Advertisement
Regular article| Volume 36, ISSUE 1, P75-86, January 2009

Patient predictors of alcohol treatment outcome: A systematic review

Published:July 28, 2008DOI:https://doi.org/10.1016/j.jsat.2008.05.007
Patient predictors of alcohol treatment outcome: A systematic review
Previous ArticleEvaluation of two Web-based alcohol interventions for mandated college students
Next ArticleAlcoholics Anonymous attendance following 12-step treatment participation as a link between alcohol-dependent fathers' treatment involvement and their children's externalizing problems

      Abstract

      Patient characteristics as predictors of alcohol use disorder treatment outcome were examined on three levels, identifying whether or not variables were significant predictors of drinking-related outcome in univariate analysis, in multivariate analysis, and in multivariate analyses limited to studies including several “key predictors.” Also, a model was developed to predict total percentage of variance in treatment outcome accounted for in each study using each of the key predictors and a range of methodological factors. The most consistent univariate predictors were baseline alcohol consumption, dependence severity, employment, gender, psychopathology rating, treatment history, neuropsychological functioning, alcohol-related self-efficacy, motivation, socioeconomic status/income, treatment goal, and religion. When these key predictors were combined into multivariate analyses, baseline alcohol consumption and gender showed substantial reductions in predictive consistency whereas the remaining variables were not greatly affected. The most consistent predictors overall were dependence severity, psychopathology ratings, alcohol-related self-efficacy, motivation, and treatment goal. The two predictor variables most associated with greater variance accounted for in predictive models, when controlling for broader methodological variables, were baseline alcohol consumption and dependence severity. Few predictor variables were examined in more than a third of studies reviewed, and few variables were found to be significant predictors in a clear majority of studies. However, a subset of variables was identified, which collectively could be considered to represent a consistent set of predictors. Too few studies controlled for other important predictor variables. Attempts to synthesize findings were often hampered by lack of agreement of the best measure for predictor variables.

      Keywords

      1. Introduction

      Prediction of treatment outcome provides the opportunity to deliver three key benefits to the clinical setting: identifying specific client groups achieving poorer outcomes, identifying areas to target in treatment, and improving accuracy of prognosis.
      Identification of factors predictive of outcome may allow for identification of populations predicted to have poorer outcome, such as for demographic predictors, or clinical variables with relative stability over time, such as is the case with a number of diagnoses, for example, social phobia or comorbid substance use disorders. Treatment may consequently be tailored to better meet the needs of such groups to reduce the disparity with respect to outcome.
      In contrast to predictive variables that are generally stable, those that may be considered more malleable can be identified as targets for manipulation as part of the treatment process. For example, where unemployment is found to predict poorer outcome, it is possible that treatment would be enhanced by actively addressing employment.
      Improved prognostication allows a clinician to better inform the client and his or her family as to what may lie ahead. It allows for improved treatment planning with respect to intervention type, duration, and intensity (
      • Kadden R.M.
      • Skerker P.M.
      Treatment decision making and goal setting.
      in: McCrady B.S. Epstein E.E. Addictions: A comprehensive guidebook. Oxford University Press, New York1999
      ). It should enable the clinician to set more realistic treatment goals and guide the client in his or her own goal setting.
      The search for predictors of treatment outcome may be seen to encompass attempts to achieve two different functions. One is to identify causal risk factors, while the other is to find consistent correlates of outcome irrespective of causality. This latter function identifies risk factors that may not be causal as their correlation with outcome is merely a product of their correlation with a strong causal risk factor. Such risk factors are thus termed “proxy risk factors” (
      • Kraemer H.C.
      • Stice C.
      • Kazdin A.
      • Offord D.
      • Kupfer D.
      How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors.
      American Journal of Psychiatry. 2001; 158: 848-856
      ). Two implications of these different functions are the timing of potential predictor variable measurement and the method of analysis.
      For predictive purposes, potential predictors could be measured at different times relative to the delivery of treatment. If prediction is to be used to determine the allocation of treatment resources, then, clearly, this is best achieved by predicting prognosis early in this process. On the other hand, to understand the process of relapse, one could measure the independent variables at a much wider variety of times, with measurement closest to the point of relapse likely to be most informative (
      • Connors G.J.
      • Maisto S.A.
      • Zywiak W.H.
      Understanding relapse in the broader context of post-treatment functioning.
      Addiction. 1996; 91: S173-S189
      ).
      The second implication relates to the role of covariates. To successfully identify the unique contribution of a variable to predicting treatment outcome, one must be able to partial out known covariates of relapse. Such covariates can either be proxy risk factors, which would no longer predict outcome if appropriate covariates were included in the analysis, or overlapping risk factors, which would be found to have diminished predictive strength when examined concurrently.
      Better understanding of consistent predictors of treatment outcome is important, as clinical prediction of human behavior and traits has been shown to be inferior to that produced by mechanical or computational means (
      • Breslin F.C.
      • Sobell M.B.
      • Sobell L.C.
      • Buchan G.
      • Cunningham J.A.
      Toward a stepped care approach to treating problem drinkers: The predictive utility of within-treatment variables and therapist prognostic ratings.
      Addiction. 1997; 92: 1479-1489
      ,
      • Grove W.M.
      • Zald D.H.
      • Lebow B.S.
      • Snitz B.E.
      • Nelson C.
      Clinical versus mechanical prediction: A meta-analysis.
      Psychological Assessment. 2000; 12: 19-30
      ). Reasons suggested for the poorer performance of clinicians, even those with substantial experience and training, included ignoring base rates, assigning nonoptimal weights to cues, failure to take into account regression toward the mean, and failure to properly assess covariation. Grove et al. (2000) also point out that clinicians often do not receive sufficient feedback on the accuracy of their judgments with which they could modify judgment bias.
      • Gibbs L.
      • Flanagan J.
      Prognostic indicators of alcoholism treatment outcome.
      The International Journal of the Addiction. 1977; 12: 1097-1141
      examined 45 studies reporting on the prediction of alcoholism treatment outcome, collating 208 candidate predictor variables. They concluded that no stable predictor characteristics could be identified, although by taking a more liberal definition of a “somewhat stable” predictor, they were able to identify psychopathology, Wechsler arithmetic scores, steady work history, being married/cohabiting, employment, higher status occupation, fewer arrests, history of Alcoholics Anonymous (AA) contact, and higher social class as predictors of better outcome. The authors go on to highlight methodological weaknesses that they believed had hampered their task, including sampling error, high attrition, alcoholism poorly defined, heterogeneous treatments, inconsistent outcome criteria, varying time to follow-up, and small samples.
      Two reviews of baseline patient predictors of outcome for broader substance use treatment are worthy of mention.
      • McKay J.R.
      • Weis R.V.
      A review of temporal effects and outcome predictors in substance abuse treatment studies with long-term follow-ups.
      Evaluation Review. 2001; 25: 113-161
      identified 12 studies with a minimum 2-year follow-up period, 7 of which were for alcohol treatment studies. The most consistent baseline predictors of outcome were pretreatment level of substance use and psychiatric severity, although, in both cases, these predicted worse outcome for approximately 65% of significant reports and better outcome for 20% of reports, with the remainder relating to treatment interaction effects. Motivation and coping were also found to be relatively consistent predictors of outcome but were much less frequently studied. Demographic variables were poor predictors. The authors did not comment on whether or not there were differences across studies that might be ascribed to the substance treatment modality (i.e., alcohol vs. other drug).
      In a meta-analysis of 69 studies predicting continued drug use by opioid treatment clients,
      • Brewer D.D.
      • Catalano R.F.
      • Haggerty K.
      • Gainey R.R.
      • Fleming C.B.
      A meta-analysis of predictors of continued drug use during and after treatment for opiate addiction.
      Addiction. 1998; 93: 73-92
      identified 10 significant predictors of future drug use. These were high levels of pretreatment opioid/other drug use, prior treatment for opioid addiction, no prior abstinence from opioids, abstinence from/light use of alcohol, depression, high stress, unemployment/employment problems, association with substance abusing peers, short length of treatment, and leaving treatment prior to completion.
      While these reviews may be applicable to alcohol misuse populations, there is clearly the need to systematically examine the large number of studies specific to alcohol treatment that have identified patient predictors of treatment outcome over the past 30 years. In addition, such a review provides the opportunity to extend earlier work by taking a closer look at how multivariate analysis can aid in identifying more potent causal risk/protective factors (
      • Kraemer H.C.
      • Stice C.
      • Kazdin A.
      • Offord D.
      • Kupfer D.
      How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors.
      American Journal of Psychiatry. 2001; 158: 848-856
      ).

      2. Method

      2.1 Study identification and selection

      Online databases searched were PsychLit, Medline, and Embase. Keywords, mapped onto the subject heading Alcoholism, were prognosis, prediction, prospective study or risk factor, and follow-up, treatment outcome or recurrence. The search was restricted to human subjects, English language, adult (18–64 years) population, and publication date 1977–2005. Additional papers were identified via citations in other reviewed papers.
      Studies identified in the above searches were included only if they met the following criteria:
      • 1.
        Most participants must have been identified as alcohol dependent or alcoholic, either using formal diagnostic criteria, by virtue of scores on a dependence severity scale, or as a general requirement of entry to the treatment program from which participants were drawn, or have been identified as having problematic drinking of a severity requiring treatment. Three included studies investigated hazardous drinking treatment samples, but each was deemed to contain a large majority likely to meet criteria for a diagnosis of alcohol abuse or dependence based on dependence scale mean scores and mean weekly drinking levels (
        • Heather N.
        • Brodie J.
        • Wale S.
        • Wilkinson G.
        • Luce A.
        • Webb E.
        • et al.
        A randomized controlled trial of moderation-oriented cue exposure.
        Journal of Studies on Alcohol. 2000; 61: 561-570
        ,
        • Heather N.
        • Rollnick S.
        • Bell A.
        Predictive validity of the Readiness to Change Questionnaire.
        Addiction. 1993; 88: 1667-1677
        ,
        • Kavanagh D.J.
        • Sitharthan T.
        • Sayer G.P.
        Prediction of results from correspondence treatment for controlled drinking.
        Addiction. 1996; 91: 1539-1545
        ).
      • 2.
        Participants must have undergone some form of treatment for their alcohol misuse.
      • 3.
        Studies must have attempted to predict drinking status at a point at least 3 months following the completion of treatment.
      • 4.
        Prediction must have been based on data gathered prior to or during treatment.
      • 5.
        A minimum follow-up rate of 65% was required. Studies not reporting a follow-up rate were excluded, whereas studies employing survival analysis that code all participants lost to follow-up as relapsed were retained.
      • 6.
        Finally, a minimum sample size of 40 at follow-up was required.

