PROTOCOL: Cognitive-Behavioural Therapies for Young People in Outpatient Treatment for Non-Opioid Drug Use

Behavioural therapy

Behavioural therapy was developed from the ideas of classical and operant conditioning (Poulsen, 2006, McGuire, 2000). In classical conditioning, behaviour is believed to be affected by stimulus-response mechanisms in the immediate surroundings of the individual; for instance, urges and cravings for drugs can be perceived as responses to external stimuli cues (Sherman, Jorenby & Baker, 1988). Identifying external stimuli cues would enable the individual to avoid settings that work as triggers to drug taking (Carroll, 2008). Operant conditioning is based on associations within a context of events (e.g.,an antecedent stimulus) and a given behaviour and its consequences, whereby perceived rewards can (negatively or positively) reinforce such behaviour (Skinner, 1988). For example, negative reinforcement is when peers do not condone drug use and positive reinforcementis for example when the psychological effect of a drug is experienced as pleasurable (Waldron & Kaminer, 2004). In a treatment context non-drug using behaviour is rewarded and thus reinforced.

Cognitive therapy

The assumption in cognitive therapy is that thoughts shape feelings and thereby, behaviour, so that it is hypothesized that by changing thought patterns, behaviour can be changed as well (Beck, Wright, Newman, & Liese 1993; Kendal l 2006; Nielsen & Thomsen, 2005). In the early 1960s cognitive therapy was aimed at treating depression, and has since been extensively modified and adapted to deal with a wide range of clinical problems and populations including people with drug use issues (Beck, 2008; Holmbeck et al., 2006; Weisz & Hawley, 2002).

Cognitive and behavioural therapy

The foundation and premise of CBT for drug use is that cognitive techniques and skills training can tackle drug-related beliefs and automatic thoughts that lead to urges and craving, while additional behavioural techniques can deal with actions that interact with the individual's cognitive processes that trigger and maintain drug using behaviour (Beck et al., 1993). Irrational and erroneous assumptions can cause and/or maintain undesirable behaviour (ibid.). CBT callsspecific attention to the propensity among substance users to mistakenly believe that the perceived advantages of using drugs (e.g., pleasure, anxiety relief) are greater than the disadvantages (e.g., financial, interpersonal) as such misconceptions help sustain the avoidance of a realistic assessment of the disadvantages (ibid.; Carroll, 2008; Nielsen & Thomsen, 2005). Thus, it is believed that the users' assessment of the possibilities for ceasing to use drugs might be based on cognitive distortions. In CBT, clients are helped to identify and challenge dysfunctional beliefs (such as ‘I cannot be happy unless I am using’ or ‘the withdrawal will be too painful‘), because thinking that one is incapable of controlling the urge to use drugs will create a self-fulfilling prophecy, as users who believe they are incapable will not even try (Beck et al., 1993).

The common denominator in all CBT interventions is to make and support continuous positive change in the client's feelings and behaviour by examining and reframing the basic maladaptive assumptions and thoughts underlying drug use (Beck, 2008; Carroll, 2008, Moos, 2007; McGuire, 2000).

CBT outlines a pattern and series of phases of drug use from the first stimulating cue to the actual act of drug using that is specific to the client. The activating stimulus can be both external (e.g., a gathering of friends using cocaine) or internal (e.g., anxiety or boredom) (Beck et.al., 1993; Beck, 2008; McGuire, 2000; Nielsen & Thomsen, 2005; Carroll, 2008). These stimuli can trigger basic assumptions (e.g., ‘I am socially isolated’) that trigger automatic thoughts (e.g. ‘A little cocaine will make me feel better’), which in turn trigger cravings and permissive beliefs that make it easier for the person to engage in the behaviour (e.g., ‘It is okay as long as I don't inject’). The individual would then form a mental strategy for obtaining the drugs and the actual drug using act could then take place. CBT would tackle this pattern of drug use by enlisting a number of techniques and strategies. Through problem solving, coping strategies, rehearsal, social skills and communication training, as well as helping young people to respond to criticism and refusing drugs, the therapist can help the young person to identify stimulating cues, discuss how to cope, and avoid drug taking behaviour. However, some stimulating cues (e.g., emotional states) may be unavoidable and consequently modifying maladaptive beliefs and automatic thought patterns that maintain drug using behaviour would be equally important (Beck et al., 1993; Beck, 2008; McGuire, 2000)

CBT components and therapy sessions

CBT interventions couldincludepermutations of various components such as thought diaries, social skills training, problem solving strategies, coping strategies, self-control and stress management techniques, and relapse prevention training. CBT has different modalities and can be implemented in an individual and/or group setting (Moos, 2007).

