western Hospital | 1707 (73.1) | 3348 (44.4) | |
| Medical condition | |||
| Chronic obstructive pulmonary disease | 403 (17.3) | 1318 (17.5) | .004 |
| Pneumonia | 658 (28.2) | 1950 (25.8) | |
| Heart failure | 928 (39.7) | 3285 (43.5) | |
| Hip fracture | 347 (14.8) | 992 (13.2) | |
| Charlson scorea | <.001 | ||
| 0 | 1265 (54.2) | 4749 (62.9) | |
| 1 | 652 (27.9) | 1481 (19.6) | |
| ≥2 | 419 (17.9) | 1315 (17.5) | |
| Income quintile | |||
| 1 (Lowest) | 657 (28.1) | 1711 (22.7) | <.001 |
| 2 | 691 (29.6) | 1549 (20.5) | |
| 3 | 531 (22.7) | 1398 (18.5) | |
| 4 | 274 (11.7) | 1152 (15.3) | |
| 5 (Highest) | 183 (7.9) | 1735 (23.0) | |
| Languageb | |||
| English | 7545 (100) | <.001 | |
| Portuguese | 833 (35.7) | ||
| Italian | 546 (23.4) | ||
| Cantonese, Mandarin, or Chinese | 336 (14.4) | ||
| Greek | 129 (5.5) | ||
| Spanish | 60 (2.6) | ||
| Other | 432 (18.5) | ||
The Charlson Comorbidity Index assigns a score measuring comorbidity; a higher score indicates a greater likelihood of death from comorbid disease.
Only the top 6 preferred languages are presented; language variable excludes American sign language (n = 6), deaf (n = 1), Chinese and English (n = 1), Greek and English (n = 1), not provided (n=2), to be confirmed (n = 32), unable to answer (n = 10), and unknown (n = 55).
Table 2. Adjusted Relative Risk of Study Outcomes by Condition and Preferred Language.
| Conditions | No. (%) of Patients | Unadjusted Difference (95% CI) | Adjusted Relative Risk (95% CI)a | |
|---|---|---|---|---|
| Limited English Proficient (n=2336) | English Proficient (n=7545) | |||
| 30-Day ED visit | ||||
| COPD (n = 1721) | 71 (17.6) | 191 (14.5) | 3.1 (−1.1 to 7.3) | 1.25 (0.95-1.66) |
| Pneumonia (n = 2608) | 99 (15.1) | 282 (14.5) | 0.6 (−2.6 to 3.8) | 1.11 (0.89-1.40) |
| Heart failure (n = 4213) | 200 (21.6) | 482 (14.7) | 6.9 (4.0 to 9.8) | 1.32 (1.12-1.55)b |
| Hip fracture (n = 1339) | 33 (9.5) | 91 (9.2) | 0.3 (−3.3 to 3.9) | 1.12 (0.76-1.66) |
| 30-Day readmission | ||||
| COPD | 63 (15.6) | 155 (11.8) | 3.8 (−0.2 to 7.8) | 1.51 (1.11-2.06)b |
| Pneumonia | 72 (10.9) | 231 (11.9) | −1.0 (−3.8 to 1.8) | 1.00 (0.77-1.31) |
| Heart failure | 168 (18.1) | 457 (13.9) | 4.2 (1.5 to 7.0) | 1.29 (1.08-1.54)b |
| Hip fracture | 22 (6.3) | 72 (7.3) | −1.0 (−4.0 to 2.0) | 1.05 (0.64-1.74) |
| 90-Day readmission | ||||
| COPD | 105 (26.1) | 272 (20.6) | 5.5 (0.7 to 10.3) | 1.32 (1.06-1.65)b |
| Pneumonia | 127 (19.3) | 380 (19.5) | −0.2 (−3.7 to 3.3) | 1.02 (0.84-1.23) |
| Heart failure | 280 (30.2) | 844 (25.7) | 4.5 (1.2 to 7.8) | 1.24 (1.09-1.40)b |
| Hip fracture | 45 (13.0) | 116 (11.7) | 1.3 (−2.8 to 5.4) | 1.23 (0.88-1.72) |
Abbreviations: ED, emergency department; COPD, chronic obstructive pulmonary disease.
Adjusted for age, sex, Charlson Comorbidity Index, imputed income quintile from 6-digit postal code, fiscal year, and hospital.