      2.2 Data analysis

      Several studies are represented by more than one paper included in this review. When the same predictor variable is examined in more than one publication from the same study, only one instance is reported. If the two publications report on the predictive power of a variable for different follow-up intervals, for different outcome measures, or using more than one instrument to measure the predictor variable, then finding a significant effect takes precedence over nonsignificant findings, as is also the case for multiple analyses within a single paper. Hence, for example, a predictor variable is deemed to be significant if a significant predictive relationship is found for any follow-up period of 3 months or more. This increases the probability of finding a significant association. However, conclusions are only drawn on aggregate findings and care has been taken to ensure that where a single predictor variable is disproportionately examined in studies employing multiple follow-up points or measures, this is highlighted.
      Significant interaction effects between baseline characteristics are considered to be significant findings for each variable included in the interaction, as the two (or more) variables are contributing to a significant prediction even if they do not have significant main effects.
      In multivariate analyses, if a variable was found to be significant with a limited number of covariates but was no longer significant when further variables were added, then this is recorded as not significant in multivariate analysis, as the intent is to test the significance of individual variables with the benefit of maximum covariation.
      The current review examined predictors of treatment outcome on three levels: All variables were first examined as significant predictors of drinking-related outcome in univariate analysis and, secondly, multivariate analysis. Thirdly, multivariate results were reexamined for variables (“key predictors”) found to be significant in approximately half or more studies and were limited to studies including a minimum of four key predictors to provide a stricter test of predictive stability.

