Accuracy of preferred language data in a multi-hospital electronic health record in Toronto, Canada

Abstract

Accurate preferred language data is a prerequisite for providing high-quality care. We investigated the accuracy of preferred language data in the electronic health record (EHR) of a large community hospital network in Toronto, Canada. We conducted a point-prevalence audit of patients admitted to intensive care, internal medicine, and nephrology services at three hospitals. We asked each patient "What is your preferred language for health care communication?" and reported on agreement (with 95% confidence intervals [CI]) between interview-based and EHR-based preferred language. We used Bayesian multilevel logistic regression to analyze the association between patient factors and the accuracy of the EHR for patients who preferred a non-English language. Between June 17, 2024, and July 19, 2024, we interviewed 323 patients, of whom 124 (38%) preferred a non-English language. Median age was 77 years and 46% were female. EHR accuracy was 86% for all patients. The probability of the EHR correctly identifying a patient with non-English preferred language (sensitivity) was 69% (CI 60-77), specificity was 97% (CI 94-99), positive predictive value was 95% (CI 88-98), and negative predictive value was 83% (CI 79-87). There were 26 different non-English preferred languages, most commonly Cantonese (27%) and Tamil (14%). Accuracy was better for patients who were female or older, and varied by hospital and medical service. Mechanisms to improve accuracy for language preference data are needed to improve the validity of research studying preferred language, mitigate algorithmic bias, and overcome language-based inequities.

Fig 1. This figure shows a forest plot of odds ratios for correctly identifying EHR-based non-English language preference.

This model was fit using all patients with an interview-based non-English language preference, and also included random intercepts by non-English preferred language (Fig 2). The point estimate for each odds ratio is shown as a black dot, with a horizontal black line denoting the width of the 95% credible interval. Odds ratios above 1 denote factors associated with better EHR language preference accuracy, while odds ratios below 1 denote factors associated with worse EHR language preference accuracy. * We modeled the log (base 2) of length of stay, which means that the odds ratio corresponds to each doubling of the length of stay. **Age in decades was modeled as a linear predictor.

Fig 2. Association between non-English preferred language and the probability of the EHR correctly identifying non-English preference.

Caption: This figure shows a forest plot of odds ratios for EHR-based non-English language preference, showing the value of the random intercepts. This model was fit on all patients with an interview-based non-English language preference, and also included fixed predictors (Fig 1). The point estimate for each odds ratio is shown as a black dot, with a horizontal black line denoting the width of the 95% credible interval. Odds ratios above 1 denote factors associated with better EHR language preference accuracy, while odds ratios below 1 denote factors associated with worse EHR language preference accuracy. Note that no languages are convincingly associated with higher or lower probability of correctly identifying non-English preference.