Background: Psychotherapy is an effective treatment for generalized anxiety disorder in comparison to no-treatment controls. Instead of creating more and more new overall treatment-packets within a medical meta-model, a complementary approach to investigate clinical research designs may lie into the understanding of already effective psychotherapies. Treatment manuals and protocols allow a relatively high degree of freedom of therapists' behaviors on how to implement the overall treatment manuals. There is a lack of systematical knowledge of how therapists have to customize these overall protocols. The present design experimentally examines 3 types of conducting a 15 session time-limited cognitive-behavioral therapy (CBT) protocol and its relation to the therapists' protocol adherence and treatment efficacy.

Methods/design: This trial investigates 3 different types of how to customize a well-introduced CBT-protocol using dyadic peer tutoring methodology (primings). The individuals with GAD are randomly assigned to 3 priming conditions (resource priming vs. supportive resource priming vs. adherence priming). Participants' treatment allocation is performed randomly, therapist's assignment to the peer tutoring partner and the priming condition is based on mutual agreement. Treatment outcomes are assessed at following levels: Observer based in-session outcomes, post-session outcomes from session 1 to 15, treatment outcome at post assessment and at 6-months follow-up assessments.

Aims of the trial. This trial investigates 3 different types of how to customize a well-introduced CBT-protocol (Zinbarg, Craske, & Barlow, 2006) using dyadic peer tutoring methodology (primings). The participants are randomly assigned to 3 priming conditions: (a) Resource priming (Flückiger, Wüsten et al., 2010; Flückiger & Grosse Holtforth, 2008a); (b) Supportive resource priming and (c) Adherence priming.

The main research questions are:

1. In-session outcomes. Using the videotapes, observer-based videoanalyses will be conducted: (a) Do the resource priming conditions show comparable observer-based therapist's adherence in comparison to the adherence priming condition (Zinbarg et al., 2007)? (b) Do the resource priming conditions show more resource activating micro interventions than the adherence priming condition (ROMA; Flückiger & Grosse Holtforth, 2008a/b)? (c) Are the in-session processes predictors and mediators of session and therapy outcomes (Flückiger et al., 2009; Flückiger et al., 2013)?
2. Post-session outcomes. Are there differences in the post-session outcomes of the 3 priming conditions? Furthermore: Are the post-session outcomes along with symptom change predictors and mediators of therapy outcome (Flückiger et al., 2009; Flückiger et al., 2012)?
3. Treatment outcomes. Do show the resource priming conditions comparable efficacy on GAD outcomes, general outcomes and dropout rates in comparison to the adherence priming condition?

Methods/Design. This study is a randomized controlled trial with three active treatment arms. This trial is conducted at the Swiss psychotherapy outpatient clinic from the Department of Psychology at the University of Zürich. The treatments are based on the same treatment protocol (Zinbarg et al., 2006). Overall, a 3 x 4 design with one between-subject factor (resource priming vs. supportive resource priming vs. problem priming) and one within-subject factor (time: pre, mid (session 6), post (session 14), 6-months follow-up).

Participants. Inclusion/Exclusion Criteria. Participants will be included in the study if they: (a) are 18 years or older; (b) agree to the informed consent, (c) have sufficient knowledge of German; and (d) fulfill the diagnostic criteria of GAD DSM-IV criteria. Participants will be excluded if they have: (a) a score of 2 or higher on the suicide item of the BDI and/or with active suicidal plans according to the diagnostic screening interview, (b) a current medication of psychotic or bipolar disorder, (c) a current treatment by a professional psychotherapist. Further, prescribed medications for anxiety or depressive disorders do not lead to exclusion if the dosage has remained constant for at least one month. The presence of a comorbidity does not result in exclusion if GAD is in the foreground according to the severity rating of the Diagnostic Interview for DSM-diagnoses.

Recruitment. Participants will be recruited by means of advertisements in newspapers and through internet forums. The high-circulation newspapers are delivered for free in the Swiss public transport systems. Interested individuals in participation will contact the study office via SMS, e-mail or phone. Screened positive patients are invited for an intake assessment to verify the inclusion and exclusion criteria based on a standardized interview. Participants that are not screened positive are informed about more appropriate treatments via phone call or if requested by a further face-to-face contact.

Randomization and treatment allocation. After meeting the inclusion criteria, patients are randomized to the 3 conditions (resource priming, supportive resource priming, adherence priming). Treatment allocation is performed randomly by online application of a full randomization. In this way, we aim to ensure that trial arms are balanced with respect to the patients' baseline characteristics. Randomization procedure is conducted by two independent research assistants. Because all patients are treated by the same CBT-manual, patients are blinded for the treatment allocation and are not informed about the randomization procedure.

GAD treatment protocol. CBT-manual (Zinbarg et al., 2006): Traditional cognitive-behavioral therapy (CBT) for GAD typically consists of psycho education of generalized anxiety disorder, relaxation training (RT), cognitive restructuring (CR) and with some in-vivo situational exposure for patients with overt behavioral avoidance (e.g., Barlow, Rapee, & Brown, 1992; Borkovec & Costello, 1993). Furthermore, imagery exposure as a GAD-specific form of in-sensu exposition will be applied to reduce experiential avoidance based on a well-introduced CBT-manual (Zinbarg et al., 2006). The manualized therapy follows an usual time-limited treatment format of 14 sessions to 50 minutes and a booster session after 6-months (total 15 sessions).

Tandem peer tutoring (priming).