      3. Results

      The literature search yielded 63 published papers describing findings from 51 unique treatment outcome studies. All potential predictor variables were initially examined. Only those reported for four or more studies were included in this review. Variables excluded due to their low frequency included social support, locus of control, coping behavior, childhood disruptive disorders, sexual and physical abuse history, alcoholism typology, general self-esteem, and outcome expectancies. Thirty-one baseline predictors were identified from these papers. The ability of these variables to predict drinking-related treatment outcome is shown in Table 1. Variables are grouped as demographic and social functioning measures, substance-related measures, and other clinical measures and then ranked in order from those most investigated to those least investigated. The ability of each variable to predict a drinking-related outcome in univariate and multivariate analyses is displayed. The nature of the significant associations is summarized in simplified form. The final column displays the percentage of studies that found the variable to be a statistically significant predictor in either univariate or multivariate analysis.
      Table 1Univariate and multivariate predictors of alcohol-consumption-related treatment outcome
      VariableUnivariateMultivariateBetter outcome predicted byTotal studiesPercent significant in univariate or multivariate analysis
      PredictorNonpredictorPredictorNonpredictor
      Demographic and social functioning measures
       Ageg, w, Dc, o, p, y, M, O, Sz, I, K, Ua, c, h, m, o, p, q, s, u, x, y, A, J, M, R, T, XOlder × 52627
      Younger × 2
       Employmentf, p, G, Ic, r, wa, h, n, I, J, Ud, p, q, r, u, v, x, G, R, TEmployed × 51947
      Better history × 2
      Lower ASI employment problems × 2
       Genderc, k, o, Qg, pc, z, K, N, Q, U, Xa, d, o, p, u, v, A, J, TFemale × 81850
      Male × 1
       Marital statusp, Dc, g, r, w, Mp, Ua, d, q, r, u, x, J, M, R, TUnmarried × 11619
      Married × 2
       Educationc, pg, o, r, w, Da, h, p, Uc, d, o, r, s, A, I, JMore educated × 51533
       Socioeconomic status/incomeMgm, n, q, Md, T, WHigher socioeconomic status/income × 3850
      Not stated × 1
       Ethnicity–w, D–I, J, T–50
       Religion––q, Na, WMore attendance × 1450
      Religious background and beliefs × 1
       Social functioning––a, Nn, THigher functioning × 1450
      Lower functioning × 1
       Living circumstances–rTd, q, rNot stated × 1425
      Substance-related measures
       Baseline alcohol consumptionk, p, r, Nc, o, Gd, k, r, A, B, E, N, U, V, W, Xa, n, o, p, s, u, x, y, z, G, KLower consumption × 122352
       Dependence severityi, k, If, ta, d, h, v, C, N, Q, T, Ub, k, m, n, s, t, u, A, B, I, J, RLower severity × 102352
      Greater severity × 1
      Not stated × 1
       Treatment historyc, fg, k, o, y, Ic, o, I, J, N, Wa, d, h, k, q, x, yFewer treatments × 41547
      More treatments × 1
      Completed past treatments × 1
      Less AA participation × 1
      Greater AA participation × 2
       Other substance usercr, J, Q, W, Xd, h, w, I, T, XLess substance use × 41242
      More substance use × 1
       Alcohol-related self-efficacyp, F, G, R, V–p, A, C, E, F, G, N, RVHigher self-efficacy × 89100
      Lower self-efficacy × 1 (for females only)
       Motivations, A KDs, u, K, N, W, YA, EHigher motivation × 6977
      Motivation × personality disorder × 1
       Duration of alcohol misuseg, rc, o, xr, Jo, xShorter duration × 1650
      Non-linear × 1
      Not stated × 1
       Onset of alcohol misusekc, g, p, Sk, pLater onset × 1617
       Alcohol-related problems–––a, b, m, s, A, U–60
       Treatment goalGkd, k, G, UsAbstinence goal × drank during580
      Treatment × 1
      Abstinent goal × 3
       Craving/impaired controlBctA, B, PLower craving/less impaired control × 2540
       Family history of alcohol/drug problems–c, g, pno, pNo family history × 1520
       Alcohol expectanciesy, K–x, y, z, K–Higher negative alcohol expectancies × 44100
      Lower positive alcohol expectancies × 1
      Higher sexual enhancement expectancy × 1
      Other clinical meaures
       Psychopathology ratingc, o, r, G, I, Qw, D, Ho, G, N, W, Xa, c, h, JLower rating × 81560
      Reduced rating over treatment × 1
       Depressiono, pb, c, H, Sn, o, Qh, j, p, u, v, w, T, XAbsence/lower depression × 21527
      Depression (for females only) × 2
       ASPD/Criminalityf, wc, r, SI, Nh, J, Q, T, XLower ASPD/criminality × 41233
       Neuropsychological functioningoE, w, Da, b, i, l, qo, N, TBetter functioning × 51155
      Worse functioning × 1
       Anxietyo, wC, h, Hj, Xn, o, u, P, TLess anxiety × 31136
      Greater anxiety × 1
       Physical health–Dh, In, J, WPoorer health × 1633
      Better health × 1
       Personalityf, H, S–f, m, L, S–Greater persistence × 25100
      Greater extroversion × 1
      Lower novelty seeking × 1
      Lower sensation seeking × 1
       Other personality disorder or PD global ratingS, YC, e, rYSNo personality disorder × 1540
      Higher obsessionality score × 1
      Letters represent the following studies: a,
      • Abbott M.W.
      • Gregson R.A.M.
      Cognitive dysfunction in the prediction of relapse in alcoholics.
      Journal of Studies on Alcohol. 1981; 42: 230-243
      ; b,
      • Allsop S.
      • Saunders B.
      • Phillips M.
      The process of relapse in severely dependent male problem drinkers.
      Addiction. 2000; 95: 95-106
      ; c,
      • Bottlender M.
      • Soyka M.
      Outpatient alcoholism treatment: Predictors of outcome after 3 years.
      Drug and Alcohol Dependence. 2005; 80: 83-89
      ; d,
      • Breslin F.C.
      • Sobell M.B.
      • Sobell L.C.
      • Buchan G.
      • Cunningham J.A.
      Toward a stepped care approach to treating problem drinkers: The predictive utility of within-treatment variables and therapist prognostic ratings.
      Addiction. 1997; 92: 1479-1489
      ; e,
      • Burtscheidt W.
      • Wolwer W.
      • Schwarz R.
      • Strauss W.
      • Gaebel W.
      Outpatient behaviour therapy in alcoholism: Treatment outcome after 2 years.
      Acta Psychiatrica Scandinavica. 2002; 106: 227-232
      ; f,
      • Cannon D.S.
      • Keefe C.K.
      • Clark L.A.
      Persistence predicts latency to relapse following inpatient treatment for alcohol dependence.
      Addictive Behaviors. 1997; 22: 535-543
      ; g,
      • Canton G.
      • Giannini L.
      • Magni G.
      • Bertinaria A.
      • Cibin M.
      • Gallimberti L.
      Locus of control, life events and treatment outcome in alcohol dependent patients.
      Acta Psychiatrica Scandinavica. 1988; 78: 18-23
      ; h,
      • Curran G.M.
      • Booth B.M.
      Longitudinal changes in predictor profiles of abstinence from alcohol use among male veterans.
      Alcoholism, Clinical and Experimental Research. 1999; 23: 141-143
      ; i,
      • Donovan D.M.
      • Kivlahan D.R.
      • Walker R.D.
      Clinical limitations of neuropsychological testing in predicting treatment outcome among alcoholics.
      Alcoholism, Clinical and Experimental Research. 1984; 8: 470-475
      ,
      • Kivlahan D.R.
      • Sher K.J.
      • Donovan D.M.
      The Alcohol Dependence Scale: A validation study among inpatient alcoholics.
      Journal of Studies on Alcohol. 1989; 50: 170-175
      ; j,
      • Driessen M.
      • Meier S.
      • Hill A.
      • Wetterling T.
      • Lange W.
      • Junghanns K.
      The course of anxiety, depression and drinking behaviours after completed detoxification in alcoholics with and without comorbid anxiety and depressive disorders.
      Alcohol and Alcoholism. 2001; 36: 249-255
      ; k,
      • Duckert F.
      Predictive factors for outcome of treatment for alcohol problems.
      Journal of Substance Abuse. 1993; 5: 31-44
      ; l,
      • Eckardt M.J.
      • Rawlings R.R.
      • Graubard B.I.
      • Faden V.
      • Martin P.R.
      • Gottschalk L.A.
      Neuropsychological performance and treatment outcome in male alcoholics.
      Alcoholism, Clinical and Experimental Research. 1988; 12: 88-93
      ; m,
      • Edwards G.
      • Taylor C.
      A test of the matching hypothesis: Alcohol dependence, intensity of treatment, and 12-month outcome.
      Addiction. 1994; 89: 553-561
      ; n,
      • Ellis D.
      • McClure J.
      In-patient treatment of alcohol problems—Predicting and preventing relapse.
      Alcohol & Alcoholism. 1992; 27: 449-456
      ; o,
      • Glenn S.W.
      • Parsons O.A.
      Prediction of resumption of drinking in posttreatment alcoholics.
      International Journal of the Addictions. 1991; 26: 237-254
      ,
      • Parsons O.A.
      • Schaeffer K.W.
      • Glenn S.W.
      Does neuropsychological test performance predict resumption of drinking in posttreatment alcoholics?.
      Addictive Behaviors. 1990; 15: 297-307
      ; p,
      • Greenfield S.F.
      • Hufford M.R.
      • Vagge L.M.
      • Muenz L.R.
      • Costello M.E.
      • Weiss R.D.
      The relationship of self-efficacy expectancies to relapse among alcohol dependent men and women: A prospective study.
      Journal of Studies on Alcohol. 2000; 61: 345-351
      ,
      • Greenfield S.F.
      • Kolodziej M.E.
      • Sugarman D.E.
      • Muenz L.R.
      • Vagge L.M.
      • He D.Y.
      • et al.
      History of abuse and drinking outcomes following inpatient alcohol treatment: A prospective study.
      Drug and Alcohol Dependence. 2002; 67: 227-234
      ,
      • Greenfield S.F.
      • Sugarman D.E.
      • Muenz L.R.
      • Patterson M.D.
      • He D.Y.
      • Weiss R.D.
      The relationship between educational attainment and relapse among alcohol-dependent men and women: A prospective study.
      Alcoholism, Clinical and Experimental Research. 2003; 27: 1278-1285
      ; q,
      • Gregson R.A.
      • Taylor G.M.
      Prediction of relapse in men alcoholics.
      Journal of Studies on Alcohol. 1977; 38: 1749-1760
      ; r,
      • Haver B.
      Comorbid psychiatric disorders predict and influence treatment outcomes in female alcoholics.
      European Addiction Research. 2003; 9: 39-44
      ,
      • Haver B.
      • Dahlgren L.
      • Willander A.
      A 2-year follow-up of 120 Swedish female alcoholics treated early in their drinking career: Prediction of drinking outcome.
      Alcoholism, Clinical and Experimental Research. 2001; 25: 1586-1593
      ; s,
      • Heather N.
      • Rollnick S.
      • Bell A.
      Predictive validity of the Readiness to Change Questionnaire.
      Addiction. 1993; 88: 1667-1677
      ; t,
      • Heather N.
      • Brodie J.
      • Wale S.
      • Wilkinson G.
      • Luce A.
      • Webb E.
      • et al.
      A randomized controlled trial of moderation-oriented cue exposure.
      Journal of Studies on Alcohol. 2000; 61: 561-570
      ,
      • Heather N.
      • Dawe S.
      Level of impaired control predicts outcome of moderation-oriented treatment for alcohol problems.
      Addiction. 2005; 100: 945-952
      ; u,
      • Hernandez-Avila C.A.
      • Burleson J.A.
      • Kranzler H.R.
      Stage of change as a predictor of abstinence among alcohol-dependent subjects in pharmacotherapy trials.
      Substance Abuse. 1998; 19: 81-91
      ; v,
      • Hodgins D.C.
      • el-Guebaly N.
      • Armstrong S.
      • Dufour M.
      Implications of depression on outcome from alcohol dependence: A 3-year prospective follow-up.
      Alcoholism, Clinical and Experimental Research. 1999; 23: 151-157
      ; w,
      • Hunter E.E.
      • Powell B.J.
      • Penick E.C.
      • Nickel E.J.
      • Liskow B.I.
      • Cantrell P.J.
      • et al.
      Comorbid psychiatric diagnosis and long-term drinking outcome.
      Comprehensive Psychiatry. 2000; 41: 334-338
      ,
      • Powell B.J.
      • Penick E.C.
      • Nickel E.J.
      • Liskow B.I.
      • Riesenmy K.D.
      • Campion S.L.
      • et al.
      Outcomes of co-morbid alcoholic men: A 1-year follow-up.
      Alcoholism, Clinical and Experimental Research. 1992; 16: 131-138
      ; x,
      • Jones B.T.
      • McMahon J.
      Negative and positive alcohol expectancies as predictors of abstinence after discharge from a residential treatment program: A one-month and three-month follow-up study in men.
      Journal of Studies on Alcohol. 1994; 55: 543-548
      ; y,
      • Jones B.T.
      • McMahon J.
      Negative alcohol expectancy predicts post-treatment abstinence survivorship: The whether, when and why of relapse to a first drink.
      Addiction. 1994; 89: 1653-1665
      ; z,
      • Jones B.T.
      • McMahon J.
      Changes in alcohol expectancies during treatment relate to subsequent abstinence survivorship.
      British Journal of Clinical Psychology. 1996; 35: 221-234
      ; A,
      • Kavanagh D.J.
      • Sitharthan T.
      • Sayer G.P.
      Prediction of results from correspondence treatment for controlled drinking.
      Addiction. 1996; 91: 1539-1545
      ; B,
      • Kranzler H.R.
      • Mulgrew C.
      • Modesto-Lowe V.
      • Burleson J.A.