CBT is a highly structured intervention and is organised closely around well-specified and individualized treatment goals (Carroll, 2008). Each CBT session is structured by an articulated agenda and discussions remain focused around issues directly related to substance use. In some cases, the therapist may lead thetherapy session with ‘motivational interviews‘ ⁶ (Carroll, 2008; Nielsen & Thomsen, 2005).

Typical Therapy Sessions

To exemplify:a therapy session typically (but not invariably)includes the following three parts: First, aclient's substance use and general functioning would be assessed (and would vary according to degree of dependency and individual conditions). A specific cognitive technique that can help identify and modify drug-related beliefs is an'advantages-disadvantages' analysis (Beck et al.,1993). In this analysis, the therapist guides the client through the process of listing and re-evaluating the advantages and disadvantages of drug use to help the young person gain a more accurate, objective and balanced view of drug use.

The second part of the therapy session is typically didactic in structure and devoted to skills training, coping and problem-solving strategies and practice. One technique for examining beliefs and considering their validity in a more systematic way is ‘The Daily Thought Record‘. Clients are asked to record their thoughts and feelings and then re-evaluate their validity, identify possible patterns of cognitive distortions and develop strategies for change (Beck et al., 1993, Nielsen & Thomsen, 2005; Carroll, 2008). The therapist may also encourage the client to try new behaviours through role playing, for the purpose of teaching the client new effective interpersonal skills, e.g., how to handle interpersonal conflicts without drug taking and develop effective repertoires of social behaviour to reduce undesirable drug use and deal with relapse if it occurs (Beck et al., 1993; Kaminer & Waldron, 2006).

Finally, the third part of the therapy session is usually dedicated to plan for the week ahead and discuss how new skills and strategies could be implemented (Carroll, 2008). This kind of collaborative empiricism that characterises CBT is particularly important when dealing with young substance users, to assist them in learning self-regulation and to exert self-control. However this kind of collaboration may also be a point of concern for the intervention's effectiveness, as participation in CBT demands a certain (above average) level of verbal articulation and self-awareness (Nielsen & Thomsen, 2005).

CBT interventions can range from 5 to 24 weeks in duration and delivery settings can vary from outpatientto community facilities, and can be delivered to individuals, groups, families and a combination of these (Dennis et al., 2004; Carroll, 2008). Purely behavioural (e.g., a standalone contingency intervention) will not be considered in this review.

Existing research

Along with a handful of other interventions, CBT is one of the most researched treatment forms (Becker & Curry, 2008; Carroll, 2008).CBT has shown promising potential for young drug users in the number of primary studies (Kaminer et. al., 1998; Kaminer & Burleson, 1999;Waldron, Slesnick, Brody, Peterson, & Turner, 2001; Kaminer, Burleson & Goldberger 2002; Dennis et al., 2004; Azrin, Donohue, Teichner, Crum, Howell & DeCato 2001; Liddle et al., 2001; Liddle, Dakof, Turner, Henderson & Greenbaum, 2008; Latimer, Winters, D'Zurilla & Nichols, 2003).

Several reviews (that for the most lack pre-published protocols) ⁷ on CBT interventions targeting young drug users already exist (Waldron & Kaminer, 2004; Vaughn & Howard, 2004; Becker & Curry 2008; Waldron & Turner 2008; Lipsey, Tanner-Smith, & Wilson, 2010). However, with only one exception (Waldron & Kaminer, 2004), all of the abovefocus broadly on psychosocial therapies in general, rather than CBT specifically. Generally, the most recent reviews conclude that CBT is associated with reduced drug use in young people (Waldron & Kaminer, 2004; Waldron & Turner, 2008; Lipsey, Tanner-Smith, & Wilson et al., 2010).

