Subtypes of Treatment Use for Alcohol Use Disorders in the Nesarc: A Latent Class Analysis

Background and Purpose: Much social work research considers treatment for alcohol use disorders (AUDs) a binary event, yet addiction treatment is often a patterned and multifaceted lifetime process. It is known that individuals access a broad array of alcohol treatment services in their community (Weisner, 2001), which suggests the need to identify treatment use patterns across a wide range of available service sectors. In the current study, we use Latent Class Analysis (LCA) and a large, nationally representative dataset to identify whether subtypes of treatment use for AUDs exists in the U.S. general population. We employed the Andersen and Aday framework (1974) to consider predisposing, enabling, and need characteristics associated with treatment use subtypes to inform efforts (e.g. Healthy People 2020) that aim to boost alcohol treatment receipt in the U.S..

Methods: We analyzed Waves 1 and 2 of the National Epidemiologic Survey of Alcohol Related Conditions, a nationally representative survey of non-institutionalized U.S. adults (2001-2005). Psychiatric diagnoses, sociodemographic characteristics, and treatment use were ascertained via the AUD and Associated Disabilities Interview Schedule-IV (Grant, 2003). Our sample included individuals (n=1,667) who ever met criteria for AUD and reported prior alcohol treatment. Alcohol treatment variables, which were our latent class indicators, included participants’ use of 13 types of common alcohol treatment services. Sociodemographic (age, gender, household income, marital status & insurance status) and psychiatric diagnoses (lifetime abuse vs. dependence AUD diagnosis, any non-nicotine or caffeine lifetime drug use disorder, and any lifetime mental illness) were covariates. With Mplus, we specified LCAs with and without covariates, then considered model fit statistics and substantive theory to identify the appropriate latent class solution. We employed survey design variables (e.g. weights) to obtain accurate standard error estimates for the population.

Results: Four subtypes of treatment users emerged through LCA; Multiservice Users (9.7% of the sample), Alcoholics Anonymous (AA) Alone (36.4%), AA Paired with Specialty Addiction Service (29.7%), and Private Medical Professionals (24.1%). Estimated means of model covariates showed divorce, public insurance and high rates of co-occurrence characterized the Multiservice Users class. Young men with low prevalence of alcohol dependence and low rates of co-occurrence characterized the Alcoholics Anonymous Alone class. Older aged men with lower income and no insurance characterized membership of AA Paired with Specialty Addiction Service class. Married persons, high income, and private insurance characterized the Private Medical Professionals class.

Conclusions and Implications: Subtypes of treatment use patterns clearly exist among those who seek treatment for alcohol use disorders. Age, income, and illness severity were associated with treatment profiles, revealing potential disparities (e.g. access to private medical professionals). Longitudinal analyses and improved measurement is needed to assess whether treatment patterns reflect service preference versus availability. Future research associated with treatment use for AUDs needs to move beyond simply examining treatment use as a binary event. Through identifying patterns of treatment use for AUDs, social workers may be able to develop targeted interventions that refer and facilitate access to appropriate AUD services that match client needs.