Assessing the need for a medical interpreter: are all questions created equal?

Language preference is currently being used in clinical practice to determine whether an interpreter is needed. The concordance of ability to communicate and language proficiency with each other and to language preference was measured with kappa agreement scores, sensitivity and specificity among 1,000 patients surveyed in Montreal, Canada. Though concordance between language preference and language proficiency or ability to communicate was moderate, both variables had low sensitivity (69 and 55 % respectively). A total of 25 % of persons with limited language proficiency and 15 % of those with limited ability to communicate were not identified to have a language preference for their mother tongue. Also, 31 and 45 % of those who preferred to be served in their mother tongue had good language proficiency and good ability to communicate. When assessing a patients' need for an interpreter, language preference is insufficient as a stand-alone question.