The findings of the aforementioned studies and reviews indicate that CBT can reduce drug use in young people in treatment. However, closer interpretation of findings reveals a complex picture that is far from clear cut. CBTs reduction in drug use is relative to the comparison interventions used in the individual studies (Lipsey, Tanner-Smith, & Wilson, 2010) and dependent on the types of CBT interventions and modalities used in the studies.

CBT Mechanisms

Lack of research on mechanisms of change specifically underpinning CBT (Waldron & Kaminer, 2004)makeany identification of key mechanisms speculative. Nevertheless, problem solving and coping strategy skills may be a key to change. Myers and Brown (1990) found that young drug abstainers and minor relapsers had higher levels of these skills than major relapsers and non-abstainers. The particular focus of CBT for substance abuse on problem solving, coping strategies, communication and social skills may support younger people positively in abstaining and dealing with possible relapse.

Moderators

Whether certain population characteristics moderate CBT outcomes for non-opiates remains largely unknown (Morgenstern & McKay, 2007). In a study including 13 to 18 year olds Kaminer, Burleson and Goldberger (1998) found that only older males in the CBT group had a significant reduction in drug use in comparison to the psychoeducational therapy group. This could indicate that CBT is more appropriate forthe older males in the study (i.e., 16 to 18 year olds). Alternatively, the group delivery aspect may provide an additional explanation. Study findings suggest that group CBT has a greater effect in reducing drug use than individual CBT (Waldron, Slesnick, Brody, Peterson & Turner, 2011; Liddle, Rowe, Dakof, Ungaro, & Henderson, 2004). The group aspect may be a more conducive and realistic setting for practicing new skills and strategies with peers in the same situation. The group environment may also contribute to the support and promotion of cognitive and behavioural change among participants (Waldron & Kaminer, 2004).

Finally, the clients' motivation also plays an important role, as the more motivated the client is to change, the better the engagement, attendance and outcome of the therapy should be (Waldron & Turner, 2008), although this finding seems to apply to all drug treatment therapies. The duration of therapy may also moderate treatment outcomes and several studies found that shorter CBT interventions were more than, or just as effective as longer durations (Dennis et al., 2004; Kaminer, 2008).

3.1.1 Types of studies

The study designs included in the review will be: ○

RCT - randomized controlled trials

The rationale for including NRCTs is as follows:

The aim of this review is to be as comprehensive as possible. There may be information that is contained in NRCTs that may be of relevance to this review that are not captured in RCTs.

3.1.2 Types of participants

The population to be included in this reviewis young people aged 13-21years, enrolled in a CBT outpatient drug treatment for non-opioid drug use (e.g., cannabis, amphetamine, ecstasy or cocaine).

Definitions of young people, and the age in which a person is considered a young person and may be entitled special services, such as drug treatment varies internationally (United Nations, 2011). Age group distinctions for young people are unclear as the boundaries are fluid and culturally specific (Weller, 2006). Furthermore young people start experimenting with illegal drugs at different ages in different countries (Hibell et al., 2009). Patterns of young people's independence from parents and independent living patterns likewise vary internationally. In order to capture international differences we have set the age range from 13 to 21 (Hibell et al., 2009; United National, 2011; SAMHAS 2010; Danish Youth Council, 2011).A study with age groups well beyond the 13 to 21 age threshold, for example a study with 13 to 65 year olds will only be included if they report findings by age group for the intervention and control group.

No universal international consensus exists concerning what categories to use when classifying drug users, and different assessment tools and ways of classifying the severity of drug use are applied in different research studies (American Psychiatric Association, 2000; WHO, 2011; Nordegren, 2002). We include participants regardless of formal drug use diagnosis. The main criterion for inclusion is that the young person is enrolled in treatment for drug use (i.e., intervention or comparison condition). Referral to and enrolment in treatment requires a level of drug use, such that the young person, his/her parent or significant other, or a representative of a statutory authority found it necessary to solicit or require treatment. We therefore define the population as young peoplereferred to or in treatment for using non-opioiddrugs.

