The multisite field trial of the consultation-liaison psychiatry assessment instrument.

A multisite field trial was conducted at 11 institutions to test the clinical reliability of a 29-item consultation-liaison (C-L) psychiatry assessment instrument. Twenty-five raters viewed videotapes of two "trainees" conducting clinical interviews with a simulated patient. One trainee was a medical student, the other was a fellow in psychiatry. Raters completed the 29-item assessment instrument for each trainee. The mean value scores reflected the skill of each trainee. The medical student had a mean score of 1.93, whereas the C-L fellow had a mean score of 3.13 which parallels the expected level of skill for the two interviewers. Eighty-six percent of the items (25/29) had a standard deviation (SD) of less than 1.0. Each of the remaining four items (14%) had a SD minimally greater than 1.0. These results reflect clear wording of items with measurable parameters defined for assessing trainees' skills. The authors present different uses for the assessment instrument, including giving feedback to trainees regarding interviewing techniques and skills; setting "gold" and "lead" standards for clinical C-L interviewing skills; and training supervisors in evaluation using a standardized assessment instrument.