Variance In Treatment Outcome Research Articles

Why do interpretations account for relatively little variance in treatment outcome?

Why do interpretations account for relatively little variance in treatment outcome?

The variability of client, therapist and clinic in psychotherapy outcomes: A three‐level hierarchical model

AbstractThis article investigates the percentage of variance in psychotherapy outcomes explained by the case‐mix variables for individual cases, by the therapist (therapist effects), and what additional variance is explained by the clinic with which the therapist is affiliated. While there has been substantial recent research regarding therapist effects, very little has been published regarding clinic‐level effects after controlling for therapist effects. The study utilised the largest sample reported to date, using data from 28 clinics with a minimum of 2,000 cases in a clinical range of severity of symptoms with pre‐post change scores on an outcome questionnaire. Only cases treated by a therapist with at least 30 cases were included. These selection criteria resulted in a case count of 156,258 clients treated by 874 therapists located at 28 clinics. After controlling for differences in case mix using diagnosis and intake score (severity of symptoms), the resulting analyses indicated that 5.21% of the variance in treatment outcome was explained by the therapist, while another 1.13% of the variance was associated with the clinic. Findings are discussed with implications for practice and policy.

Preliminary evidence of contextual factors' influence on weight loss treatment outcomes: implications for future research.

Background/Objectives:Behavioral health interventions, including behavioral obesity treatment, typically target psychosocial qualities of the individual (e.g., knowledge, self-efficacy) that are largely treated as persistent, over momentary contextual factors (e.g., affect, environmental conditions). The variance in treatment outcomes that can be attributable to these two sources is rarely quantified but may help inform future research and treatment development efforts.Subjects/Methods:The intraclass correlation coefficient (ICC) for weekly weight loss was calculated in three studies involving 10–12 weeks of behavioral obesity treatment delivered to adults via in-person group sessions, mobile application, or website. The ICC explains the proportion of variance between versus within individuals, and was used to infer the contribution of individual versus contextual factors to weekly weight loss. The analytic approach involved unconditional linear mixed effect models with a random effect for subject.Results:The ICCs were very low, ranging from 0.01 to 0.06, suggesting that momentary contextual factors may influence obesity treatment outcomes to a substantial degree.Conclusions:This study yielded preliminary evidence that the influence of contextual factors in behavioral obesity treatment may be underappreciated. Future research is needed to simultaneously identify and quantify sources of within- and between-subjects variance to optimize treatment approaches.

How therapists' emotion recognition relates to therapy process and outcome.

Empathy is an essential characteristic for therapists that explained 9% of the variance in treatment outcomes. Many measures of empathy are based on self-report questionnaires. Therefore, they reflect how a person perceives his empathic abilities, which can be biased from his true abilities. The ability to recognize emotions has been found to be related to empathy and serves as a measure for the use of empathic abilities. In this study, we examined therapists' empathic abilities, using a novel task for recognizing emotions, and looked at how they related to the therapy process and outcomes. The study included 33 patient-therapist dyads. Therapist empathy was assessed with an emotion detection task, the JeFEE. Clients filled questionnaires after each session regarding therapy progress and their symptoms. We found that emotion recognition moderated the change in (a) client secure attachment to therapist, (b) client avoidant attachment to therapist, (c) client working alliance, (d) client rate of tense or upset they felt during the session, (e) client lack of emotional clarity of emotions, (f) client non-acceptance of emotional responses, (g) client overall emotion regulation and (h) client main target complaint. Implication for therapy and therapists' selection and training are discussed.

Brain markers predicting response to cognitive\u2010behavioral therapy for social anxiety disorder: an independent replication of Whitfield-Gabrieli et al. 2015

Predictive brain markers promise a number of important scientific, clinical, and societal applications. Over 600 predictive brain markers have been described in published reports, but very few have been tested in independent replication attempts. Here, we conducted an independent replication of a previously published marker predicting treatment response to cognitive-behavioral therapy for social anxiety disorder from patterns of resting-state fMRI amygdala connectivity1. The replication attempt was conducted in an existing dataset similar to the dataset used in the original report, by a team of independent investigators in consultation with the original authors. The precise model described in the original report positively predicted treatment outcomes in the replication dataset, but with marginal statistical significance, permutation test p = 0.1. The effect size was substantially smaller in the replication dataset, with the model explaining 2% of the variance in treatment outcomes, as compared to 21% in the original report. Several lines of evidence, including the current replication attempt, suggest that features of amygdala function or structure may be able to predict treatment response in anxiety disorders. However, predictive models that explain a substantial amount of variance in independent datasets will be needed for scientific and clinical applications.