The focus of this review is on non-opioid use to avoid duplication of effect, as psychosocial interventions for the treatment of youth opioid use have been evaluated in Cochrane reviews (Amato et al., 2011; Minozzi et al., 2010). We will include participants with poly-drug use as long as the majority of drug users in the study are non-opioid users. Study populations with severe mental illnesses (e.g., schizophrenia, psychoticillness) will be excluded. We expect that some study populations may include young people with ‘common’ non-severe comorbid conditions (e.g.,behavioral, emotional, mental health issues)(Hawkins, 2009). These studies will not be excluded as long as the CBT intervention's focus is on treating drug use ⁹ . A study will be excluded if the primary intervention focus is to treat the comorbid condition (e.g., depression) in young people who also use drugs.

3.1.3 Types of interventions

The review will include outpatient CBT interventions(as defined in section 1.2,Description of the intervention) of any duration delivered to young people individually or in groups (e.g., peers or families), described by the authors as CBT or judged by the review authors to represent CBT. We will only include studies with CBT interventions specifically directed at treating ‘young people‘ for non-opioid drug use.

The intervention must be an outpatient intervention that does not include overnight stays in a hospital or other treatment facility. The CBT intervention can take place in the home, at community centres, in a therapist's office or at outpatient facilities, and can be delivered to individuals, groups, families and a combination of these

CBT interventions conducted by non-professionals (e.g., lay volunteers) will be excluded. Interventions in restrictive environments, such as prisons or other locked institutions (e.g., detention centres. institutions for sentence-serving juvenile delinquents) will be excluded. Interventions focusing exclusively on treating mental disorders will also be excluded

Studies where CBT is delivered in combination with add-on components (such as motivationalinterviewing) will be included as long as CBT is the primary intervention.

Eligible control and comparisons will include no intervention, waitlist controls and alternative interventions, as we are interested in both absolute and relative effects. Due to ethical considerations and nature of the problem (i.e., young peoples' drug use) the likelihood of a no-treatment control group is small. We expect that the most frequent comparison will be alternative interventions (Lipsey, Tanner-Smith, & Wilson, 2010). Alternative interventions as the comparison will complicate synthesis possibilities and analyses. We will pay careful attention to the types of comparison as is reflected in the Data Synthesis section (3.4).

3.1.4 Types of outcomes

3.2.1 Electronic searches

Relevant studies will be identified through electronic searches of the following bibliographic databases and government policy databanks. No language or date restrictions will be applied to the searches.

MEDLINE

EMBASE

CINAHL

Web of Science

SocIndex

PsycINFO

Cochrane Controlled Trial Register (CENTRAL)

Bibliotek.dk

LIBRIS

BIBSYS

Social Care Online

ERIC

SweMed+

Criminal Justice Abstracts

Bibliography of Nordic Criminology (up to summer 2008)

3.2.2 Search terms

An example of the search strategy for MEDLINE on the OVID platform is listed below. The strategy will be modified for the different databases. We will report full details of the modifications used for other databases in the completed review.

3.2.3 Searching other resources

The review authors will check the reference lists of other relevant reviews and included primary studies to identify new leads. Citation searching in the Web of Science will also be considered. In addition, we will contact international experts to identify unpublished and on-going studies.

3.2.4 Grey literature

We will use Google and Google Scholar search engines and the advanced search options to search the web to identify potential unpublished and/or studies in progress. We will check the first 150 hits.OpenSIGLE( http://opensigle.inistfr/ ) and OpenGrey (http://www.opengrey.eu/) will also be used to search for European grey literature. Sites such as NCJRS: National Criminal Justice Reference Service will be searched. Copies of relevant documents will be storedand we will record the exact URL and date of access.