      Validity of the obsessive compulsive drinking scale (OCDS): Does craving predict drinking behavior?.
      Alcoholism, Clinical and Experimental Research. 1999; 23: 108-114
      ; C,
      • Langenbucher J.
      • Sulesund D.
      • Chung T.
      • Morgenstern J.
      Illness severity and self-efficacy as course predictors of DSM-IV alcohol dependence in a multisite clinical sample.
      Addictive Behaviors. 1996; 21: 543-553
      ,
      • Langenbucher J.
      • Chung T.
      • Morgenstern J.
      • Labouvie E.
      • Nathan P.E.
      • Bavly L.
      Physiological alcohol dependence as a “specifier” of risk for medical problems and relapse liability in DSM-IV.
      Journal of Studies on Alcohol. 1997; 58: 341-350
      ; D,
      • Lemke S.
      • Moos R.H.
      Outcomes at 1 and 5 years for older patients with alcohol use disorders.
      Journal of Substance Abuse Treatment. 2003; 24: 43-50
      ; E,
      • Litt M.D.
      • Kadden R.M.
      • Cooney N.L.
      • Kabela E.
      Coping skills and treatment outcomes in cognitive–behavioral and interactional group therapy for alcoholism.
      Journal of Consulting and Clinical Psychology. 2003; 71: 118-128
      ; F,
      • Long C.G.
      • Hollin C.R.
      • Williams M.J.
      Self-efficacy, outcome expectations, and fantasies as predictors of alcoholics' posttreatment drinking.
      Substance Use and Misuse. 1998; 33: 2383-2402
      ; G,
      • Long C.G.
      • Williams M.
      • Hollin C.R.
      Alcoholism treatment: Intake variables as outcome predictors.
      Addiction Research. 1998; 6: 295-305
      ,
      • Long C.G.
      • Williams M.
      • Midgley M.
      • Hollin C.R.
      Within-program factors as predictors of drinking outcome following cognitive–behavioral treatment.
      Addictive Behaviors. 2000; 25: 573-578
      ; H,
      • Marra D.
      • Warot D.
      • Payan C.
      • Hispard E.
      • Dally S.
      • Puech A.J.
      Anhedonia and relapse in alcoholism.
      Psychiatry Research. 1998; 80: 187-196
      ; I,
      • McLellan A.T.
      • Luborsky L.
      • Woody G.E.
      • O'Brien C.P.
      • Druley K.A.
      Predicting response to alcohol and drug abuse treatments. Role of psychiatric severity.
      Archives of General Psychiatry. 1983; 40: 620-625
      ; J,
      • McLellan A.T.
      • Alterman A.I.
      • Metzger D.S.
      • Grissom G.R.
      • Woody G.E.
      • Luborsky L.
      • et al.
      Similarity of outcome predictors across opiate, cocaine, and alcohol treatments: Role of treatment services.
      Journal of Consulting and Clinical Psychology. 1994; 62: 1141-1158
      ; K,
      • McMahon J.
      • Jones B.T.
      Post-treatment abstinence survivorship and motivation for recovery: The predictive validity of the readiness to change (RCQ) and negative alcohol expectancy (NAEQ) questionnaire.
      Addiction Research. 1996; 4: 161-176
      ; L,
      • Meszaros K.
      • Lenzinger E.
      • Hornik K.
      • Fureder T.
      • Willinger U.
      • Fischer G.
      • et al.
      The Tridimensional Personality Questionnaire as a predictor of relapse in detoxified alcohol dependents. The European Fluvoxamine in Alcoholism Study Group.
      Alcoholism, Clinical and Experimental Research. 1999; 23: 483-486
      ; M,
      • Noda T.
      • Imamichi H.
      • Kawata A.
      • Hirano K.
      • Ando T.
      • Iwata Y.
      • et al.
      Long-term outcome in 306 males with alcoholism.
      Psychiatry and Clinical Neurosciences. 2001; 55: 579-586
      ; N,
      Project MATCH Research Group
      Matching alcoholism treatments to client heterogeneity: Project MATCH three-year drinking outcomes.
      Alcoholism, Clinical and Experimental Research. 1998; 22: 1300-1311
      ,
      • Zywiak W.H.
      • Longabaugh R.
      • Wirtz P.W.
      Decomposing the relationship between pretreatment social network characteristics and alcohol treatment outcome.
      Journal of Studies on Alcohol. 2002; 63: 114-121
      ; O,
      • Rice C.
      • Longabaugh R.
      • Beattie M.
      • Noel N.
      Age group differences in response to treatment for problematic alcohol use.
      Addiction. 1993; 88: 1369-1375
      ; P,
      • Rohsenow D.J.
      • Monti P.M.
      • Rubonis A.V.
      • Sirota A.D.
      • Niaura R.S.
      • Colby S.M.
      • et al.
      Cue reactivity as a predictor of drinking among male alcoholics.
      Journal of Consulting and Clinical Psychology. 1994; 62: 620-626
      ; Q,
      • Rounsaville B.J.
      • Dolinsky Z.S.
      • Babor T.F.
      • Meyer R.E.
      Psychopathology as a predictor of treatment outcome in alcoholics.
      Archives of General Psychiatry. 1987; 44: 505-513
      ; R,
      • Rychtarik R.G.
      • Prue D.M.
      • Rapp S.R.
      • King A.C.
      Self-efficacy, aftercare and relapse in a treatment program for alcoholics.
      Journal of Studies on Alcohol. 1992; 53: 435-440
      ; S,
      • Sellman J.D.
      • Joyce P.R.
      Does depression predict relapse in the 6 months following treatment for men with alcohol dependence?.
      Australian and New Zealand Journal of Psychiatry. 1996; 30: 573-578
      ,
      • Sellman J.D.
      • Mulder R.T.
      • Sullivan P.F.
      • Joyce P.R.
      Low persistence predicts relapse in alcohol dependence following treatment.
      Journal of Studies on Alcohol. 1997; 58: 257-263
      ; T,
      • Shaw G.K.
      • Waller S.
      • Latham C.J.
      • Dunn G.
      • Thomson A.D.
      The detoxification experience of alcoholic in-patients and predictors of outcome.
      Alcohol and Alcoholism. 1998; 33: 291-303
      ; U,
      • Sobell M.B.
      • Sobell L.C.
      • Gavin D.R.
      Portraying alcohol treatment outcomes: Different yardsticks of success.
      Behavior Therapy. 1995; 26: 643-669
      ; V,
      • Solomon K.E.
      • Annis H.M.
      Outcome and efficacy expectancy in the prediction of post-treatment drinking behaviour.
      British Journal of Addiction. 1990; 85: 659-665
      ; W,
      • Staines G.
      • Magura S.
      • Rosenblum A.
      • Fong C.
      • Kosanke N.
      • Foote J.
      • et al.
      Predictors of drinking outcomes among alcoholics.
      American Journal of Drug and Alcohol Abuse. 2003; 29: 203-218
      ; X,
      • Tomasson K.
      • Vaglum P.
      Psychopathology and alcohol consumption among treatment-seeking alcoholics: A prospective study.
      Addiction. 1996; 91: 1019-1030
      ; Y,
      • Verheul R.
      • Van Den Brink W.
      • Hartgers C.
      Personality disorders predict relapse in alcoholic patients.
      Addictive Behaviors. 1998; 23: 869-882
      .
      Demographic variables were for the most part widely investigated, with age being the sole variable in Table 1 to be utilized as a predictor variable in more than half of the studies reviewed. Gender was a significant predictor for half of the reviewed studies, as was religion, albeit investigated in only four studies, whereas age, marital status, education, and ethnicity were generally poor predictors. Social functioning was captured by a variety of measures, with employment and higher socioeconomic status predicting better outcome. Global measures of social functioning produced inconsistent results in two studies reporting significant results, and the variable living circumstances was also an inconsistent predictor in this category.
      Substance-related measures included several consistent predictors. Those with the highest rate of successful prediction were alcohol-related self-efficacy, motivation to change, treatment goal, and alcohol-related expectancies, ranging from 77% (motivation) to 100% (self-efficacy and expectancies). Motivation was measured using a wide range of instruments, whereas self-efficacy and alcohol expectancies were more consistent in using versions of the Situational Confidence Questionnaire (
      • Annis H.M.
      • Graham J.M.
      Situational Confidence Questionnaire (SCQ 39): User's guide.
      Addiction Research Foundation, Toronto1988
      ) and the Alcohol Expectancies Questionnaire (
      • Brown S.A.
      • Goldman M.S.
      • Inn A.
      • Anderson L.R.
      Expectations of reinforcement from alcohol: Their domain and relation to drinking patterns.
      Journal of Consulting and Clinical Psychology. 1980; 48: 419-426
      ) or Negative Alcohol Expectancies Questionnaire (
      • McMahon J.
      • Jones B.T.
      The Negative Alcohol Expectancy Questionnaire.
      Journal of the Association of Nurses in Substance Abuse. 1993; 12: 17
      ). The utility of treatment goal as a predictor was achieved using a nonabstinent outcome measure in all cases. Moderate success was identified for baseline alcohol consumption, dependence severity, treatment history, and duration of alcohol misuse and somewhat lower rates for other substance use and craving/impaired control. Onset age of alcohol misuse, alcohol-related problems, and family history of alcohol or drug problems were all poor predictors of outcome.
      Baseline alcohol consumption and dependence severity were two of the three most widely investigated predictor variables. Both categories consisted of a range of measures. For baseline alcohol consumption, the most commonly used were percentage days abstinent (PDA), drinks per drinking day (DDD), and total consumption. PDA and DDD were somewhat more consistently predictive while all three were more consistent than the miscellaneous and poorly defined other measures of drinking behavior and consumption, which predicted outcome in only three of nine studies. Four studies utilized both PDA and DDD as predictors but did not show a clear advantage for either. For dependence severity, the most commonly used measures were the Severity of Alcohol Dependence Questionnaire/Severity of Alcohol Dependence Questionnaire—Form C (SADQ/SADQ-C;
      • Stockwell T.
      • Murphy D.
      • Hodgson R.
      The severity of alcohol dependence questionnaire: Its use, reliability and validity.
      British Journal of Addiction. 1983; 78: 145-155
      ), Alcohol Dependence Scale (ADS;
      • Skinner H.A.
      • Allen B.A.
      Alcohol dependence syndrome: Measurement and validation.
      Journal of Abnormal Psychology. 1982; 91: 199-209
      ), and Addiction Severity Index (ASI;
      • McLellan A.T.
      • Luborsky L.
      • Woody G.E.
      • O'Brien C.P.
      An improved diagnostic evaluation instrument for substance abuse patients. The Addiction Severity Index.
      Journal of Nervous and Mental Disease. 1980; 168: 26-33
      ). The ADS and ASI each predicted outcomes in approximately two thirds of studies (eight and six studies, respectively). An additional six miscellaneous measures, employed in a total of eight studies, were equally consistent predictors. In contrast, the SADQ/SADQ-C was found to be a significant predictor for only one of seven studies.
      • Langenbucher J.
      • Sulesund D.
      • Chung T.
      • Morgenstern J.
      Illness severity and self-efficacy as course predictors of DSM-IV alcohol dependence in a multisite clinical sample.
      Addictive Behaviors. 1996; 21: 543-553
      compared the ADS, ASI, and DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition) alcohol dependence symptom count in predicting outcome at 6 months, using six measures of relapse, consumption pattern, and dependence severity. The ADS and ASI successfully predicted four and two measures, respectively, and the DSM-IV symptom count successfully predicted one measure.
      More general psychopathological ratings were found to be approximately twice as likely as were variables looking specifically at depression or anxiety to predict outcome and were more consistent in finding that less psychopathology predicted better outcome. Rating scales were varied, with the most common being the psychological severity scale of the ASI. Neither antisocial personality disorder (ASPD), the most commonly measured coexisting personality disorder, nor other personality disorders were consistent predictors of outcome.
      Neuropsychological functioning was a moderately consistent predictor while personality was a significant predictor for all five studies in which it was examined. For both neuropsychological functioning and personality, however, these results often reflected studies in which a large number of variables in each category were examined with only a small subset actually being identified as being significant. Table 1 shows the range of personality measures identified in five studies while the neuropsychiatric variables were too varied and numerous to be meaningfully summarized, with inconsistent results across studies for the often small number of studies examining specific neuropsychological domains. The other remaining clinical variable, physical health, was a generally inconsistent predictor.
      A number of variables produced contradictory findings, although, perhaps with the exception of depression, none showed a clear pattern of divergent findings, with typically only one finding at odds with the remainder of significant findings for that variable.

