Effects of medical interpreter utilization on healthcare outcomes : an examination of California

Abstract

This study examines the effects of different medical interpreters on the self-assessed health and medical visits of patients with limited English proficiency. Individual-level, pooled cross-sectional data were obtained from the California Health Interview Survey. Interpreter groups included professionals, bilingual medical staff, other interpreters, and no interpreter available. Results from multinomial logistic regression suggest patients using professional interpreters have a greater preference for healthcare. Estimates from a hurdle model report that having an interpreter increases access to healthcare, but those with no interpreter or a professional interpreter are the most frequent visitors. Ordered probit model estimates indicate that patients with professional interpreters report having the greatest health. A fuzzy regression discontinuity design is used to attempt to correct for the potential selection bias in choice of interpreters; these results suggest that patients using professional interpreters have less medical visits and worse health than patients using another interpreter or no interpreter at all. Findings underscore the need for (quasi-) experimental design in future research assessing the effectiveness of different medical interpreters in facilitating healthcare outcomes.