To investigate various types of how to conduct a standardized CBT-protocol, all therapists are tutored in peer dyads (tandem peer tutoring). Immediate before sessions 1 to 5, the therapists are required to contact the tandem-partner face-to-face or via self-phone to deliberate the forthcoming session by a 5 to 10 minute brief communication (primings; for a comparable procedure see Flückiger & Grosse Holtforth, 2008). The tandem peer tutoring are grouped in three conditions:

1. Resource priming: Immediately before sessions 1 to 5, therapists have a five-minute conversation about how to implement strengths-based micro-interventions in the forthcoming session. Strengths-based micro-interventions addresses therapists explicit focus on patients' preexisting strengths and abilities, subtle changes and improvements during therapy (potential recources) as well as motivational preparedness, readiness and goals (motivational resources; Grawe, 2006; Flückiger et al., 2010).
2. Supportive resource priming: The supportive resource priming condition has the very same protocol as the resource priming condition (5 brief tandem peer tutorings). The only difference in the procedure is that the therapists are allowed to integrate a helpful significant person of the patients (such as the partners or the best friends) around session 1 and 7 to encourage and support the patient to realize their treatment plans (active integration of interpersonal resources). However, the integration of a significant other person does not touch the CBT-treatment protocol.
3. Adherence priming: Immediately before sessions 1 to 5, therapists have a five-minute conversation about how to implement the disorderspecific interventions that are described in the treatment protocol. These communications are focused on therapists' understanding of patients GAD and the related comorbidities and how these issues can be addressed in the prescriptive treatment protocol.

Therapists. Twelve advanced trainees with at least 2 years post-graduate training are recruited from local (CBT)-training centers. The majority of the therapists have experience as study therapists in a prior randomized controlled trial. All the therapists participate in an initial 16-hour workshop presented by the developer of the treatment manual (Zinbarg et al., 2006). In addition to the peer-tutorings (primings), the therapists have regular supervision in small groups on a 14-day basis. The supervision is conducted in mixed groups over the 3 priming conditions. All the supervisors also participate in the initial 16-hour workshop. To respect and coordinate the therapists' preferences (e.g. preferences in working days and time schedules) they were assigned in a joint face-to-face session at study start. All the therapists gave verbal and written consent to the selected peer-tutoring partner and priming condition.

Assessments. At intake, GAD-diagnosis and its core symptomatology are identified according to the structured interview section for GAD (DIPS; Margraf, Schneider, & Ehlers, 1994). Furthermore, GAD-criteria are assessed by self-reports. The individual worries are identified using the Penn State Worry Questionnaire (PSWQ) and the Worry Domain Questionnaire (WDQ). Mental disorders on Axis I are assessed using face-to-face diagnostic interviews for Axis I (Diagnostisches Interview für Psychische Störungen für DSM-IV, SCID-I) and for Axis II (Strukturiertes Klinisches Interview für DSM-IV, Achse-II, Persönlichkeitsstörungen, SKID-II; Wittchen, Zaudig, & Fydrich, 1997).

PSWQ, Beck Anxiety Inventory (BAI) and the State-Trait Anxiety (STAI) are used as GAD-outcomes. Premature termination from the trial, the interpersonal problems and strengths inventories, the behavioral inhibition and activation scale, as well as the resource potential questionnaire are used as general outcomes. The outcome measures are taken at intake, directly after session 6 (end of primings), session 14 (post-assessment), and at 6- month follow-up.

The following process measures are examined: (a) Post-session reports (Flückiger et al., 2010), working alliance (Munder et al., 2009) and symptom change (BAI) are conducted from session 1 to 15 and (b) In-session processes: Using the recorded videotapes, adherence ratings (Zinbarg et al., 2007), the Resource-Oriented Microprocess Analysis (ROMA-P/T, Flückiger & Grosse Holtforth, 2008) and the level of explication scale (Sachse, 1992) are conducted at session 2, 5 and 8.

Study from the participants view. After contaciting the study administration, the participant's information contained a precise description of the inclusion and exclusion criteria, the structure of the treatment manual as well as the data collection procedure including the video recordings. Confidentiality of the collected data is confirmed. Volontarines of participation is emphasized by the opportunity to terminate the treatment at every timepoint during treatment and the cancelation of all conducted data. Furthermore, an insurance for possible negative outcomes of clinical trials is contracted.

Statistical Analysis. In order to handle the hierarchical data structures (sessions at Level 1 nested within patients at Level 2 and therapists at Level 3), hierarchical linear model (HLM) with time as a within-groups factor and treatment condition as a between-groups factor will be used for the main research questions. The analyses will be conducted on the intention-to-treat sample as well as on the completer sample. To investigate if possible patients' and therapists' effects have an impact to the therapy outcomes, patients' and therapists' pretreatment characteristics are investigated as outcome predictors at Level 2 and 3. Clinical significance will be defined as (a) a statistically reliable change according to the Reliable Change Index (RCI), and (b) being within two standard deviations of a nonclinical group at post-assessment or follow-up (Jacobson & Truax, 1991). All the analyses will be run by HLM and R packet.

Sample size. Based on a power analysis with G*Power the optimal sample size with an Alpha-error of 5 %, and a Beta-error of 80% as well as a correlation coefficient of r = .30 of the repeated assessments are 60 participants, i.e. 20 participants in each priming condition.

Inclusion Criteria:

• are 18 years or older
• agree to the informed consent,
• have sufficient knowledge of German
• fulfill the diagnostic criteria of GAD DSM-IV criteria

Exclusion Criteria:

• a score of 2 or higher on the suicide item of the BDI and/or with active suicidal plans according to the diagnostic screening interview,
• a current medication of psychotic or bipolar disorder,
• a current treatment by a professional psychotherapist. (Further, prescribed medications for anxiety or depressive disorders do not lead to exclusion if the dosage has remained constant for at least one month. The presence of a comorbidity does not result in exclusion if GAD is in the foreground according to the severity rating of the Diagnostic Interview for DSM-diagnoses.)