      3.1 Identifying key predictors

      The most consistent predictor would be one that had been tested in a large number of studies and was found to be significant, in a consistent direction, in both univariate and multivariate analyses. Furthermore, consistent predictors would be best identified when they were included in multivariate models with other competing consistent predictors.
      A subset of the variables that predicted outcome in 50% or more of studies was singled out to identify the most consistent predictors from the candidates presented in Table 1. Employment and treatment history were added to this list as both were successful predictors in 47% of studies, were widely examined, and constitute variables routinely collected in clinical settings, whereas social functioning was excluded as the significant findings were contradictory and based on only four studies (see Table 1). Next, only studies reporting multivariate analysis employing a minimum of four of these variables were selected. Variables then needed to be reported in four or more studies. The results are displayed in Table 2 for the 12 variables meeting these criteria, drawn from 19 studies.
      Table 2Key predictors of alcohol-consumption-related treatment outcome in multivariate analysis
      VariablePredictorNonpredictorTotal studiesPercentage significant
      Demographic and social functioning measures
       Employmenta, h, n, J, Ud, p, q, u, x, G, T1242
       Genderz, K, N, Ua, d, o, p, u, A, J, T1233
       Socioeconomic status/Incomen, qd, T, W540
       Religionq, Na, W450
      Substance-related measures
       Baseline alcohol consumptiond, A, N, U, Wa, n, o, p, s, u, x, z, G, K1533
       Dependence severitya, d, h, N, T, Un, s, u, A, J1155
       Treatment historyo, J, N, Wa, d, h, q, x944
       Alcohol-related self-efficacyp, A, G, N–4100
       Motivations, u, K, N, WA683
       Treatment goald, G, Us475
      Other clinical measures
       Psychopathology ratingo, G, N, Wa, h, J757
       Neuropsychological functioninga, qo, N, T540
      Letters represent studies as shown in the legend to Table 1.
      When subjected to the stricter test embodied by Table 2, baseline alcohol consumption and gender showed significant reductions in predictive consistency whereas the remaining variables were not greatly affected. The most consistent predictors remain dependence severity, psychopathology ratings, alcohol-related self-efficacy, motivation, and treatment goal. Treatment history showed significant variability in the direction of association in Table 1, with better outcome predicted by less treatment in five studies, by more treatment in three studies, and by completing treatment in one study. In contrast, when studies were restricted to the more rigorous group in Table 2, three of the four studies found better outcome to be predicted by more treatment, with one predicting the reverse.