Treatment Evaluation in Inpatient Psychiatry: Transdiagnostic Factors as Correlates of Treatment Outcome - Implications for Clinical Practice and for Effectiveness Research

An inpatient psychiatric-psychotherapeutic treatment was evaluated in a naturalistic setting (effectiveness research). Transdiagnostic factors were examined as correlates of treatment outcome. Self-report questionnaires were administered at the beginning and the end of the inpatient treatment. The sample (N = 378) consisted of women and men (16-80 years, M = 40.4, SD = 15.8) with primary diagnoses of depressive, anxiety or eating disorders. The average treatment duration was 8.2 weeks (SD = 3.8). Primary outcome variables were severity of symptoms and level of psychosocial functioning. As transdiagnostic correlates, self-esteem, self-efficacy, and emotion regulation were included in the regression analyses. The change in transdiagnostic factors explained a significant proportion of the treatment outcome (explained variance: 8-27 % self-efficacy, 23-42 % self-esteem, 10-26 % emotion regulation). The present study shows that the evaluation of a treatment in a naturalistic clinical setting is feasible. Transdiagnostic factors explained a substantial part of the statistical variance in treatment outcome across all disorders. Even if the present study design does not allow conclusions about causality, the results underline the relevance of transdiagnostic trainings (e. g., in emotion regulation) to improve effectiveness of treatment approaches.

Sex Differences in Neuromodulation Treatment Approaches for Traumatic Brain Injury: A Scoping Review.

Neuromodulatory brain stimulation interventions for traumatic brain injury (TBI)-related health sequelae, such as psychiatric, cognitive, and pain disorders, are on the rise. Because of disproportionate recruitment and epidemiological reporting of TBI-related research in men, there is limited understanding of TBI development, pathophysiology, and treatment intervention outcomes in women. With data suggesting sex-related variances in treatment outcomes, it is important that these gaps are addressed in emerging, neuromodulatory treatment approaches for TBI populations. Four research databases (PubMED, EMBASE, CINAHL, and PsycINFO) were electronically searched in February 2020. This PRISMA Scoping Review (PRISMA-ScR)-guided report contextualizes the importance of reporting sex differences in TBI + neuromodulatory intervention studies and summarizes the current state of reporting sex differences when investigating 3 emerging interventions for TBI outcomes. Fifty-four studies were identified for the final review including 12 controlled trials, 16 single or case series reports, and 26 empirical studies. Across all studies reviewed, 68% of participants were male, and only 7 studies reported sex differences as a part of their methodological approach, analysis, or discussion. This review is hoped to update the TBI community on the current state of evidence in reporting sex differences across these 3 neuromodulatory treatments of post-TBI sequelae. The proposed recommendations aim to improve future research and clinical treatment of all individuals suffering from post-TBI sequelae.

Low-Dose Tacrolimus Prevents Dysregulated Peri-Conceptional Ovarian and Systemic Immune Cellular Homeostasis in Subjects with PCOS

Polycystic ovary syndrome (PCOS) is characterized by failure of ovulation and is associated with obesity and chronic inflammation. Recent evidence suggests that anomalous activation of ovarian macrophages and numerical and functional deficits in the Th17 (CD4+IL17A+) and the CD4+CD25+CD127low Tregs plays crucial role in PCOS. We have shown that the pre-pregnancy use of tacrolimus prevents adverse reproductive outcomes in a mouse model of PCOS. Here we used the HFD-NONcNZO mice to test a hypothesized beneficial use of tacrolimus relative to metformin in favorably influencing the ovarian and systemic immune milieux conducive to gestational success in subjects with PCOS. Compared to normative controls, our data revealed an aberrant peri-conceptional suppression of the CD4+CD25+CD127low Tregs together with an overexpression of the Th17 T cells and lack of coordinated activation of ovarian macrophages in untreated HFD-dNONcNZO mice. Significant variances in treatment outcomes favoured the use of tacrolimus over metformin in treated mice. Consistent with the human fertility studies, this investigation reveals a beneficial systemic use of tacrolimus (0.1 mg/kg) in promoting early pregnancy in individuals with PCOS and suggests the need for further research into the selective inhibition of IL17A as a plausibly alternative immunotherapeutic approach in the clinical management of infertile individuals with PCOS.

