Elsevier

Social Science & Medicine

Volume 64, Issue 2, January 2007, Pages 462-476
Social Science & Medicine

A multilevel decomposition approach to estimate the role of program location and neighborhood disadvantage in racial disparities in alcohol treatment completion

https://doi.org/10.1016/j.socscimed.2006.08.032Get rights and content

Abstract

Large racial disparities in completion rates from substance abuse treatment programs in urban settings remain largely unexplained, although evidence is accumulating that neighborhood conditions may influence individual substance abuse patterns and consequences. Understanding racial disparities in alcohol treatment completion, in particular, is crucial to resolving health disparities because racial/ethnic minorities bear a disproportionate burden of alcohol-related health consequences. Patient records for all non-homeless African American (N=1677), Hispanic (N=1635), and white (N=1216) alcohol outpatients, ages 18 or older, discharged during 1998–2000 from publicly funded treatment programs in Los Angeles County, the second largest system of publicly funded substance abuse treatment in the United States, were combined with census data. We tested the hypothesis that racial differences in treatment completion are related to differences in neighborhood context, particularly neighborhood-level disadvantage. Estimates from multilevel statistical models indicate that treatment neighborhood disadvantage is independently associated with treatment completion after controlling for patient characteristics and facility- and zip code-level random effects. Results of a Oaxaca decomposition of the regression estimates indicate that racial differences in treatment neighborhood disadvantage account for 32.3% of African American–white differences in treatment completion. Hispanic–white differences in completion, and the effect of home neighborhood disadvantage on completion, were non-significant. We conclude that the location of publicly funded alcohol treatment programs is related to racial disparities in treatment completion, but additional research is necessary to understand the mechanism behind this association.

Introduction

Racial and ethnic disparities in rates of retention in alcohol treatment programs have been documented in several recent studies. In the Project MATCH clinical trial, relatively lower completion rates by African American and Hispanic patients were detected, but lost statistical significance after controlling for occupation (Tonigan, 2003). Yet in a random sample of treatment clients from a large California health maintenance organization, significant differences in retention persisted after controlling for addiction severity, self-reported treatment goals, sex, age, income, and employment among African-American women (but not men) compared to whites (Mertens & Weisner, 2000). At public substance abuse treatment programs in Los Angeles, differences in retention between African American and white patients net of patient-level controls have been present at least since the early 1990s (Hser, Joshi, Maglione, Chou, & Anglin, 2001; Jacobson, Robinson, & Bluthenthal, 2006) and have been documented more recently between Hispanic and white patients (Longshore et al., 2004). However, the cause of these disparities remains unclear.

Understanding racial disparities in alcohol treatment outcomes is imperative because historically African Americans and other minority populations generally have experienced up to five times the rate of alcohol morbidity and mortality as whites in the United States, despite similar rates of lifetime heavy drinking (Caetano, 2003; Grant et al., 2004; Group for the Advancement of Psychiatry, 1996). Considering that alcohol consumption is causally related to more than 60 medical conditions (Room, Babor, & Rehm, 2005), reducing racial/ethnic disparities in alcohol-related problems is of special relevance to the national priority of reducing disparities in health (US Department of Health and Human Services, 2000). Disparities in treatment effectiveness could contribute to differences in alcohol-related health outcomes.

Although treatment for alcohol abuse and alcoholism has generally been shown to be an effective intervention to reduce health problems and expenditures (Holder & Blose, 1992; Holder et al., 2000), retention in treatment for alcohol and other classes of substance abuse is widely considered to be “one of the greatest problems interfering with treatment effectiveness” (Stahler, Cohen, & Shipley, 1993). To date, retention studies have examined a broad range of factors including patient demographics, employment, living situation, family background, addiction severity, and treatment program process and components (e.g. Cacciola, Dugosh, Foltz, Leahy, & Stevens, 2005; De Leon, Hawke, Jainchill, & Melnick, 2000; Hser, Evans, & Huang, 2005; Hser et al., 2001; Joe, Simpson, & Broome, 1998; Maglione, Chao, & Anglin, 2000; Mertens & Weisner, 2000; Morrissey et al., 2005; Simpson, Joe, Broome et al., 1997). Research has only recently ventured beyond the patient and program to understand the influence of the larger community context that patients face as they attempt recovery. But in attempts to do so, attention has focused on the important, though narrow issues of distance-to-treatment and proximity to primary health and mental health services (Beardsley, Wish, Fitzelle, O’Grady, & Arria, 2003; Friedmann, D’Aunno, Jin, & Alexander, 2000; Friedmann, Lemon, Stein, Etheridge, & D’Aunno, 2001; Marsh, D’Aunno, & Smith, 2000; Umbricht-Schneiter, Ginn, Pabst, & Bigelow, 1994).