P < .05.
Patients with LEP and heart failure had an increased risk of a 30-day ED visit (21.6% vs 14.7%; RR, 1.32; 95% CI, 1.12-1.55) compared with EP patients. Patients with LEP and heart failure experienced greater risk of readmission at 30 days (18.1% vs 13.9%; RR, 1.29; 95% CI, 1.08-1.54) and at 90 days (30.2% vs 25.7%; RR, 1.24; 95% CI, 1.09-1.40). Patients with LEP and COPD also had greater risk of readmission at 30 days (15.6% vs 11.8%; RR, 1.51; 95% CI, 1.11-2.06) and at 90 days (26.1% vs 20.6%; RR, 1.32; 95% CI, 1.06-1.65) but did not have significantly increased risk of a 30-day ED visit (17.6% vs 14.5%; RR, 1.25; 95% CI, 0.95-1.66) than did EP patients. For patients discharged with pneumonia or hip fracture, there was no significant difference in ED visits or readmissions between patients with LEP or EP (Table 2).
Discussion
In 2 Toronto hospitals, patients with LEP and heart failure were more likely to return to the ED after discharge than EP patients. Patients with LEP and heart failure or COPD were also more likely to be readmitted.
In a previous study, patients with LEP and myocardial infarction had similar lengths of stay and mortality rates as EP patients.3 Myocardial infarction was hypothesized to be a “minimally communication sensitive” condition, as its protocolized care may mitigate the effect of language. Hip fracture and pneumonia may also be minimally communication sensitive. Hip fracture care is protocolized at the hospitals studied, and pneumonia is typically treated with a defined duration of antibiotics.
In contrast, COPD and heart failure require complex chronic disease management after discharge. Patients with LEP report poorer comprehension of discharge instructions, lower medication adherence, and difficulty with care transitions.4,5,6 COPD and heart failure may represent communication-sensitive conditions, and strategies to improve discharge communication and postdischarge support may be required.
This study has several limitations. First, residual confounding is possible, as patient factors such as illness severity, health literacy, education, and race/ethnicity were not measured, and imbalances between groups favored greater revisits and readmissions in patients with LEP. Second, not all individuals with a non-English preferred language have LEP. Defining LEP is complex and definitions vary, but the definition in this study has been commonly used. Third, although both hospitals offer medical interpretation, patient-level data on interpreter use was not available. Fourth, only 2 Toronto hospitals in the same network were studied.
Section Editor: Jody W. Zylke, MD, Deputy Editor.
References
- 1.Karliner LS, Kim SE, Meltzer DO, Auerbach AD. Influence of language barriers on outcomes of hospital care for general medicine inpatients. J Hosp Med. 2010;5(5):276-282. doi: 10.1002/jhm.658 [DOI] [PubMed] [Google Scholar]
- 2.Divi C, Koss RG, Schmaltz SP, Loeb JM. Language proficiency and adverse events in US hospitals: a pilot study. Int J Qual Health Care. 2007;19(2):60-67. doi: 10.1093/intqhc/mzl069 [DOI] [PubMed] [Google Scholar]
- 3.Grubbs V, Bibbins-Domingo K, Fernandez A, Chattopadhyay A, Bindman AB. Acute myocardial infarction length of stay and hospital mortality are not associated with language preference. J Gen Intern Med. 2008;23(2):190-194. doi: 10.1007/s11606-007-0459-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Karliner LS, Auerbach A, Nápoles A, Schillinger D, Nickleach D, Pérez-Stable EJ. Language barriers and understanding of hospital discharge instructions. Med Care. 2012;50(4):283-289. doi: 10.1097/MLR.0b013e318249c949 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Graham CL, Ivey SL, Neuhauser L. From hospital to home: assessing the transitional care needs of vulnerable seniors. Gerontologist. 2009;49(1):23-33. doi: 10.1093/geront/gnp005 [DOI] [PubMed] [Google Scholar]
- 6.Wilson E, Chen AH, Grumbach K, Wang F, Fernandez A. Effects of limited English proficiency and physician language on health care comprehension. J Gen Intern Med. 2005;20(9):800-806. doi: 10.1111/j.1525-1497.2005.0174.x [DOI] [PMC free article] [PubMed] [Google Scholar]