Better patients' treatment experiences are associated with better postoperative results in Dupuytren's disease.

This prospective study investigates the extent to which a better experience with healthcare delivery is associated with better postoperative treatment outcomes after surgery for Dupuytren’s contracture. Patients undergoing limited fasciectomy or percutaneous needle fasciotomy for Dupuytren’s contractures completed the Michigan Hand Outcomes Questionnaire before and 3 months after surgery, together with a patient reported experience measure, while hand therapists assessed the straightness of the finger with a goniometer. Regression analyses were used to examine associations. We found that a better experience with healthcare delivery was associated with better patient-reported outcomes, while association with residual extension deficit was minimal. Strongest associations were seen with communication of the physician, postoperative care and information about the treatment. Experience with the treatment explained up to 12% of the variance in treatment outcome. These findings suggest that patient reported treatment outcomes in Dupuytren’s disease can be improved by improving the treatment context.Level of evidence: II

A multi-service practice research network study of large group psychoeducational cognitive behavioural therapy

BackgroundThis was a multi-service evaluation of the clinical and organisational effectiveness of large group psychoeducational CBT delivered within a stepped care model. MethodClinical outcomes for 4451 participants in 163 psychoeducational groups delivered across 5 services were analysed by calculating pre-post treatment anxiety (GAD-7) effect sizes (Cohen's d). Overall and between-service effects were compared to published efficacy benchmarks. Multilevel modelling was used to examine if variability in clinical outcomes was explained by differences in service, group and patient-level (case-mix) variables. ResultsThe pooled GAD-7 (pre-post) effect size for all services was d = 0.70, which was consistent with efficacy benchmarks for guided self-help interventions (d = 0.69). One service had significantly smaller effects (d = 0.48), which was explained by differences in group treatment length and case-mix. Variability between groups (i.e., group effects) explained up to 3.6% of variance in treatment outcomes. ConclusionsLarge group psychoeducational CBT is clinically effective, organisationally efficient and consistent with a stepped care approach to service design. Clinical outcome differences between services were explained by group and patient variables.

An inflammatory biomarker as a differential predictor of outcome of depression treatment with escitalopram and nortriptyline.

Major depressive disorder has been linked with inflammatory processes, but it is unclear whether individual differences in levels of inflammatory biomarkers could help match patients to treatments that are most likely to be beneficial. The authors tested the hypothesis that C-reactive protein (CRP), a commonly available marker of systemic inflammation, predicts differential response to escitalopram (a serotonin reuptake inhibitor) and nortriptyline (a norepinephrine reuptake inhibitor). The hypothesis was tested in the Genome-Based Therapeutic Drugs for Depression (GENDEP) study, a multicenter open-label randomized clinical trial. CRP was measured with a high-sensitivity method in serum samples from 241 adult men and women with major depressive disorder randomly allocated to 12-week treatment with escitalopram (N=115) or nortriptyline (N=126). The primary outcome measure was the score on the Montgomery-Åsberg Depression Rating Scale (MADRS), administered weekly. CRP level at baseline differentially predicted treatment outcome with the two antidepressants (CRP-drug interaction: β=3.27, 95% CI=1.65, 4.89). For patients with low levels of CRP (<1 mg/L), improvement on the MADRS score was 3 points higher with escitalopram than with nortriptyline. For patients with higher CRP levels, improvement on the MADRS score was 3 points higher with nortriptyline than with escitalopram. CRP and its interaction with medication explained more than 10% of individual-level variance in treatment outcome. An easily accessible peripheral blood biomarker may contribute to improvement in outcomes of major depressive disorder by personalizing treatment choice.