This paper examines whether broader differences in community context experienced by racial and ethnic minorities could be one factor contributing to lower rates of alcohol treatment completion among minority patients. With respect to drug use behavior, Lillie-Blanton, Anthony, and Schuster (1993) found that racial/ethnic differences in crack cocaine use lose their statistical significance when controlling for neighborhood clustering, although the specific neighborhood characteristics that underlie this result were not investigated. More recent studies have shown that drinking and illicit substance abuse in several urban contexts are related to neighborhood levels of poverty and disorganization (i.e. disadvantage) net of individual-level controls (Boardman, Finch, Ellison, Williams, & Jackson, 2001; Hill & Angel, 2005; Saxe et al., 2001). These studies have drawn on social ecology theory (Bronfenbrenner, 1996; Kelly, 1966; Moos, 1973) and strain and social disorganization theories (Agnew, 1992; Merton, 1938) to argue that substance abuse can be explained in part as a response to neighborhood environments that cause stress or strain and lack the community-level organization or collective efficacy to sanction substance abuse behavior.

However, it is questionable whether racial disparities in neighborhood context in the general population (Kasarda, 1993) are mirrored in subpopulations accessing publicly funded alcohol treatment programs. Furthermore, while neighborhood conditions associated with patterns of substance abuse could conceivably influence rates of treatment entry, it is not necessarily the case that they also influence patterns of treatment completion in the same way. This issue was recently examined by Jacobson (2004), who suggested a number of causal mechanisms that could link neighborhood disadvantage to treatment completion based on a review of related literature. We briefly summarize the principal hypotheses forwarded in this review below. For a more detailed discussion, readers are referred to the full article.

First, it is argued that because psychological stress is independently associated with dropout at the individual level (Brown et al., 1998; Hiller, Knight, & Simpson, 1999), the same neighborhood-level psychological stressors thought to be associated with substance abuse—less access to reliable employment, higher levels of criminal victimization, harsher living conditions, and social stigmatization from living in a disadvantaged area (Boardman et al., 2001)—may impede treatment completion. Second, Tucker, Vuchinich, and Gladsjo (1990–1991) argue that the relative paucity in disadvantaged areas of establishments that facilitate everyday tasks—grocery stores and other retail establishments, financial institutions, and health care facilities—may contribute to relapse by increasing the burden of “daily hassles” on residents. Relapse, in turn, is inversely related to treatment completion (Simpson, Joe, Rowan-Szal, & Greener, 1997). Additionally, lower expected education and earnings in disadvantaged areas (Borjas, 1995; Cutler & Glaeser, 1997; Massey, 1990) could affect patients’ self-efficacy and expected benefit from treatment, and thus their willingness to stay on when attendance becomes difficult (Hiller et al., 1999). Finally, greater physical availability of alcohol in disadvantaged areas (Alaniz, 1998; Storr, Chen, & Anthony, 2004) may expose patients in these areas to more environmental triggers for relapse to the extent that it increases actual availability, audiovisual cues for use, or the expectation of consumption (Rosenhow, Niaura, Childress, Abrams, & Monti, 1991).

These arguments provide a theoretical basis for expecting an inverse association between neighborhood disadvantage and alcohol treatment completion. In this paper, we do not test these causal mechanisms directly, but instead estimate a reduced-form relationship and the contribution of that relationship to racial disparities in treatment completion. Our analysis is guided by three research questions:

  • (1)

    Relative to white patients, are African American and Hispanic patients exposed to higher levels of neighborhood disadvantage either where they live or attend treatment?

  • (2)

    If so, are these neighborhood differences independently related to individual alcohol treatment completion?

  • (3)

    How much of differences in treatment completion between these groups can be explained by patient-, facility-, and neighborhood-level differences, and specifically by neighborhood disadvantage?

In addressing the third research question, this paper employs a post-regression decomposition—to date uncommon in health disparities research—to directly estimate the share of the racial disparity associated with individual-, facility-, and neighborhood-level race differences.

