Are Alcoholism Treatments Effective? The Project MATCH Data: Response

William R Miller

doi:10.1186/1471-2458-5-76

Abstract

A response to Cutler RB, Fishbain, DA: Are alcoholism treatments effective? The Project MATCH data BMC Public Health 2005, 5:75.

Highlights

Disappointment with study findings, depends upon what the beholder had hoped to see in the first place
The surprise, was not the absence of between-treatment differences on the two primary outcome variables, but that on at least one other time-honored outcome measure – the percentage of patients maintaining complete abstinence – those in the Twelve-Step Facilitation treatment fared significantly better at all follow-up points than did patients in the other two conditions – a substantial advantage of about 10 percentage points that endured across 3 years [5]
It is true that most of the primary patient-treatment matching predictions were not confirmed in early followup; instead, some large predicted matching effects were found with a priori hypothesis tests that had been assigned to secondary status [7], or emerged at 3-year follow-up [3]

Summary

Introduction

Disappointment with study findings, depends upon what the beholder had hoped to see in the first place. To support their inaccurate claim that "essentially no patient-treatment matches were found," Cutler and Fishbain observed that "some 504 hypotheses were tested." The primary and secondary matching hypotheses numbered 19, all of which had been carefully specified a priori, with statistical protection against Type I error.

Results

Conclusion