The Importance of Group Cohesion in Inpatient Treatment of Combat-Related PTSD

Post-traumatic stress disorder (PTSD) is the most widespread mental illness resulting from exposure to combat, necessitating an increase in the provision of group therapy. This pilot study examined the efficacy of, and treatment outcome predictors associated with, group inpatient treatment of combat-related PTSD. Participants included 38 active duty military personnel deployed during Operation Enduring Freedom (OEF) and Operation Iraqi Freedom (OIF), diagnosed with PTSD, and consecutive admissions to an inpatient PTSD treatment facility. A paired samples t-test revealed significant change in symptom severity and global functioning between pre- and post-treatment. Multiple regression analyses supported the predictive utility of baseline symptomatology and group cohesion (> 50% of the variance in treatment outcome), highlighting the importance of group cohesion in the efficacy of group treatment for combat-related PTSD.

Homework Compliance Counts in Cognitive-Behavioral Therapy

Background: Prior research has demonstrated that there is some association between treatment engagement and treatment outcome in behavioral therapy for anxiety disorders. However, many of these investigations have been limited by weak measurement of treatment engagement variables, failure to control for potentially important baseline variables, and failure to consider various treatment engagement variables simultaneously. The purpose of the present study is to examine the relationship between two treatment engagement variables (treatment expectancy and homework compliance) and the extent to which they predict improvement from cognitive-behavioral therapy (CBT) for anxiety disorders. Methods: 84 adults with a DSM-IV-defined principal anxiety disorder took part in up to 12 sessions of CBT or acceptance and commitment therapy. Pre- and post-treatment disorder severity was assessed using clinical severity ratings from a semi-structured diagnostic interview. Participants made ratings of treatment expectancy after the first session. Homework compliance was assessed each session by the treating clinician. Results: Contrary to hypotheses, treatment expectancy and homework compliance were poorly correlated. Regression analyses revealed that homework compliance, but not treatment expectancy, predicted a significant portion of the variance in treatment outcome (10%). Conclusions: The present research suggests that although treatment expectation and homework compliance likely represent unique constructs of treatment engagement, homework compliance may be the more important treatment engagement variable for outcomes. The present research suggests that improvement of homework compliance has the potential to be a highly practical and effective way to improve clinical outcomes in CBT targeting anxiety disorders.

Predictors for Treatment Outcome of Binge Eating With Obesity: A Naturalistic Study

This study examines predictors of short-term treatment outcome for obese individuals with binge eating disorder (BED). A battery of assessment questionnaires was given to 212 patients on admission of a CBT day-treatment program for BED. Treatment outcome assessed by changes in eating disorder symptomatology was measured in 182 completers. Linear regression analyses indicated that a combination of variables at baseline predicted 26% of the variance in treatment outcome. High social embedding and higher scores on openness (NEO-PI-R) were significantly related to more improvement after treatment. Higher scores on depressive symptoms (BDI), agoraphobia (SCL-90) and extraversion (NEO-PI-R) were significantly related to less improvement. The analyses show that the level of social embedding and psychopathological comorbidity (state and trait) are predictors for treatment outcome. This study confirms the notion that social context and comorbidity need to be taken into account as described in treatment guidelines of NICE and APA for BED.

A Perspective on the Effectiveness of Interventions for Alcohol and other Substance Use Disorders

