Evaluation of Brief Intervention

Abstract

I have several questions regarding the article on the evaluation of a brief intervention for alcohol use by Bazargan-Hejazi et al. 1Bazargan-Hejazi S. Bing E. Bazargan M. et al.Evaluation of a brief interventaion in an Inner-City Emergency Department.Ann Emerg Med. 2005; 46: 67-76Abstract Full Text Full Text PDF PubMed Scopus (91) Google ScholarFirst, the data in Table 3, on which the major claims of the paper rests, is inconsistent. The absolute numbers and the percentages for the moderate-risk group do not match. If the percentages in the table (34% v. 13%) are correct, then the difference between the intervention and the control groups is not statistically significant (diff = .21, 95% CI -.01 to .43).Second, although the authors have shown at entry, the intervention and the control groups were similar in most baseline characteristics (Table 1), it is not clear that the patients in the final analysis were also similar at baseline. This may be important given the attrition rate of 37%.Third, several variables (eg, marital status, housing status) were omitted from the logistic regression model. In particular, patients without social supports may have a more difficult time with at-risk drinking, and represent confounding variables. Thus, the results of the logistic model may be flawed.Fourth, was there a difference (as stratified by intervention group v. control group) in the patients who increased their AUDIT scores at 3 months? If the intervention group resulted in as many increases in AUDIT scores as in decreases, then the intervention may not have a beneficial effect overall.Finally, of 1058 patients screened, the intervention appears to have resulted in 5 patients who improved their AUDIT scores, a success rate of <1%. Thus, I question the effectiveness of the program as it stands, especially since the outcome (reduction of AUDIT scores in 3 months) may not be clinically meaningful. I have several questions regarding the article on the evaluation of a brief intervention for alcohol use by Bazargan-Hejazi et al. 1Bazargan-Hejazi S. Bing E. Bazargan M. et al.Evaluation of a brief interventaion in an Inner-City Emergency Department.Ann Emerg Med. 2005; 46: 67-76Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar First, the data in Table 3, on which the major claims of the paper rests, is inconsistent. The absolute numbers and the percentages for the moderate-risk group do not match. If the percentages in the table (34% v. 13%) are correct, then the difference between the intervention and the control groups is not statistically significant (diff = .21, 95% CI -.01 to .43). Second, although the authors have shown at entry, the intervention and the control groups were similar in most baseline characteristics (Table 1), it is not clear that the patients in the final analysis were also similar at baseline. This may be important given the attrition rate of 37%. Third, several variables (eg, marital status, housing status) were omitted from the logistic regression model. In particular, patients without social supports may have a more difficult time with at-risk drinking, and represent confounding variables. Thus, the results of the logistic model may be flawed. Fourth, was there a difference (as stratified by intervention group v. control group) in the patients who increased their AUDIT scores at 3 months? If the intervention group resulted in as many increases in AUDIT scores as in decreases, then the intervention may not have a beneficial effect overall. Finally, of 1058 patients screened, the intervention appears to have resulted in 5 patients who improved their AUDIT scores, a success rate of <1%. Thus, I question the effectiveness of the program as it stands, especially since the outcome (reduction of AUDIT scores in 3 months) may not be clinically meaningful. Evaluation of a Brief Intervention in an Inner-city Emergency Department - reply to #2005-677Annals of Emergency MedicineVol. 47Issue 2PreviewWe thank the author of this letter for his interest in the topic of brief intervention to reduce high-risk drinking, and we are happy to address the questions that are raised. Full-Text PDF