Accuracy of Self-Assessed Spanish Fluency in Medical Students

Abstract

Background: Non-English language fluency is increasingly important in patient care. Fluency self-assessment is easily obtained, but its accuracy is unknown. Purposes: The purpose is to determine accuracy of medical students' self-assessed Spanish fluency. Methods: Four matriculating classes assessed their own oral fluency as (“none”:“novice”;“intermediate”;“advanced”;“native-speaker”). Participants who rated themselves greater than “novice” and who expressed interest in medical Spanish coursework took a standardized fluency test (Spoken Language Evaluation, scaled 1–12). Using predetermined test categories (1–5 = novice, 6–8 = intermediate, 9–12 = advanced/native), we determined the predictive value of self-assessment for predicting the same or greater fluency on the test. Results: Of 102 participants, 12 (12%) tested below their self-assessed level, 77 (75%) tested at their self-assessed level, and 13 (13%) tested above. The predictive value of self-assessment for having at least that fluency level was 88% (95% CI = 80, 94). Conclusions: In medical students reporting greater than “novice” capability and interest in medical Spanish coursework, fluency self-assessment was a good indicator of scores on a standardized fluency test.

This study was approved by the University of North Carolina Biomedical Institutional Review Board. Funding was provided by The Duke Endowment, Health Care division. Dr. Reuland was also supported by National Research Service Award Fellowship, T32 HP14001. These data were initially presented at the annual meeting of the Society of General Internal Medicine, Pittsburgh, PA, April 12, 2008. Please note that ALTA Language Services, Inc. administered the language tests under a contract with our educational program. They provided technical advice about the administration and interpretation of their language test to authors and to anonymous peer reviewers. They played no role in the funding or design of study, interpretation of results, or decision to publish. We thank Gina Horne and Patricia Neff for their help with data collection for this project. We thank Rosalie Dominik, DrPH for biostatistical advice.

Notes

^a Column percentages do not sum to 100% due to rounding.

^b Fluency test descriptions provided by the language testing service, ALTA Language Services, Inc. (Atlanta, GA).