In addition we will search the following websites:

National Institute on Drug Abuse (NIDA)

http://www.nida.nih.gov/nidahome.htm

The European Monitoring Centre for Drugs and Drug Addiction (EMCDDA)

http://www.emcdda.europa.eu/index.cfm

Substance abuse and Mental Health Services administration (SAMHSA)

http://www.samhsa.gov/

3.2.5 Hand searching journals

The following journals that we consider most likely to include relevant primary studies will also be hand searched from 2012: •

Addiction

3.3.1 Selection of studies

Two members of the review team (AK & SF ¹⁰ ) will independently screen titles and abstracts in order to exclude studies that are clearly irrelevant under the supervision of the first author (KK). Studies considered eligible by at least one of the review authors will be retrieved in full text. The full texts will then be screened by two members of the review team to determine study eligibility based on the inclusion criteria. Any disagreements about eligibility will be resolved by a third review author (TF). We will check for multiple publications of studies (i.e., whether several studies are based on the same data source).

Reasons for exclusion will be documented for each study that is retrieved in full text. The study inclusion coding sheet will be piloted and adjusted if required by the review authors (see Appendix 6.1). The overall search and screening process will be illustrated in a flow-diagram.

3.3.2 Data extraction and management

Two review authors (KK & AK) will independently code and extract data from the included studies. A data extraction sheet will be piloted on several studies and revised as necessary (see Appendix 6.2). Extracted data will be stored electronically. Any disagreements will be resolved by consulting a third reviewer with extensive content and methods expertise (TF). Analysis will be conducted in RevMan5 and/or STATA. Data and information will be extracted on: characteristics of participants (e.g., age, gender, drug use severity and history), intervention characteristics and control conditions, research design, sample size, outcomes, and results.

3.3.3 Assessment of risk of bias in included studies

We will assess the methodological quality of studies using a risk of bias model developed by Prof. Barnaby Reeves in association with the Cochrane Non-Randomised Studies Methods Group (Reeves, Deeks, Higgins & Wells, 2011). ¹¹ This model, an unpublished extension of the existing Cochrane Collaboration's risk of bias tool (Higgins & Green, 2008), covers risk of bias in RCTs, quasi-randomised trials and non-randomised studies that have a well-defined control group. The extended model is organised and follows the same steps as the existing Risk of Bias model according to the Cochrane Hand book, chapter 8 (Higgins and Green, 2008). The extension to the model is explained thus: 1)

The existing Cochrane risk of bias tool needs elaboration when assessing non-randomised studies because, for non-randomised studies, particular attention must be paid to selection bias / risk of confounding. The extended model therefore specifically incorporates a formalised and structured approach for the assessment of selection bias in non-randomised studies ¹² by adding an explicit item about confounding (Reeves, Deeks, Higgins & Wells, 2011). It is based on a list of confounders considered important and defined in the protocol for the review. The assessment of confounding is made using a worksheet where for each confounder it is marked whether the confounder was considered by the researchers, the precision with which it was measured, the imbalance between groups and the care with which adjustment was carried out (see 12.3 Appendix 3, p. 50). This assessment will inform the final risk of bias score for confounding.

The refined assessment is pertinent when thinking of data synthesis as it operationalizes the identification of studies (especially in relation to non-randomised studies) with a very high risk of bias. The refinement increases transparency in assessment judgments and provides justification for not including a study with a very high risk of bias in the meta-analysis.

Risk of bias judgment items and assessment

The risk of bias model used in this review is based on 9 items (for guidelines and coding sheets see Appendices 3 and 4).

The 9 items refer to •

sequence generation (Judged on a low/high risk/unclear scale - NRCTs will automatically have high risk of bias)

The assessment will be based on pre-specified questions (see Appendix 3).“Yes” indicates a low risk, “No” indicates a high risk of bias, “and ”Unclear“ indicates an unclear or unknown risk of bias. In the 5 point scale 1 corresponds to No/Low risk of bias (e.g., 1 = a high quality RCT) and 5 corresponds to Yes/High risk of bias (e.g., 5= too risky, too much bias, e.g., a poor quality study). A judgment of five on any item assessedtranslates to a risk of bias so high that the findings will not be considered in the data synthesis (because they are more likely to mislead than inform).

Confounding

An important part of the risk of bias assessment of non-randomised studies is how the studies deal with confounding factors. Selection bias is understood as systematic baseline differences between interventionvs.control (or comparison) groups that can compromise their comparability. We will code baseline equivalence of groups for the NRCTs.

