[Effects of skill level on individual behavioral counseling for smoking cessation]

The relationship between smokers' characteristics and success of smoking cessation has been examined in smoking intervention studies, but counselors' characteristics have generally not been considered. The purpose of the present study was to examine the relationship between counselor's skill level and intervention outcome.

The subjects were 858 smokers identified at annual health checkup at 6 worksites from January 1998 to September 1999. Smokers received stage-matched individual counseling by 23 trained counselors with feedback of results for carbon monoxide testing and nicotine metabolites in urine. If smokers set a quit date at the initial counseling session they received 4 follow-up telephone calls over 3 months. We further conducted 6 month and 1 year follow-up surveys using self-administered questionnaires. Smoking cessation status was validated by carbon monoxide test or family or colleague witnesses. We evaluated counselors' intervention skills using a structured evaluation form (total score: 0-24) based on videotaped interactions between each counselor and a simulated smoker. The intervention skills were classified into 3 levels, Low (0-14), Middle (15-17) and High (18-24). We performed multiple logistic regression analysis to adjust for smokers' characteristics and multi level analysis to adjust for counselors' characteristics other than counseling skills.

Among the 23 counselors, 4, 11 and 8 counselors were assigned to the Low, Middle and High groups, respectively, which counseled 190, 344 and 324 smokers. The 7 day point prevalence abstinence rates at 6 months were 2.1%, 4.7% and 7.4%, respectively (cumulative chi-square test: P < 0.01). The 6 month sustained abstinence rates at one year were 1.1%, 3.2% and 4.6%, respectively (P < 0.05). The odds ratios of abstinence rates at 6 months and one year in Middle group were 2.33 (95% CI: 0.75-7.28) and 3.07 (0.65-14.54) respectively, compared to the Low group by multiple logistic regression analysis. In the High group they were 3.66 (1.21-11.04) and 4.86 (1.06-22.28). The results of the multi level analysis were approximately equal to those in the multiple logistic regression analysis.

This result suggests that a higher counseling skill level can lead to higher abstinence rates.