Competence of psychiatric interpreters poses an important problem. In the United States, refugees, immigrants, foreign students, guest workers, illegal migrants, and persons from other ethnolinguistic groups who are not fluent in English account for over 15% of the population. Some of these groups have higher rates of psychiatric disorders than the general population. In the pilot study reported on page XXXX, Kilian and colleagues (1) found that the skills of untrained interpreters in a South African psychiatric hospital ranged from poor to acceptable (1). These authors could well have been describing scenarios in many U.S. hospitals and clinics.

The authors have made a valuable contribution to cross-language psychiatric care by examining the skills of six hospital staff members serving as ad hoc interpreters (all had other jobs in the hospital). Researchers asked the interpreters to translate typical queries used in psychiatric evaluation and to define terms, such as depression and psychosis. Independent translators unfamiliar with the mental health field then provided back-translations, which revealed a range of competence, from interpreters who barely spoke one of the two languages to those who demonstrated command of both languages and an understanding of psychiatric concepts.

It is hoped that future studies of interpreters will identify the type and severity of translation problems, so that they can be addressed on a national and international level. Such studies could provide a basis for standards, training to meet these standards, and certification in specific languages. An initial study to assess interpreter skills would ideally have the following features. Interpreters should come from a variety of institutions from several regions, to ensure a random, or at least a representative, sample. Butcher and Garcia (2) have described a minimal level of experience in each language and culture (at least five to seven years, with day-to-day use of the language, rather than classroom exposure alone) as a prerequisite for assessing linguistic and psychometric equivalence of terms and phrases in two languages. Back-translation serves as a standard for assessing written tests—and tests exist for some interpreters. However, to my knowledge a standard assessment for psychiatric interpreters does not exist (although a few countries require training and certification). An assessment method would need to be developed. For example, a "gold standard" measure would need to assess two tasks: whether the meaning of the clinician's query or counsel is accurately conveyed to the patient (a clinician's judgment) and whether the meaning of the patient's response or query and the level of the patient's understanding are accurately conveyed to the clinician (a patient's judgment).

In their report Kilian and colleagues suggested that demographic variables such as age, gender, education, and occupation may affect the inherent skills of untrained interpreters. To examine these and other variables, such as the number of years spent living in both languages and cultures, whether the individual works full-time or part-time as an interpreter, and the type and level of training), the sample would need to be sufficiently large—more than 100 and perhaps as many as 200 interpreters.

Larger contextual issues would also need to be addressed. One involves the "model" or function of the interpreter (3). In emergency care, a "black box" model, in which the interpreter has no relationship with the clinician, may suffice. Some clinicians who work with highly skilled clinician-interpreters may adapt a "supervisor" model in which they train the interpreter in psychiatric concepts, diagnoses, and therapy—much like training a physician's assistant. However, other clinicians may simply want their queries and counsel and the patient's responses to be accurately translated. They do not want interpreters who are trained in diagnosis lest they begin to make decisions beyond their ken. These interpreters must learn to accurately translate psychiatric symptoms, ranging from headache to hallucinations to dissociation.

As Kilian and colleagues pointed out, common psychiatric experiences, such as hallucinations, require more than literal translations. Although many somatic symptoms and even some psychological symptoms (such as nightmares) can be readily translated between many languages, most psychological symptoms (such as anxiety, anhedonia, and hopelessness) require a paragraph or two of explanation for patients to understand the symptom and then decide whether they manifest it (4).

The authors have provided a valuable insight by documenting at their institution the extent and severity of this widespread problem. The next step rests with national and international institutions to create cost-effective standards and solutions to correct this deplorable inequity.