Interpretation Accuracy Between Orthopaedic Surgeons and Qualified Remote Spanish Interpreters in a Trauma Setting

To assess the interpretation accuracy of remote qualified Spanish medical interpreters (QMI) in an orthopaedic trauma setting and determine whether interpretation method (telephonic or video), delivery speed, and subject complexity were associated with interpretation accuracy.

Three clinical vignettes-simple, moderate, and complex-describing fictitious orthopaedic trauma injuries, treatments, complications, and outcomes were written by the investigators. An investigator read each vignette type to a simulated patient through 20 randomly selected QMIs (10 through telephone and 10 through video), all of whom were trained, proficiency-tested, and blinded to the study design, yielding 60 total interpretations. Data collected included interpretation method, English delivery speed (words per minute), and vignette complexity. One orthopaedic traumatologist and 1 bilingual physician assistant evaluated transcriptions of recorded QMI Spanish interpretations using a nonvalidated, criterion-referenced, expert-derived categorical assessment reflecting the degree to which an interpreter preserved the fidelity of the original vignette for patient ascertainment of diagnosis, treatment, risks, and prognosis, termed overall interpretation designation (OID). OID was classified, "acceptable," "borderline," or "unacceptable." A second orthopaedist, blinded to the first 2 investigators' preliminary translation designations, independently evaluated disputed translations and adjudicated them to consensus OID.

Of the 60 interpretations, 46.7% were designated acceptable, 18.3% as borderline, and 35.0% unacceptable. Interpretation accuracy was not associated with vignette complexity (moderate odds ratio [OR]: reference; simple OR: 0.81, 95% CI, 0.21-2.95, P = 0.75; complex OR: 1.49, 95% CI, 0.36-6.74, P = 0.59) or delivery speed (Avg: 148 words per minute, range: 129.1-169.0, OR 0.98, 95% CI, 0.92-1.04, P = 0.53). Odds of an acceptable interpretation were higher with video compared with telephonic interpretation (OR 6.21, 95% CI, 2.17-19.48, P < 0.01).

Less than half of Spanish interpretations delivered to simulated orthopaedic trauma patients by qualified medical interpreters were designated as acceptable by provider investigators, and 35% were unacceptable. The odds of an acceptable interpretation were more than 6 times greater with remote video interpretation compared with telephonic interpretation.