Beyond the neighborhoods where patients live, this research also advances the health disparities literature by investigating the location where services are delivered, the “treatment context”. Jacobson (2004) emphasized attention to both residential and treatment contexts as part of a broader “treatment ecology” view encompassing all of the “principle geographic contexts experienced by the client over the course of treatment and the interrelationships among them.”

Section snippets

Data

The study area is Los Angeles County (LAC), California, the second largest publicly funded substance abuse treatment system in the United States in terms of treatment admissions (US Department of Health and Human Services, 2003). Levels of racial residential segregation with respect to both Hispanics and African Americans in LAC are high (McConville, Ong, Houston, & Rickles, 2001), providing a context to examine neighborhood differences.

Patient-level data are standardized intake and discharge

Analytic strategy

An important concern for studies of neighborhood influences on outcomes is confounding from facility- and individual-level confounders. We use a multilevel statistical framework to guard against such confounding and a post-regression decomposition to estimate the share of the racial disparity associated with racial differences at each level.

Three stages of analysis correspond to the three research questions. First, χ2 tests (for dichotomous variables) and t-tests (for continuous variables) were

Differences between white, African American, and Hispanic Patients

Table 1 presents the bivariate findings. Mean neighborhood disadvantage for African Americans is approximately 13 times higher in home neighborhoods and 7 times higher in treatment locations compared to whites. Hispanic patients fall between these groups. The three groups also differ with respect to patient characteristics that could be related to completion. Because Hispanic–white differences in treatment completion were not detected, we describe only the African American–white differences

Discussion

What explains observed racial/ethnic differences in completion rates at publicly funded alcohol treatment programs? Previous analysis of treatment outcomes in Los Angeles found that differences in patient characteristics explained 40% of white–African American differences in treatment completion among alcohol outpatients (Jacobson et al., 2006). This paper extends the inquiry to the role of neighborhood context both where patients live and attend treatment, in a subsample of Los Angeles public

Acknowledgments

This study was supported by National Institutes of Health Grants NIAAA R21AA013813, NCRR G12RR0302618, and CH05-DREW-616 to Charles R. Drew University of Medicine and Science. We thank Rick Rawson of UCLA Integrated Substance Abuse Programs for supporting this research and John Bacon and Tom Tran of the Los Angeles County Alcohol and Drug Programs Administration for assistance with data preparation.

References (71)

  • D.D. Simpson et al.

    Drug abuse treatment process components that improve retention

    Journal of Substance Abuse Treatment

    (1997)
  • J.J. Sudano et al.

    Explaining US racial/ethnic disparities in health declines and mortality in late middle age: The roles of socioeconomic status, health behaviors, and health insurance

    Social Science & Medicine

    (2006)
  • R. Agnew

    Foundation for a general strain theory of crime and delinquency

    Criminology

    (1992)
  • M.L. Alaniz

    Alcohol availability and targeted advertising in racial/ethnic minority communities

    Alcohol Health and Research World

    (1998)
  • Costs of alcohol abuse to Los Angeles County (2003): Technical report

    (2004)
  • M.D. Anglin et al.

    Treatment of drug abuse

  • R.J. Blendon et al.

    Access to medical care for black and white Americans: A matter of continuing concern

    Journal of the American Medical Association

    (1989)
  • R.N. Bluthenthal et al.

    Characteristics of malt liquor beer drinkers in a low-income, racial minority community sample

    Alcoholism: Clinical and Experimental Research

    (2005)
  • J.D. Boardman et al.

    Neighborhood disadvantage, stress, and drug use among adults

    Journal of Health and Social Behavior

    (2001)
  • G.J. Borjas

    Ethnicity, neighborhoods, and human-capital externalities

    The American Economic Review

    (1995)
  • U. Bronfenbrenner

    The ecology of human development: Experiments by nature and design

    (1996)
  • R.A. Brown et al.

    Depression among cocaine abusers in treatment: Relation to cocaine and alcohol use and treatment outcome

    American Journal of Psychiatry

    (1998)
  • R. Caetano

    Alcohol-related health disparities and treatment-related epidemiological findings among whites, blacks and hispanics in the United States

    Alcoholism: Clinical and Experimental Research

    (2003)
  • P. Congdon

    Bayesian statistical modelling

    (2001)
  • P. Congdon

    Applied Bayesian modelling

    (2003)
  • D.M. Cutler et al.

    Are ghettos good or bad?