Together, the 2 papers in this In Review series'- on the effectiveness of interventions for alcohol use disorders (AUDs) represent a comprehensive synthesis of the relevant literature and will serve as the go-to resource for some years for Canadian and international researchers, as well as for government and community decision-makers involved in funding, developing policy, and administering and delivering treatment and support services. As comprehensive as the 2 papers are, Dr Bernard Le Foil and colleagues (see Lev-Ran S et al') and Dr Garth W Martin and Dr Jurgen Rehnr acknowledge several limitations that the reader should keep in mind, most notably the limited attention given to special populations, such as youth, women, older adults, and diverse communities. For some of these subpopulations (for example. First Nations or subpopulations based on gender diversity) the research findings are not substantive enough to draw firm conclusions regarding the most effective interventions. In other instances, such as the treatment of adolescents and youth, it was a matter of keeping the reviews in reasonable scope and managing space limitations. Certainly the literature on the treatment of adolescents and youth is very well developed-' and strong enough to warrant the same general conclusion as presented in the 2 review papers'2; namely, that substance use treatment works and returns the economic investment through reductions in health care use and criminal justice and other costs at the societal level. Today's most relevant questions for intervention research focus on what works best from a comparative point of view, under what modifying or mediating conditions, and with what degree of cost-effectiveness. Aside from issues related to subpopulations, there are 5 broad contextual factors that the reader must consider when reading and distilling key learnings from the 2 reviews.'1,2 First, the charge given to the authors was to synthesize the treatment literature related to alcohol abuse and dependence and not other psychoactive drugs of abuse. While there is some legitimacy to the development of substance-specific interventions (for example, specific pharmacologic interventions for AUDs). the trend during the last several years in most specialized treatment settings has been toward a diverse, multiple drug-using client population. The treatment population has diversified to the extent that programs are often challenged to identify the primary drug of abuse. In Ontario, for example, the most recent data from the provincial Drug and Alcohol Treatment Information System (commonly referred to as DATIS) yield the following breakdown - alcohol only 33%. drug only 31%, and combined alcohol and drug 36% (excluding family members who are also seen as system clients in their own right). These data have held steady for the past several years. By implication, interventions need to be as adaptable as possible across substances, while recognizing the need for some specialization for pharmacological intervention. It also means that treatment research samples, even if they are cleaned to reduce contamination by substances other than the drug of primary interest for the research question, yield study results that may or may not apply in real-life, mixed drug-using treatment populations. Thus both efficacy trials and field effectiveness studies are needed to advance the substance use field. The second important factor to keep in mind in applying the results of these 2 important reviews is that, while they focus on the interventions, someone has to engage the client and deliver them with competence, respect, and empathy. The paper by Dr Martin and Dr Rehm2 notes the importance of therapist effectsp 352 and these effects are by no means trivial. In the adult and youth literature alike it is surprising how few clinically significant differences emerge in studies comparing the effectiveness of different interventions. The psychotherapy literature suggests that therapeutic alliance may account for as much as 30% of the variance in treatment outcome. …

Alcohol expectancies and social self-efficacy as mediators of differential intervention outcomes for college hazardous drinkers with social anxiety

The current pilot study examined the roles of two cognitive factors – positive alcohol expectancies of social anxiety reduction and drink refusal self-efficacy relevant to social situations – in mediating greater reduction in alcohol behaviors by the Brief Intervention for Socially Anxious Drinkers (BISAD; n = 21) compared to an alcohol psychoeducation ( n = 20) in a sample of college hazardous drinkers with social anxiety. Mediation analysis results indicated that decreased positive alcohol expectancies and increased drink refusal self-efficacy relevant to social situations accounted for an average of 67% of the variance in treatment outcomes as measured by total quantity of alcohol consumption, heavy drinking days and problems related to alcohol use in the past month. Study results may enhance the understanding of cognitive factors' role in alcohol treatment outcomes, which could in turn improve the efficacy of interventions aimed to reduce hazardous drinking and comorbid social anxiety.