For this review, the following confounding factors are considered to be the most relevant: age, gender, and drug history (including drug severity). If other confounders are considered by study investigators in the included studies they will be assessed in the same manner (Appendix 4).

We focus on the three confounders - age, gender and drug use history -as they are major predictors of drug use. Young people are in a transitional and developmental life phase, and their patterns of drug use are connected to age (Labouvie & White 2002; Kaminer 2008; Waldron & Kaminer 2004). Gender is also identified as a confounding factor, becausemales generally have higher drug use than females (Østergaard & Bastholm Andrade, 2011; McCabe et al., 2007). And finally, history of drug use and persistent patterns of use affect treatment outcomes (Labouvie & White, 2002; Kaminer, 2008).

Review authors (at least two, AK & KK) will independently assess the risk of bias for each included study as described in the previous sections. Disagreements will be solved by a third review author with content and statistical expertise (TF). We will report the risk of bias assessments in risk of bias tables for each included study in the completed review. These assessments will also informthe data synthesis.

3.3.4 Measures of treatment effect

3.3.5 Unit of analysis issues

We will take into account the unit of analysis of the studies to determine whether individuals were randomised in groups (i.e., cluster randomised trials), whether individuals may have undergone multiple interventions, and whether there were multiple treatment groups.

3.3.6 Dealing with missing data and incomplete data

The review authors will assess missing data and attrition rates in the included studies. In the case of missing data (e.g., valid Ns, means and standard deviations) the reviewers will contact primary study authors for missing data. The review authors will record attrition rates and (if possible) reasons for attrition from included studies.

Information on intention to treat analysis (ITT) will be recorded. We will run separate meta-analyses with studies that did not use ITT. We will perform sensitivity analysis to examine influences on effectsin studies using ITT analysis vs. studies not using ITT analysis.

3.3.7 Assessment of heterogeneity

Statistically significant heterogeneity among primary outcome studies will be assessed with Chi-squared (Q) test, tau-squared and I-squared (Higgins et al., 2003). A significant Q or tau-squared (P<.05)and I-squared of at least 50% will be considered as statistical heterogeneity.

3.3.8 Assessment of publication bias

Reporting bias refers to both publication bias and selective reporting of outcome data and results. Selective reporting will be dealt with in the risk of bias assessment and any concerns will be reported. We will use funnel plots for information about possible publication bias if we find sufficient studies (Higgins & Green, 2008). However, asymmetric funnel plots are not necessarily caused by publication bias (and publication bias does not necessarily cause asymmetry in a funnel plot). If asymmetry is present, we will consider and report possible reasons for this.

3.4.1 Subgroup analysis/moderator analysis and investigation of heterogeneity

We will investigate the following study-level covariates (if possible) with the aim of explaining observed heterogeneity: intervention characteristics (e.g., treatment duration, treatment intensity), participants' characteristics (e.g., gender, age, family composition, ethnicity, co-morbidity, and history of drug use) and comparison intervention characteristics.

If the number of included studies is sufficient (dependent on the spread of the study means of the covariates and study sizes, see Borenstein, Hedges, Higgins & Rothstein,2009 and Simmonds & Higgins, 2007), we will perform moderator analyses (meta-regression) to explore how observed variables are related to heterogeneity using a mixed model. Otherwise, single factor subgroup analysis will be performed.

3.4.2 Sensitivity analysis

Sensitivity analysis will be used to evaluate whether the pooled effect sizes are robust across study design and components of methodological quality. For methodological quality, we will consider sensitivity analysis for each major component of the risk of bias checklists.

To check for the possible influence on developer bias effect sizes we will run sensitivity analysis on studies conducted by program developers vs. studies conducted by independent researchers. Developer bias can occur in studies conducted by developers who may unconsciously influence the success of an intervention(Petrosino&Soydan, 2005; Eisner, 2009; Sherman & Strang, 2009).

We will also consider sensitivity analysis for program fidelity, i.e. compliance with program manual and requirements for therapist training.