      3.2 Predicting prediction: Associations with total variance accounted for

      While most of the studies reviewed here examined prediction using multivariate analysis, not all reported on the combined predictive power of the models developed. For those that do, the total percentage of variance accounted for varies widely from a nonsignificant R2 = .03 in predicting drinks per day (
      • Donovan D.M.
      • Kivlahan D.R.
      • Walker R.D.
      Clinical limitations of neuropsychological testing in predicting treatment outcome among alcoholics.
      Alcoholism, Clinical and Experimental Research. 1984; 8: 470-475
      ) to R2 = .62 in predicting DDD (
      • Sobell M.B.
      • Sobell L.C.
      • Gavin D.R.
      Portraying alcohol treatment outcomes: Different yardsticks of success.
      Behavior Therapy. 1995; 26: 643-669
      ), with a mean R2 = .30.
      All R2 values for drinking-related outcomes reported in the reviewed papers were entered into SPSS version 13.0. In total, 21 studies with 41 R2 values were used, and the associations between these and a number of methodological factors were explored. The methodological factors can be divided into those relating to the individual predictor variables employed, which is the focus of this review, and a range of wider factors, such as sample size, selection, and composition. This second set of factors is examined first, with any significant association with resultant R2 values used to allow for a more conservative exploration of which key predictor variables are associated with higher multivariate R2 values.
      In univariate analysis, higher R2 values were predicted by mixed gender samples (t = 3.00, p = .005), samples not limited to those meeting criteria for alcohol dependence (t = 3.93, p < .001), and outpatient or mixed samples compared to exclusively inpatient samples (t = 3.42, p = .001). A trend was found for including variables measured after baseline (t = 1.69, p = .099). Variables not predictive of higher R2 values were limiting the sample to treatment completers, publication date, sample size, follow-up interval, and follow-up rate. Entering the four variables significant to p < .10 into a conditional stepwise regression produced a model accounting for 43.6% of variance in R2 values. This model indicated that samples not limited to those meeting criteria for alcohol dependence (standardized β = .485, t = 3.63, p = .001), including variables measured after baseline (standardized β = .388, t = 3.21, p = .003), and mixed gender samples (standardized β = .274, t = 2.07, p = .045) were all independently associated with more predictive models. Identical results were generated from both forward and backward conditional models.
      When key predictors were examined, higher R2 values were predicted in univariate analysis by studies using baseline alcohol consumption (t = 5.38, p < .001), dependence severity (t = 2.19, p = .034), and treatment goal (t = 3.09, p = .004) and those not using neuropsychological functioning variables (t = 3.20, p = .003), while there was a trend for those studies not using psychopathology ratings (t = 1.88, p = .069). Not surprisingly, it was found that the number of key predictor variables used in a study was positively correlated with total variance accounted for (r = .348, p = .026). These six variables and the four methodological variables identified as significant/trends in univariate analysis were entered into stepwise conditional regression models. In the backward conditional model, R2=.533, from including variables measured after baseline (standardized β = .340, t = 2.81, p = .008), baseline alcohol consumption (standardized β = .555, t = 5.01, p < .001), not using a psychopathology rating (standardized β = −.299, t = −2.46, p = .019), and dependence severity (standardized β = .236, t = 2.15, p = .038). In the forward conditional model, R2=.397, from baseline alcohol consumption alone (standardized β = .642, t = 5.23, p < .001).
      The negative association between use of psychopathology ratings and variance accounted for was examined, appearing to confirm the suspicion that studies employing such measures may recruit samples with greater psychiatric severity, while those excluding patients with significant psychiatric comorbidity would be less inclined to employ such measures: Of the 6 studies using these measures, only 1 excluded patients based on psychiatric status compared to 7 of the 15 studies not employing such measures.
      The two original methodological variables removed from both forward and backward regression were mixed gender samples and samples not limited to those meeting criteria for alcohol dependence. Post hoc exploration revealed that both were highly correlated with the use of baseline alcohol consumption measures. None of the 13 studies employing all male samples measured baseline alcohol consumption compared to 22 of the remaining 28 (χ2 = 22.0, p < .001), while baseline alcohol consumption measures were also rare for studies limited to dependent populations (7 of 25) compared to more inclusive studies (15 of 16; χ2 = 17.0, p < .001). In contrast, the use of baseline alcohol consumption measures was not correlated with timing of predictor measurement (χ2 = 0.15, ns).
      Whereas choice of outcome measure was too diverse to be examined for the small sample of studies providing R2 values, the frequency with which different outcome measures were associated with the various predictor values (i.e., the mirror image if the primary question for this review) was examined and showed that continuous consumption measures (DDD, PDA, and combined consumption measures) were more often predicted by baseline variables than were categorical measures (usually abstinence status) or time to lapse/relapse measures.

