A research coding method to evaluate a smoking cessation model for training residents-A preliminary report.

Abstract

Develop a reliable coding method of a cigarette cessation model used to train residents - a preliminary report. Two trained (30h) undergraduates coded videotaped interviews from 161 resident-simulated patient (SP) interactions. To establish reliability, coders coded 33 (20%) of 161 study set tapes for the BHTM. Cohen's Kappa and percent of agreement were used to measure coders' reliability in unitizing and coding residents' skills for eliciting 5 variables: Educating, Informing, and Motivating (3 items); Commitment and Goals (3 items); Negotiate Plan (7 items); Patient-Centered Skills (9 items); Emotional Skills (6 items). 50 items were dichotomized a priori from analysis of the training model and were reduced to 28 during training. Kappa ranged from 0.73 to 0.87 for the 5 variables and 28 individual items. The overall kappa was 0.84, and percent of agreement was 93%. Percent of agreement by item ranged from 82 to 100%. A highly reliable coding method, weighted (by no. of items) to highlight the key elements of the teaching, is recommended for investigators wishing to better focus on the partnership, emotions, and planning. This is a unique way to integrate patient-centered skills into motivational interviewing.