A Multilingual Chatbot Can Effectively Engage Arthroplasty Patients Who Have Limited English Proficiency

Abstract

BackgroundIn the United States, English language proficiency is widely accepted as a key social determinant of health. For patients with limited English proficiency (LEP), language barriers can make the delivery of perioperative instructions challenging. The purpose of this study was to evaluate whether a multilingual chatbot could effectively engage LEP patients and improve their outcome after total joint arthroplasty (TJA). MethodsWe identified 1,282 TJA patients (705 knees, 577 hips) who enrolled in a short message service (SMS) chatbot from 2020-2022. Forty-seven patients enrolled in the chatbot received their messages in a language other than English. A historical control of 68 LEP patients not enrolled in the chatbot was identified. Chi-squared, Fisher’s exact test, and t-tests were performed to measure the effect that conversational engagement had on emergency department (ED) visits, hospital readmissions, and reoperations. ResultsThere was no difference in the conversational engagement between LEP patients and those with English as their primary language (EPL) (12.3 versus 12.2 text responses, P = .959). The LEP cohort who enrolled in the chatbot had fewer readmissions (0% versus 8.3%, P = .013) and a near significant reduction in ED visits (0.9% versus 8.0%, P = .085) compared to those not enrolled. There was no difference in reoperations between the 2 cohorts. ConclusionLEP and EPL patients engaged equally with the multilingual chatbot. LEP patients who enrolled in the chatbot had fewer readmissions and a near significant reduction in ED visits. Multilingual platforms such as this chatbot may provide more equitable care to our frequently encountered LEP patients.