    The Quarterly Journal of Economics

    (1997)
  • I.A. de la Rosa

    Perinatal outcomes among Mexican Americans: A review of an epidemiological paradox

    Ethnicity and Disease

    (2002)
  • K. Fiscella et al.

    Health disparities based on socioeconomic inequities: Implications for urban health care

    Academic Medicine

    (2004)
  • M.E. Ford et al.

    Conceptualizing and categorizing race and ethnicity in health services research

    Health Services Research

    (2005)
  • P.D. Friedmann et al.

    Medical and psychosocial services in drug abuse treatment: Do stronger linkages promote client utilization?

    Health Services Research

    (2000)
  • P.D. Friedmann et al.

    Linkage to medical services in the drug abuse treatment outcome study

    Medical Care

    (2001)
  • E. Fuentes-Afflick et al.

    Low birth weight and Latino ethnicity: Examining the epidemiologic paradox

    Archives of Adolecent Medicine

    (1997)
  • Alcoholism in the United States: Racial and ethnic considerations

    (1996)
  • J. Hewko et al.

    Measuring neighbourhood spatial accessibility to urban amenities: Does aggregation error matter?

    Environment and Planning A

    (2002)
  • M.L. Hiller et al.

    Risk factors that predict dropout from corrections-based treatment for drug abuse

    The Prison Journal

    (1999)
  • Cited by (51)

    • Race, psychosocial characteristics, and treatment outcomes among individuals undergoing treatment for cannabis use disorder: A latent profile analysis based on preferred method of using cannabis

      2021, Journal of Substance Abuse Treatment
      Citation Excerpt :

      African Americans and Hispanic Americans are known to have less access to substance abuse treatment relative to White Americans (Alegria et al., 2011; Guerrero et al., 2013). Furthermore, disparities exist within treatment, with racial/ethnic minorities being less likely to complete treatment (Cooper et al., 2010; Jacobson et al., 2007). Also, neighborhood disadvantages are associated with lower rates of program completion, lack of adequate health insurance coverage of treatment, and limited availability of facilities that accept Medicaid (Cummings et al., 2014; Jacobson et al., 2007).

    • Employment after beginning treatment for substance use disorders: The impact of race/ethnicity and client community of residence

      2018, Journal of Substance Abuse Treatment
      Citation Excerpt :

      In our particular case, the community level covariates subjected to factor analysis denoted census tract percentages of the following: residents who are unemployed, residents living in poverty, residents in management/professional occupations (recoded; for employed civilian population, ages 16+), female-headed households with children, households with annual income >$75,000 (recoded). Previous studies demonstrated that these variables might affect the outcomes of substance abuse treatment (Jacobson, Robinson, & Bluthenthal, 2007; Karriker-Jaffe, 2013; Karriker-Jaffe et al., 2012; Mennis, Stahler, & Baron, 2012; Molina et al., 2012). However, in our factor analysis, the Kaiser criterion (Kaiser, 1960) led us to retain one factor, which loaded most heavily on income and occupation variables and more modestly on education and female head of household variables.

    • Does the implementation of evidence-based and culturally competent practices reduce disparities in addiction treatment outcomes?

      2017, Addictive Behaviors
      Citation Excerpt :

      This dichotomous measure was coded 1 if clients met the following criteria: (a) the client reported no alcohol or drug use during the 30 days prior to discharge, (b) the clinician reported client sobriety at discharge, and (c) the clinician coded treatment episode as successful based on the client meeting treatment goals for that episode. Using three indicators of success, this measure of treatment completion is more comprehensive than other studies that have relied on a single variable (Jacobson, Robinson, & Bluthenthal, 2007; SAMHSA, 2009a). Program staff rated implementation of CMT and implementation of MAT on 5-point Likert scales (1 = never to 5 = always) according to how often they were used in their program.

    • Health in the Tenderloin: A resident-guided study of substance use, treatment, and housing

      2017, Social Science and Medicine
      Citation Excerpt :

      For example, the myriad of neighborhood disadvantages related to residential racial segregation has been described as a fundamental cause of racial health disparities, including substance use disparities (Williams and Collins, 2001). Measures of neighborhood disadvantage such as Black and Hispanic socioeconomic status and housing instability (Saloner and Lê Cook, 2013) have been linked to racial disparities in completing substance use treatment (Jacobson et al., 2007). Together studies like these indicate that social position within an environment, defined by factors like race and class background, shape substance use and treatment experiences.

    View all citing articles on Scopus
    View full text