      4. Discussion

      A wide range of baseline patient characteristics were examined as potential predictors of treatment outcome. The most consistent predictors were identified, and several methodological factors influencing prediction were also examined.
      Dependence severity, consumption level, and treatment history are perhaps the most intuitive predictors of treatment outcome, and coexisting psychiatric conditions and motivation might perhaps be added to these. The current review has reinforced the importance of these variables, all of which were previously found to be significant (
      • Gibbs L.
      • Flanagan J.
      Prognostic indicators of alcoholism treatment outcome.
      The International Journal of the Addiction. 1977; 12: 1097-1141
      ), while adding a degree of refinement to understanding their relationship to outcome: Baseline alcohol consumption was less robust when regressed along with other key predictors; the association between baseline dependence severity and outcome was dependent on the measures used; and the two main classes of coexisting disorders, depression and anxiety, were less consistent predictors of outcome than a continuous measure of current psychopathology.
      In addition to validating these measures, several other variables emerged as potentially valuable for future research and clinical application, in particular alcohol-related self-efficacy, alcohol expectancies, and treatment goal, although the latter two of these were investigated by only a small number of studies. Self-efficacy emerges as the most consistent predictor variable, with all nine studies investigating this variable reporting a significant association, albeit with one reverse association (as discussed below).
      Alcohol consumption can be measured in a variety of ways, as demonstrated by the studies reviewed here, with frequency, intensity, and combined measures found to be the most consistent predictors and, in turn, the most consistently predicted outcome measures. In considering which measures would be most usefully employed to predict outcome, it should be acknowledged that measuring baseline alcohol consumption in a way that matches consumption measures at outcome has a natural logic to it. Babor et al. (
      • Babor T.F.
      • Longabaugh R.
      • Zweben A.
      • Fuller R.K.
      • Stout R.L.
      • Anton R.F.
      • et al.
      Issues in the definition and measurement of drinking outcomes in alcoholism treatment research.
      Journal of Studies on Alcohol—Supplement. 1994; 12: 101-111
      ), in considering the definition and measurement of drinking-related outcome, recommended drinking frequency and intensity as primary outcome measures. They argue that drinking intensity and frequency have been demonstrated to be somewhat independent measures, an assertion supported by analysis of four of their own studies. Furthermore, negative consequences are more likely with high levels of either, with some consequences more related to one than the other. Finally, they suggest that different treatments may impact to varying degrees on these two, either for the treatment population as a whole or for different subsets.
      In contrast to consumption and dependence severity being successful predictors, the variable alcohol-related problems, which might be expected to complement these, was not significant in any of the six studies in which it was investigated. This can be contrasted further with the positive results shown for negative alcohol expectancies. An awareness of the risks associated with heavy alcohol use is more predictive of treatment outcome than is the extent to which such adverse events are actually experienced.
      Marital status is surprisingly unpredictive, with contradictory findings among the small number of positive findings. This is in contrast to other measures that might be considered to reflect social stability, such as employment, socioeconomic status, and social functioning. The unreliability of marital status as a predictor in the current review is in contrast to the findings of
      • Gibbs L.
      • Flanagan J.
      Prognostic indicators of alcoholism treatment outcome.
      The International Journal of the Addiction. 1977; 12: 1097-1141
      where it was significant in 12 of 23 studies, all finding being married associated with better outcome than not being married in univariate analysis.
      • Beattie M.C.
      Meta-analysis of social relationships and posttreatment drinking outcomes: Comparison of relationship structure, function and quality.
      Journal of Studies on Alcohol. 2001; 62: 518-527
      found being married to be a weaker predictor of positive drinking outcome than social support or marital and family adjustment, and hence, it may be that this most widely examined relationship variable is in fact the least likely to predict outcome. This meta-analysis also found that the association was substantially lower for treatment populations that were not predominantly male. The contrasting finding in the current review with that of Gibbs and Flanagan may therefore be the result of the increasing number of women attending alcohol treatment, represented in published research, in more recent decades.
      A notable exception to the high rate of demographic variables utilized in prediction studies is the very small number of studies reporting on ethnicity as a predictor variable. Whether this is due to reticence to address what any ethnicity-based outcome disparity might mean or a reflection on the poor state of data collection for this variable is unclear. However, given that patterns of substance misuse are known to vary across different ethnic groups (
      • Castro F.G.
      • Proescholdbell R.J.
      • Abeita L.
      • Rodriguez D.
      Ethnic and cultural minority groups.
      in: McCrady B.S. Epstein E.E. Addictions: A comprehensive guidebook. Oxford University Press, New York1999
      ) and the increasing recognition that effective treatment for such groups is potentially undermined in settings that do not respond to such patients' cultural needs (
      • Brady M.
      Culture in treatment, culture as treatment. A critical appraisal of developments in addictions programs for indigenous North Americans and Australians.
      Social Science & Medicine. 1995; 41: 1487-1498
      ,
      • Huriwai T.
      • Sellman J.D.
      • Sullivan P.
      • Potiki T.L.
      Optimal treatment for Maori with alcohol and drug-use-related problems: An investigation of cultural factors in treatment.
      Substance Use and Misuse. 2000; 35: 281-300
      ,
      • Terrell M.D.
      Ethnocultural factors and substance abuse: Towards culturally sensitive treatment models.
      Psychology of Addictive Behaviors. 1993; 7: 162-167
      ), it is disappointing that the association between ethnicity and treatment outcome is not being more routinely reported.
      Some variables represented in Table 1 are perhaps better considered as variable categories and may constitute a wide array of more specific factors. This is particularly the case for personality and neuropsychological functioning. In such cases, caution is necessary in accepting these as useful predictors unless there is some consistency in which specific measures are significant, with such consistency not found for either of these variable categories.
      A number of variables produced contradictory findings. The presence of these divergent results highlights the desire for more sophisticated handling of study data. It may be for example that some variables are moderated by others, such as two studies (
      • Ellis D.
      • McClure J.
      In-patient treatment of alcohol problems—Predicting and preventing relapse.
      Alcohol & Alcoholism. 1992; 27: 449-456
      ,
      • Rounsaville B.J.
      • Dolinsky Z.S.
      • Babor T.F.
      • Meyer R.E.
      Psychopathology as a predictor of treatment outcome in alcoholics.
      Archives of General Psychiatry. 1987; 44: 505-513
      ) finding that depression predicted better outcome for women but not for men, while for the two studies finding depression predicted worse outcome,
      • Glenn S.W.
      • Parsons O.A.
      Prediction of resumption of drinking in posttreatment alcoholics.
      International Journal of the Addictions. 1991; 26: 237-254
      found no Gender × Depression interaction and
      • Greenfield S.F.
      • Kolodziej M.E.
      • Sugarman D.E.
      • Muenz L.R.
      • Vagge L.M.
      • He D.Y.
      • et al.
      History of abuse and drinking outcomes following inpatient alcohol treatment: A prospective study.
      Drug and Alcohol Dependence. 2002; 67: 227-234
      ,
      • Greenfield S.F.
      • Sugarman D.E.
      • Muenz L.R.
      • Patterson M.D.
      • He D.Y.
      • Weiss R.D.
      The relationship between educational attainment and relapse among alcohol-dependent men and women: A prospective study.
      Alcoholism, Clinical and Experimental Research. 2003; 27: 1278-1285
      did not examine gender interaction.
      A second anomalous finding was that of
      • Langenbucher J.
      • Sulesund D.
      • Chung T.
      • Morgenstern J.
      Illness severity and self-efficacy as course predictors of DSM-IV alcohol dependence in a multisite clinical sample.
      Addictive Behaviors. 1996; 21: 543-553
      , where higher alcohol-related self-efficacy predicted worse outcome, in contrast to the findings of eight other studies. A possible explanation for this can be found in the work of
      • Burling T.A.
      • Reilly P.M.
      • Moltzen J.O.
      • Ziff D.C.
      Self-efficacy and relapse among inpatient drug and alcohol abusers: A predictor of outcome.
      Journal of Studies on Alcohol. 1989; 50: 354-360
      who found that high self-efficacy predicted early treatment exit and more discharges for “negative” reasons for a large alcohol and drug treatment sample. Baseline self-efficacy did not predict outcome, but improved self-efficacy across treatment did. They proposed that many patients were overconfident on entering treatment, and it was these patients who were more likely to leave early and also to not make appreciable gains in self-efficacy across treatment. Thus, it is possible that a curvilinear relationship exists between self-efficacy and outcome, with some patients vulnerable to having an excess of self-efficacy, presumably inconsistent with their actual ability to resist use and prompting them in some cases to exit treatment early. Unfortunately, the studies reviewed do not provide sufficient data on treatment attendance to tease this out, while comparable scale data from two other studies (
      • Long C.G.
      • Hollin C.R.
      • Williams M.J.
      Self-efficacy, outcome expectations, and fantasies as predictors of alcoholics' posttreatment drinking.
      Substance Use and Misuse. 1998; 33: 2383-2402
      ,
      • Solomon K.E.
      • Annis H.M.
      Outcome and efficacy expectancy in the prediction of post-treatment drinking behaviour.
      British Journal of Addiction. 1990; 85: 659-665
      ) do show that the Langenbucher et al. sample had the highest mean Situational Confidence Questionnaire scores of the three.
      An examination of studies reporting the total variance accounted for by their respective predictive models was undertaken and provided some insight into influences on strength of prediction. A range of methodological variables, beyond which predictor variables were included, were found to be associated with predictor model strength. Given that these were found to covary with the use of different predictor variables, this is an important finding. A larger sample would be needed to draw firm conclusions, but the analyses appear to support the importance of baseline alcohol consumption and dependence severity as highly relevant predictor variables.
      The methodological variables identified in this analysis also highlight the potential importance of patient group and treatment setting. For example, the fact that outcome was better predicted for outpatient and mixed samples than for exclusively inpatient samples is a reminder that the reviewed studies cover a wide range of treatment types. While broad inpatient versus outpatient modality was the only treatment descriptor examined, and only in relation to total model strength, it is possible that specific treatment components may lend themselves to expressing different predictive associations; for example,
      • Beattie M.C.
      Meta-analysis of social relationships and posttreatment drinking outcomes: Comparison of relationship structure, function and quality.
      Journal of Studies on Alcohol. 2001; 62: 518-527
      found that treatments that involved significant others showed greater impact of social support. The same comment could be made for the importance of patient type, such as gender or problem severity, with few studies examining interaction between such characteristics and other predictors. Equally, the “restriction in range” often imposed by treatment setting, on these same variables, has the potential to obscure interactions or direct associations with outcome.
      It is worth considering which potential predictor variables are not present. Coercion or other features of referral, such as self-referral versus agency referral, do not feature among the studies reviewed here, reinforcing the comments of
      • Wild T.C.
      Social control and coercion in addiction treatment: Towards evidence-based policy and practice.
      Addiction. 2006; 101: 40-49
      who highlighted the assumption of the value of coercion in the absence of data. Only one reviewed study examined treatment coercion (
      • McLellan A.T.
      • Alterman A.I.
      • Metzger D.S.
      • Grissom G.R.
      • Woody G.E.
      • Luborsky L.
      • et al.
      Similarity of outcome predictors across opiate, cocaine, and alcohol treatments: Role of treatment services.
      Journal of Consulting and Clinical Psychology. 1994; 62: 1141-1158
      ). Also absent from the clinical variables listed in Table 1, with a single exception (
      • Greenfield S.F.
      • Kolodziej M.E.
      • Sugarman D.E.
      • Muenz L.R.
      • Vagge L.M.
      • He D.Y.
      • et al.
      History of abuse and drinking outcomes following inpatient alcohol treatment: A prospective study.
      Drug and Alcohol Dependence. 2002; 67: 227-234
      ), is a history of sexual abuse, despite high rates (
      • Moncrieff J.
      • Drummond D.C.
      • Candy B.
      Sexual abuse in people with alcohol problems. A study of the prevalence of sexual abuse and its relationship to drinking behaviour.
      British Journal of Psychiatry. 1996; 169: 355-360
      ) and finding that it is a strong predictor for onset of substance use disorder (
      • Monlar B.E.
      • Buka S.L.
      • Kessler R.C.
      Child sexual abuse and subsequent psychopathology: Results from the National Comorbidity Survey.
      American Journal of Public Health. 2001; 91: 753-760
      ).
      While failure to report/publish negative findings is a common concern in clinical research (
      • Easterbrook P.J.
      • Berlin J.A.
      • Gopalan R.
      • Matthews D.R.
      Publication bias in clinical research.
      Lancet. 1991; 337: 867-872
      ), we believe that this is likely to have had, at best, a modest effect on data available for the current review. This is because most reviewed papers contained information on numerous examined associations, most of which were not significant, and thus, it is only likely to have been those studies finding close to no significant predictors of outcome that may have not been published.
      A significant shortcoming for many of the studies reviewed was the failure to control for other important predictor variables, using techniques such as analysis of covariance and regression analysis. Most notably, gender, dependence severity, and baseline alcohol consumption have at times not been included as predictors or covariates. While this is acceptable if the authors only wish to explore univariate predictive power, this does not lend itself to the purpose of this review and, in fact, is often contrary to the stated purpose of the papers in which such oversights are found. Some studies reviewed here that have focused on a single measure do so because they hypothesize a conceptual link between the variable and outcome, either directly or via the treatment process. Examples of this include controlling for baseline alcohol consumption but no other variable in examining the role of efficacy expectancy and outcome expectancy in predicting outcome (
      • Solomon K.E.
      • Annis H.M.
      Outcome and efficacy expectancy in the prediction of post-treatment drinking behaviour.
      British Journal of Addiction. 1990; 85: 659-665
      ), while Greenfield et al. described, with three separate studies, examining the relationship between self-efficacy (
      • Greenfield S.F.
      • Hufford M.R.
      • Vagge L.M.
      • Muenz L.R.
      • Costello M.E.
      • Weiss R.D.
      The relationship of self-efficacy expectancies to relapse among alcohol dependent men and women: A prospective study.
      Journal of Studies on Alcohol. 2000; 61: 345-351
      ), sexual and physical abuse (
      • Greenfield S.F.
      • Kolodziej M.E.
      • Sugarman D.E.
      • Muenz L.R.
      • Vagge L.M.
      • He D.Y.
      • et al.
      History of abuse and drinking outcomes following inpatient alcohol treatment: A prospective study.
      Drug and Alcohol Dependence. 2002; 67: 227-234
      ), and educational attainment (
      • Greenfield S.F.
      • Sugarman D.E.
      • Muenz L.R.
      • Patterson M.D.
      • He D.Y.
      • Weiss R.D.
      The relationship between educational attainment and relapse among alcohol-dependent men and women: A prospective study.
      Alcoholism, Clinical and Experimental Research. 2003; 27: 1278-1285
      ) and treatment outcome in the same sample, without examining these three sets of variables in the same analyses. Both studies make clear causative statements in introducing their papers, but the absence of more rigorous attempts to control for covariates must undermine any conclusions drawn and misses the opportunity to better explore causal relationships
      Attempts to synthesize findings were often made more difficult by the wide range of scales used to measure similar phenomena, different outcome measures, differing time frames, and different populations. The impact of such heterogeneity has been well demonstrated within individual studies reviewed here, including varying predictive strength of different measures of dependence severity (
      • Langenbucher J.
      • Sulesund D.
      • Chung T.
      • Morgenstern J.
      Illness severity and self-efficacy as course predictors of DSM-IV alcohol dependence in a multisite clinical sample.
      Addictive Behaviors. 1996; 21: 543-553
      ), different predictor profiles, using the same measures over the same follow-up period, with different populations (
      Project MATCH Research Group
      Matching alcoholism treatments to client heterogeneity: Project MATCH three-year drinking outcomes.
      Alcoholism, Clinical and Experimental Research. 1998; 22: 1300-1311
      ,
      • Verheul R.
      • Van Den Brink W.
      • Hartgers C.
      Personality disorders predict relapse in alcoholic patients.
      Addictive Behaviors. 1998; 23: 869-882
      ), and changes in the strength of association, depending on the statistical technique employed (
      • Eckardt M.J.
      • Rawlings R.R.
      • Graubard B.I.
      • Faden V.
      • Martin P.R.
      • Gottschalk L.A.
      Neuropsychological performance and treatment outcome in male alcoholics.
      Alcoholism, Clinical and Experimental Research. 1988; 12: 88-93
      ).
      • Curran G.M.
      • Booth B.M.
      Longitudinal changes in predictor profiles of abstinence from alcohol use among male veterans.
      Alcoholism, Clinical and Experimental Research. 1999; 23: 141-143
      found changes in predictor profile across different follow-up periods, whereas the current analysis found that the overall strength of prediction was not a function of follow-up interval. An examination of the impact of methodological variables on total variance accounted for revealed the vulnerability of findings to methodological factors such as the gender mix of the sample. Finally, many of the studies reviewed here used multiple outcome measures, typically finding markedly different patterns of predictor profile.
      The present review aimed to limit the impact of some of the design factors identified above by
      • Gibbs L.
      • Flanagan J.
      Prognostic indicators of alcoholism treatment outcome.
      The International Journal of the Addiction. 1977; 12: 1097-1141
      as hampering elucidation of consistent outcome predictors. Specifically, review eligibility criteria limited the impact of attrition rate and small sample size while an analysis of methodological factors impacting prediction model strength was able to quantify the impact of treatment modality and found that sample size and time to follow-up did not influence total model strength.
      Future research would do well to continue to examine the most likely predictor candidates, as well as to consider those underinvestigated. With a more consistent focus on a smaller number of variables, it may also be possible to start to understand which predictors may be differentially more effective for which outcome measures and to consider interactions between predictor variables and with treatment type.

