Formal treatment, self-help, or no treatment for alcohol-use disorders? Evidence from the National Household Survey...

The purpose of this study was to examine further alcohol treatment choice by using data from a nationally representative sample of adults with alcohol-use disorders to test which of three models-sequential, multinomial, or nested best fit the data. The goals were to provide evidence about how this choice was made and to provide improved coefficient estimates, as well as to inform future analyses of treatment choice.

Data from the 2000 National Household Survey of Drug Abuse include respondents ages 18-64 reporting symptoms consistent with Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) diagnoses of alcohol abuse or dependence. A nested multinomial framework is used to determine the preferred model and to estimate the effect of respondents' characteristics on the decisions to receive help and what kind of help to receive.

A sequential model, in which the choice of whether to receive help is unaffected by the level of satisfaction afforded by the alternatives, best fit the data. Older respondents had higher odds of both receiving help and choosing self-help, and those with a DSM-IV diagnosis of abuse had lower odds of receiving help but higher odds of entering self-help.

The decision to receive help for alcohol problems appears unaffected by the perceived differences between these two broad categories of alternatives: self-help or formal treatment. This result may indicate the need to provide more information on the full range of treatment options to those for whom self-help may not be sufficient.