T1348 Predictors of Outcome of Biofeedback Therapy in Patients With Constipation

Background: Ano-rectal biofeedback therapy (BFT) is a safe and effective treatment in patients with constipation. Given the high prevalence of constipation, there is a need to prioritise patients. Therefore, the aim of this study was to further explore factors which predict success or failure of BFT. Several previous studies have examined this in part, with the only consistent finding being that the patient's willingness to participate predicts success of therapy. Methods: 102 consecutive patients (mean age 48±2 yrs, 88 females) with constipation referred for anorectal BFT were evaluated. All patients completed the Rome questionnaire, a validated constipation questionnaire, and 10 cm visual analogue scales for willingness to participate in BFT, satisfaction with bowel habit and impact on quality of life. All patients underwent comprehensive anorectal manometry and balloon expulsion test. Patients were enrolled in a 6 weekly-visit BFT program. Independent predictors were identified using a bootstrapped backward elimination procedure based on linear regression with change in bowel satisfaction as the outcome. Only predictors selected in ≥50% of bootstrap samples were considered. Results: 96 patients completed the full course of BFT. Stool consistency (p=0.009), willingness to participate (p<0.001), balloon expulsion time (p=0.02) and resting anal sphincter pressure (p=0.04) were positively correlated with an improvement in bowel satisfaction scores. On the other hand, more laxative use (p=0.049) and a greater satisfaction with bowel habit at baseline (p<0.001) correlated with no improvement in bowel satisfaction scores. The outcome of BFTwas not influenced by age, duration of symptoms or compliance with therapy. Multiple linear regression indicated that stool consistency (p=0.02) and willingness to participate (p= 0.003) independently predicted an improvement in bowel satisfaction scores after BFT, and total constipation score (p=0.006) and baseline satisfaction with bowel habit (p<0.001) independently predicted worsening in bowel satisfaction scores after BFT. The model including these 4 variables explained 52% of the variance in treatment outcome. Conclusion: In agreement with previous studies, success of BFT is predicted by a greater willingness to participate. New findings are that (1) harder stools predict a successful outcome, while more prolonged balloon expulsion and anal sphincter hypertonia at baseline are associated with greater bowel satisfaction after BFT; and (2) the presence of more severe constipation predicts failure of BFT. This information should be taken into account when prioritising patients for BFT.

Editorial: Evaluating Interventions in Child Mental Health: The Importance of Care Provider and Contextual Influences

B R U C E WA M P O L D I N H I S B O O K The great psychotherapy debate, highlights the tension between a narrow, ‘medical’ model of evaluating psychological treatments and a broader evaluative framework which takes contextual factors into account, including the influence of the therapists themselves on treatment outcome (Wampold, 2001). Such ideas have fuelled an interesting debate in the field of psychological treatment research and a special section of the journal Psychotherapy Research was recently devoted to the topic of the extent to which therapist effects had a bearing on treatment outcome (Hill, 2006). Wampold’s main contention is that therapists account for a greater proportion of the variance in treatment outcome than any particular model of psychological intervention and hence greater attention should be paid to the attributes of effective therapists. One of the studies to be examined most in relation to this issue is the NIMH Treatment of Depression Collaborative Research Program (TDCRP) Multicentre Randomized Controlled Trial (RCT). This RCT compared the outcome of Cognitive Behaviour Therapy (CBT) vs. Interpersonal Psychotherapy (IPT) vs. Antidepressant medication vs. Placebo in adults with Major Depression. Minimal differences between different models of intervention were identified, and subsequent re-analyses of the data looking for therapist effects in the CBT and IPT arms came to conflicting conclusions. One analysis found significant therapist effects (Kim, Wampold, & Bolt, 2006) while another, using different statistical methods, found no effects (Elkin, Falconnier, Martinovich, & Mahony, 2006)! What is the relevance of all this to child mental health? Intuitively as practitioners many of us will hold the view that the qualities and skills of the person delivering a particular treatment and the setting in which the treatment is undertaken may be as important as the actual model of intervention itself. However, this issue has received little attention in the field of child mental health research. This is likely to change in future following the publication of the CONSORT guidance extension for RCTs of nonpharmacologic treatments (Boutron, Moher, Altman, Schulz, & Ravaud, 2008). CONSORT, or the Consolidated Standards of Reporting Trials, is the gold standard guidance by which RCTs are reported. Researchers generally undertake to conduct and report on trials according to CONSORT. The new guidance, published this year, emphasizes that in fields such as surgery or psychotherapy, where the intervention is generally complex and multi-dimensional and influenced by the skills and expertise of the ‘care provider’ as well as the setting in which the treatment is undertaken, care must

In Search of the Active Ingredient: What Really Works in Mental Health Care?

AbstractThis brief report provides a concise overview of decades of research into what really makes a difference in treatment outcomes for the most common psychiatrics disorders. The evidence points to the conclusion that the relationship with the clinician accounts for much more of the variance in treatment outcomes than the treatment methods themselves. The discussion of research and statistical methodology are relevant to investigations of treatments for speech and language disorders.

Patient predictors of alcohol treatment outcome: A systematic review

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.