      References

        • Abbott M.W.
        • Gregson R.A.M.
        Cognitive dysfunction in the prediction of relapse in alcoholics.
        Journal of Studies on Alcohol. 1981; 42: 230-243
        View in Article
        • Allsop S.
        • Saunders B.
        • Phillips M.
        The process of relapse in severely dependent male problem drinkers.
        Addiction. 2000; 95: 95-106
        View in Article
        • Annis H.M.
        • Graham J.M.
        Situational Confidence Questionnaire (SCQ 39): User's guide.
        Addiction Research Foundation, Toronto1988
        View in Article
        • Babor T.F.
        • Longabaugh R.
        • Zweben A.
        • Fuller R.K.
        • Stout R.L.
        • Anton R.F.
        • et al.
        Issues in the definition and measurement of drinking outcomes in alcoholism treatment research.
        Journal of Studies on Alcohol—Supplement. 1994; 12: 101-111
        View in Article
        • Beattie M.C.
        Meta-analysis of social relationships and posttreatment drinking outcomes: Comparison of relationship structure, function and quality.
        Journal of Studies on Alcohol. 2001; 62: 518-527
        View in Article
        • Bottlender M.
        • Soyka M.
        Outpatient alcoholism treatment: Predictors of outcome after 3 years.
        Drug and Alcohol Dependence. 2005; 80: 83-89
        View in Article
        • Brady M.
        Culture in treatment, culture as treatment. A critical appraisal of developments in addictions programs for indigenous North Americans and Australians.
        Social Science & Medicine. 1995; 41: 1487-1498
        View in Article
        • Breslin F.C.
        • Sobell M.B.
        • Sobell L.C.
        • Buchan G.
        • Cunningham J.A.
        Toward a stepped care approach to treating problem drinkers: The predictive utility of within-treatment variables and therapist prognostic ratings.
        Addiction. 1997; 92: 1479-1489
        View in Article
        • Brewer D.D.
        • Catalano R.F.
        • Haggerty K.
        • Gainey R.R.
        • Fleming C.B.
        A meta-analysis of predictors of continued drug use during and after treatment for opiate addiction.
        Addiction. 1998; 93: 73-92
        View in Article
        • Brown S.A.
        • Goldman M.S.
        • Inn A.
        • Anderson L.R.
        Expectations of reinforcement from alcohol: Their domain and relation to drinking patterns.
        Journal of Consulting and Clinical Psychology. 1980; 48: 419-426
        View in Article
        • Burling T.A.
        • Reilly P.M.
        • Moltzen J.O.
        • Ziff D.C.
        Self-efficacy and relapse among inpatient drug and alcohol abusers: A predictor of outcome.
        Journal of Studies on Alcohol. 1989; 50: 354-360
        View in Article
        • Burtscheidt W.
        • Wolwer W.
        • Schwarz R.
        • Strauss W.
        • Gaebel W.
        Outpatient behaviour therapy in alcoholism: Treatment outcome after 2 years.
        Acta Psychiatrica Scandinavica. 2002; 106: 227-232
        View in Article
        • Cannon D.S.
        • Keefe C.K.
        • Clark L.A.
        Persistence predicts latency to relapse following inpatient treatment for alcohol dependence.
        Addictive Behaviors. 1997; 22: 535-543
        View in Article
        • Canton G.
        • Giannini L.
        • Magni G.
        • Bertinaria A.
        • Cibin M.
        • Gallimberti L.
        Locus of control, life events and treatment outcome in alcohol dependent patients.
        Acta Psychiatrica Scandinavica. 1988; 78: 18-23
        View in Article
        • Castro F.G.
        • Proescholdbell R.J.
        • Abeita L.
        • Rodriguez D.
        Ethnic and cultural minority groups.
        in: McCrady B.S. Epstein E.E. Addictions: A comprehensive guidebook. Oxford University Press, New York1999
        View in Article
        • Connors G.J.
        • Maisto S.A.
        • Zywiak W.H.
        Understanding relapse in the broader context of post-treatment functioning.
        Addiction. 1996; 91: S173-S189
        View in Article
        • Curran G.M.
        • Booth B.M.
        Longitudinal changes in predictor profiles of abstinence from alcohol use among male veterans.
        Alcoholism, Clinical and Experimental Research. 1999; 23: 141-143
        View in Article
        • Donovan D.M.
        • Kivlahan D.R.
        • Walker R.D.
        Clinical limitations of neuropsychological testing in predicting treatment outcome among alcoholics.
        Alcoholism, Clinical and Experimental Research. 1984; 8: 470-475
        View in Article
        • Driessen M.
        • Meier S.
        • Hill A.
        • Wetterling T.
        • Lange W.
        • Junghanns K.
        The course of anxiety, depression and drinking behaviours after completed detoxification in alcoholics with and without comorbid anxiety and depressive disorders.
        Alcohol and Alcoholism. 2001; 36: 249-255
        View in Article
        • Duckert F.
        Predictive factors for outcome of treatment for alcohol problems.
        Journal of Substance Abuse. 1993; 5: 31-44
        View in Article
        • Easterbrook P.J.
        • Berlin J.A.
        • Gopalan R.
        • Matthews D.R.
        Publication bias in clinical research.
        Lancet. 1991; 337: 867-872
        View in Article
        • Eckardt M.J.
        • Rawlings R.R.
        • Graubard B.I.
        • Faden V.
        • Martin P.R.
        • Gottschalk L.A.
        Neuropsychological performance and treatment outcome in male alcoholics.
        Alcoholism, Clinical and Experimental Research. 1988; 12: 88-93
        View in Article
        • Edwards G.
        • Taylor C.
        A test of the matching hypothesis: Alcohol dependence, intensity of treatment, and 12-month outcome.
        Addiction. 1994; 89: 553-561
        View in Article
        • Ellis D.
        • McClure J.
        In-patient treatment of alcohol problems—Predicting and preventing relapse.
        Alcohol & Alcoholism. 1992; 27: 449-456
        View in Article
        • Gibbs L.
        • Flanagan J.
        Prognostic indicators of alcoholism treatment outcome.
        The International Journal of the Addiction. 1977; 12: 1097-1141
        View in Article
        • Glenn S.W.
        • Parsons O.A.
        Prediction of resumption of drinking in posttreatment alcoholics.
        International Journal of the Addictions. 1991; 26: 237-254
        View in Article
        • Greenfield S.F.
        • Hufford M.R.
        • Vagge L.M.
        • Muenz L.R.
        • Costello M.E.
        • Weiss R.D.
        The relationship of self-efficacy expectancies to relapse among alcohol dependent men and women: A prospective study.
        Journal of Studies on Alcohol. 2000; 61: 345-351
        View in Article
        • Greenfield S.F.
        • Kolodziej M.E.
        • Sugarman D.E.
        • Muenz L.R.
        • Vagge L.M.
        • He D.Y.
        • et al.
        History of abuse and drinking outcomes following inpatient alcohol treatment: A prospective study.
        Drug and Alcohol Dependence. 2002; 67: 227-234
        View in Article
        • Greenfield S.F.
        • Sugarman D.E.
        • Muenz L.R.
        • Patterson M.D.
        • He D.Y.
        • Weiss R.D.
        The relationship between educational attainment and relapse among alcohol-dependent men and women: A prospective study.
        Alcoholism, Clinical and Experimental Research. 2003; 27: 1278-1285
        View in Article
        • Gregson R.A.
        • Taylor G.M.
        Prediction of relapse in men alcoholics.
        Journal of Studies on Alcohol. 1977; 38: 1749-1760
        View in Article
        • Grove W.M.
        • Zald D.H.
        • Lebow B.S.
        • Snitz B.E.
        • Nelson C.
        Clinical versus mechanical prediction: A meta-analysis.
        Psychological Assessment. 2000; 12: 19-30
        View in Article
        • Haver B.
        Comorbid psychiatric disorders predict and influence treatment outcomes in female alcoholics.
        European Addiction Research. 2003; 9: 39-44
        View in Article
        • Haver B.
        • Dahlgren L.
        • Willander A.
        A 2-year follow-up of 120 Swedish female alcoholics treated early in their drinking career: Prediction of drinking outcome.
        Alcoholism, Clinical and Experimental Research. 2001; 25: 1586-1593
        View in Article
        • Heather N.
        • Brodie J.
        • Wale S.
        • Wilkinson G.
        • Luce A.
        • Webb E.
        • et al.
        A randomized controlled trial of moderation-oriented cue exposure.
        Journal of Studies on Alcohol. 2000; 61: 561-570
        View in Article
        • Heather N.
        • Dawe S.
        Level of impaired control predicts outcome of moderation-oriented treatment for alcohol problems.
        Addiction. 2005; 100: 945-952
        View in Article
        • Heather N.
        • Rollnick S.
        • Bell A.
        Predictive validity of the Readiness to Change Questionnaire.
        Addiction. 1993; 88: 1667-1677
        View in Article
        • Hernandez-Avila C.A.
        • Burleson J.A.
        • Kranzler H.R.
        Stage of change as a predictor of abstinence among alcohol-dependent subjects in pharmacotherapy trials.
        Substance Abuse. 1998; 19: 81-91
        View in Article
        • Hodgins D.C.
        • el-Guebaly N.
        • Armstrong S.
        • Dufour M.
        Implications of depression on outcome from alcohol dependence: A 3-year prospective follow-up.
        Alcoholism, Clinical and Experimental Research. 1999; 23: 151-157
        View in Article
        • Hunter E.E.
        • Powell B.J.
        • Penick E.C.
        • Nickel E.J.
        • Liskow B.I.
        • Cantrell P.J.
        • et al.
        Comorbid psychiatric diagnosis and long-term drinking outcome.
        Comprehensive Psychiatry. 2000; 41: 334-338
        View in Article
        • Huriwai T.
        • Sellman J.D.
        • Sullivan P.
        • Potiki T.L.
        Optimal treatment for Maori with alcohol and drug-use-related problems: An investigation of cultural factors in treatment.
        Substance Use and Misuse. 2000; 35: 281-300
        View in Article
        • Jones B.T.
        • McMahon J.
        Negative alcohol expectancy predicts post-treatment abstinence survivorship: The whether, when and why of relapse to a first drink.
        Addiction. 1994; 89: 1653-1665
        View in Article
        • Jones B.T.
        • McMahon J.
        Negative and positive alcohol expectancies as predictors of abstinence after discharge from a residential treatment program: A one-month and three-month follow-up study in men.
        Journal of Studies on Alcohol. 1994; 55: 543-548
        View in Article
        • Jones B.T.
        • McMahon J.
        Changes in alcohol expectancies during treatment relate to subsequent abstinence survivorship.
        British Journal of Clinical Psychology. 1996; 35: 221-234
        View in Article
        • Kadden R.M.
        • Skerker P.M.
        Treatment decision making and goal setting.
        in: McCrady B.S. Epstein E.E. Addictions: A comprehensive guidebook. Oxford University Press, New York1999
        View in Article
        • Kavanagh D.J.
        • Sitharthan T.
        • Sayer G.P.
        Prediction of results from correspondence treatment for controlled drinking.
        Addiction. 1996; 91: 1539-1545
        View in Article
        • Kivlahan D.R.
        • Sher K.J.
        • Donovan D.M.
        The Alcohol Dependence Scale: A validation study among inpatient alcoholics.
        Journal of Studies on Alcohol. 1989; 50: 170-175
        View in Article
        • Kraemer H.C.
        • Stice C.
        • Kazdin A.
        • Offord D.
        • Kupfer D.
        How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors.
        American Journal of Psychiatry. 2001; 158: 848-856
        View in Article
        • Kranzler H.R.
        • Mulgrew C.
        • Modesto-Lowe V.
        • Burleson J.A.
        Validity of the obsessive compulsive drinking scale (OCDS): Does craving predict drinking behavior?.
        Alcoholism, Clinical and Experimental Research. 1999; 23: 108-114
        View in Article
        • Langenbucher J.
        • Chung T.
        • Morgenstern J.
        • Labouvie E.
        • Nathan P.E.
        • Bavly L.
        Physiological alcohol dependence as a “specifier” of risk for medical problems and relapse liability in DSM-IV.
        Journal of Studies on Alcohol. 1997; 58: 341-350
        View in Article
        • Langenbucher J.
        • Sulesund D.
        • Chung T.
        • Morgenstern J.
        Illness severity and self-efficacy as course predictors of DSM-IV alcohol dependence in a multisite clinical sample.
        Addictive Behaviors. 1996; 21: 543-553
        View in Article
        • Lemke S.
        • Moos R.H.
        Outcomes at 1 and 5 years for older patients with alcohol use disorders.
        Journal of Substance Abuse Treatment. 2003; 24: 43-50
        View in Article
        • Litt M.D.
        • Kadden R.M.
        • Cooney N.L.
        • Kabela E.
        Coping skills and treatment outcomes in cognitive–behavioral and interactional group therapy for alcoholism.
        Journal of Consulting and Clinical Psychology. 2003; 71: 118-128
        View in Article
        • Long C.G.
        • Hollin C.R.
        • Williams M.J.
        Self-efficacy, outcome expectations, and fantasies as predictors of alcoholics' posttreatment drinking.
        Substance Use and Misuse. 1998; 33: 2383-2402
        View in Article
        • Long C.G.
        • Williams M.
        • Hollin C.R.
        Alcoholism treatment: Intake variables as outcome predictors.
        Addiction Research. 1998; 6: 295-305
        View in Article
        • Long C.G.
        • Williams M.
        • Midgley M.
        • Hollin C.R.
        Within-program factors as predictors of drinking outcome following cognitive–behavioral treatment.
        Addictive Behaviors. 2000; 25: 573-578
        View in Article
        • Marra D.
        • Warot D.
        • Payan C.
        • Hispard E.
        • Dally S.
        • Puech A.J.
        Anhedonia and relapse in alcoholism.
        Psychiatry Research. 1998; 80: 187-196
        View in Article
        • McKay J.R.
        • Weis R.V.
        A review of temporal effects and outcome predictors in substance abuse treatment studies with long-term follow-ups.
        Evaluation Review. 2001; 25: 113-161
        View in Article
        • McLellan A.T.
        • Alterman A.I.
        • Metzger D.S.
        • Grissom G.R.
        • Woody G.E.
        • Luborsky L.
        • et al.
        Similarity of outcome predictors across opiate, cocaine, and alcohol treatments: Role of treatment services.
        Journal of Consulting and Clinical Psychology. 1994; 62: 1141-1158
        View in Article
        • McLellan A.T.
        • Luborsky L.
        • Woody G.E.
        • O'Brien C.P.
        An improved diagnostic evaluation instrument for substance abuse patients. The Addiction Severity Index.
        Journal of Nervous and Mental Disease. 1980; 168: 26-33
        View in Article
        • McLellan A.T.
        • Luborsky L.
        • Woody G.E.
        • O'Brien C.P.
        • Druley K.A.
        Predicting response to alcohol and drug abuse treatments. Role of psychiatric severity.
        Archives of General Psychiatry. 1983; 40: 620-625
        View in Article
        • McMahon J.
        • Jones B.T.
        The Negative Alcohol Expectancy Questionnaire.
        Journal of the Association of Nurses in Substance Abuse. 1993; 12: 17
        View in Article
        • McMahon J.
        • Jones B.T.
        Post-treatment abstinence survivorship and motivation for recovery: The predictive validity of the readiness to change (RCQ) and negative alcohol expectancy (NAEQ) questionnaire.
        Addiction Research. 1996; 4: 161-176
        View in Article
        • Meszaros K.
        • Lenzinger E.
        • Hornik K.
        • Fureder T.
        • Willinger U.
        • Fischer G.
        • et al.
        The Tridimensional Personality Questionnaire as a predictor of relapse in detoxified alcohol dependents. The European Fluvoxamine in Alcoholism Study Group.
        Alcoholism, Clinical and Experimental Research. 1999; 23: 483-486
        View in Article
        • Moncrieff J.
        • Drummond D.C.
        • Candy B.
        Sexual abuse in people with alcohol problems. A study of the prevalence of sexual abuse and its relationship to drinking behaviour.
        British Journal of Psychiatry. 1996; 169: 355-360
        View in Article
        • Monlar B.E.
        • Buka S.L.
        • Kessler R.C.
        Child sexual abuse and subsequent psychopathology: Results from the National Comorbidity Survey.
        American Journal of Public Health. 2001; 91: 753-760
        View in Article
        • Noda T.
        • Imamichi H.
        • Kawata A.
        • Hirano K.
        • Ando T.
        • Iwata Y.
        • et al.
        Long-term outcome in 306 males with alcoholism.
        Psychiatry and Clinical Neurosciences. 2001; 55: 579-586
        View in Article
        • Parsons O.A.
        • Schaeffer K.W.
        • Glenn S.W.
        Does neuropsychological test performance predict resumption of drinking in posttreatment alcoholics?.
        Addictive Behaviors. 1990; 15: 297-307
        View in Article
        • Powell B.J.
        • Penick E.C.
        • Nickel E.J.
        • Liskow B.I.
        • Riesenmy K.D.
        • Campion S.L.
        • et al.
        Outcomes of co-morbid alcoholic men: A 1-year follow-up.
        Alcoholism, Clinical and Experimental Research. 1992; 16: 131-138
        View in Article
        • Project MATCH Research Group
        Matching alcoholism treatments to client heterogeneity: Project MATCH three-year drinking outcomes.
        Alcoholism, Clinical and Experimental Research. 1998; 22: 1300-1311
        View in Article
        • Rice C.
        • Longabaugh R.
        • Beattie M.
        • Noel N.
        Age group differences in response to treatment for problematic alcohol use.
        Addiction. 1993; 88: 1369-1375
        View in Article
        • Rohsenow D.J.
        • Monti P.M.
        • Rubonis A.V.
        • Sirota A.D.
        • Niaura R.S.
        • Colby S.M.
        • et al.
        Cue reactivity as a predictor of drinking among male alcoholics.
        Journal of Consulting and Clinical Psychology. 1994; 62: 620-626
        View in Article
        • Rounsaville B.J.
        • Dolinsky Z.S.
        • Babor T.F.
        • Meyer R.E.
        Psychopathology as a predictor of treatment outcome in alcoholics.
        Archives of General Psychiatry. 1987; 44: 505-513
        View in Article
        • Rychtarik R.G.
        • Prue D.M.
        • Rapp S.R.
        • King A.C.
        Self-efficacy, aftercare and relapse in a treatment program for alcoholics.
        Journal of Studies on Alcohol. 1992; 53: 435-440
        View in Article
        • Sellman J.D.
        • Joyce P.R.
        Does depression predict relapse in the 6 months following treatment for men with alcohol dependence?.
        Australian and New Zealand Journal of Psychiatry. 1996; 30: 573-578
        View in Article
        • Sellman J.D.
        • Mulder R.T.
        • Sullivan P.F.
        • Joyce P.R.
        Low persistence predicts relapse in alcohol dependence following treatment.
        Journal of Studies on Alcohol. 1997; 58: 257-263
        View in Article
        • Shaw G.K.
        • Waller S.
        • Latham C.J.
        • Dunn G.
        • Thomson A.D.
        The detoxification experience of alcoholic in-patients and predictors of outcome.
        Alcohol and Alcoholism. 1998; 33: 291-303
        View in Article
        • Skinner H.A.
        • Allen B.A.
        Alcohol dependence syndrome: Measurement and validation.
        Journal of Abnormal Psychology. 1982; 91: 199-209
        View in Article
        • Sobell M.B.
        • Sobell L.C.
        • Gavin D.R.
        Portraying alcohol treatment outcomes: Different yardsticks of success.
        Behavior Therapy. 1995; 26: 643-669
        View in Article
        • Solomon K.E.
        • Annis H.M.
        Outcome and efficacy expectancy in the prediction of post-treatment drinking behaviour.
        British Journal of Addiction. 1990; 85: 659-665
        View in Article
        • Staines G.
        • Magura S.
        • Rosenblum A.
        • Fong C.
        • Kosanke N.
        • Foote J.
        • et al.
        Predictors of drinking outcomes among alcoholics.
        American Journal of Drug and Alcohol Abuse. 2003; 29: 203-218
        View in Article
        • Stockwell T.
        • Murphy D.
        • Hodgson R.
        The severity of alcohol dependence questionnaire: Its use, reliability and validity.
        British Journal of Addiction. 1983; 78: 145-155
        View in Article
        • Terrell M.D.
        Ethnocultural factors and substance abuse: Towards culturally sensitive treatment models.
        Psychology of Addictive Behaviors. 1993; 7: 162-167
        View in Article
        • Tomasson K.
        • Vaglum P.
        Psychopathology and alcohol consumption among treatment-seeking alcoholics: A prospective study.
        Addiction. 1996; 91: 1019-1030
        View in Article
        • Verheul R.
        • Van Den Brink W.
        • Hartgers C.
        Personality disorders predict relapse in alcoholic patients.
        Addictive Behaviors. 1998; 23: 869-882
        View in Article
        • Wild T.C.
        Social control and coercion in addiction treatment: Towards evidence-based policy and practice.
        Addiction. 2006; 101: 40-49
        View in Article
        • Zywiak W.H.
        • Longabaugh R.
        • Wirtz P.W.
        Decomposing the relationship between pretreatment social network characteristics and alcohol treatment outcome.
        Journal of Studies on Alcohol. 2002; 63: 114-121
        View in Article

      Article info

      Publication history

      Published online: July 28, 2008
      Accepted: May 5, 2008
      Received in revised form: April 16, 2008
      Received: December 21, 2007

      Identification

      DOI: https://doi.org/10.1016/j.jsat.2008.05.007

      Copyright

      © 2009 Elsevier Inc. Published by Elsevier Inc. All rights reserved.

      ScienceDirect

      Access this article on ScienceDirect

      The content on this site is intended for healthcare professionals.


      We use cookies to help provide and enhance our service and tailor content. To update your cookie settings, please visit the Cookie Preference Center for this site.
      Copyright © 2023 Elsevier Inc. except certain content provided by third parties.


      RELX