- CBT
- treatment planning
A CBT treatment plan for generalized anxiety, session by session: what each phase contains and which manual it comes from
Wherever you practise, a CBT plan for generalized anxiety disorder (GAD) needs what templates rarely give: a dose and a source. England’s NICE guideline CG113 states both in one recommendation, 1.2.19: CBT for GAD should “be based on the treatment manuals used in the clinical trials of CBT for GAD” and “usually consist of 12 to 15 weekly sessions (fewer if the person recovers sooner; more if clinically required), each lasting 1 hour.” It is written for step 3 of NICE’s stepped-care model (marked functional impairment, or no adequate response to a low-intensity intervention). One of those manuals, from Michel Dugas’s group, spends 12 sessions like this: one on psychoeducation, three on uncertainty, one on the usefulness of worry, three on problem solving, three on imaginal exposure and one on relapse prevention. Applied relaxation is the alternative NICE offers at the same step and the same dose. What circulates is something else: technique lists with no session count and no source, and, in Spanish, a “10 sessions” figure that a 2008 guideline from Spain passed to the IMSS guideline in Mexico, where it still appears.
What circulates: techniques without a sequence
Carepatron’s Spanish-language CBT treatment plan page (dated 17 August 2025; consulted 8 October 2026) offers a downloadable sample, in English, that is a GAD plan: cognitive restructuring, deep breathing with progressive muscle relaxation, behavioral activation, and a goal to “Reduce overall anxiety levels by 50% as measured by GAD-7 scores within six months.” It names no session count and no manual. Nothing in it targets uncertainty, problem orientation or worry exposure, and its relaxation is never applied in anxiety-provoking situations, the step NICE says defines applied relaxation.
A 2023 thesis from the Universidad Técnica Particular de Loja in Ecuador, on a 70-year-old woman with GAD, shows the other version: six techniques, including «parada de pensamiento» (thought stopping) and «técnica de relajación progresiva de Jacobson» (Jacobson’s progressive relaxation), delivered in «once sesiones en un período de tres meses» (eleven sessions over three months). The pattern is the same: the plan is built by adding techniques, not by following a tested protocol.
Where “10 sessions” comes from: a worked example from Mexico
Mexico’s social-security institute, the IMSS, publishes a clinical practice guideline for anxiety disorders in adults (institutional guidance, not a binding norm): IMSS-392-10, published in September 2010 and updated in August 2011 (consulted October 2026). Page 26 reads:
«No hay evidencia de que la TCC aplicada durante más de 6 meses (10 sesiones) consiga una mayor efectividad a más largo plazo.»
«Para la mayor parte de los pacientes, la TCC se realizará en sesiones semanales de 1 a 2 horas y deberá complementarse en un plazo máximo de 4 meses desde el inicio.»
(No evidence that CBT for more than 6 months, 10 sessions, works better in the long term; for most patients, weekly sessions of 1 to 2 hours, completed within 4 months.)
The first sentence is copied word for word from Spain’s 2008 primary-care anxiety guideline (Agencia Laín Entralgo), which based it on a Scottish long-term follow-up (Durham and colleagues, 2005) and made it a grade A recommendation: «La TCC debe ser aplicada en unas 10 sesiones (unos 6 meses) como media» (CBT should be delivered in about 10 sessions, about 6 months, on average). The second sentence, which the IMSS attributes in part to “NICE, 2007”, is a near-literal translation of a recommendation CG113 now keeps only for panic disorder: “For most people, CBT should take the form of weekly sessions of 1 to 2 hours and should be completed within a maximum of 4 months of commencement” (1.3.16, [2004]).
Neither sentence is a factual error; both are dated. For GAD, NICE reviewed the trials and in 2011 wrote 1.2.19. Spain moved too: in November 2024 its guideline portal, GuíaSalud, published a new GAD guideline that «actualiza parcialmente» (partially updates) the 2008 one, and it carries three figures: its text asks for CBT of «al menos, 12 sesiones con periodicidad semanal» (at least 12 weekly sessions, p. 45), its good-practice recommendations speak of completing «8-12 sesiones» (p. 46), and its introduction still repeats «unas 10 sesiones durante aproximadamente 6 meses» (about 10 sessions over roughly 6 months, p. 31). We did not locate a later version of IMSS-392-10. Nor is NICE new: on 11 September 2026 it concluded that CG113 needs updating. Before you take a session count from a guideline, check its date and trace the sentence to its source. The full dose table, disorder by disorder, is in how many sessions of CBT.
“The treatment manuals used in the clinical trials”: which ones
1.2.19 names no manual. The 2011 full guideline pooled 21 trials of CBT built on different models: intolerance of uncertainty (Dugas and Ladouceur), Borkovec’s CBT, the cognitive therapy of Butler and of Arntz, Hoyer’s worry exposure, Wells’s metacognitive therapy and Roemer’s acceptance-based behaviour therapy. On average they ran “approximately 15 weekly sessions (range of 8 to 20 weeks)” of about 70 minutes. To cost the recommendation, NICE assumed “12 sessions and three booster sessions, lasting 1 hour each”: a practical way to read “12 to 15”.
The consequence: pick one manual and keep its order. A plan stitched from three models is none of the tested protocols, and 1.2.21 asks for supervision “to monitor fidelity to the treatment model”, in the singular. The Dugas protocol is used below because its sequence is published phase by phase, with the sessions each phase takes, in English and in Spanish.
The Dugas protocol, session by session
The trial by Dugas and colleagues (2010), one of those NICE reviewed, compared CBT with applied relaxation, both “administered over 12 weekly 1-hour sessions.” The phases and the sessions each takes are the trial’s; the between-session tasks follow Arturo Bados’s Spanish-language description of the protocol (University of Barcelona, 2017):
| Session | Phase | Content | Between-session task |
|---|---|---|---|
| 1 | Psychoeducation and worry awareness training | The GAD model; sorting worries about current problems from worries about hypothetical situations | Worry log three times a day: content, duration, anxiety (0–8), type |
| 2–4 | Uncertainty recognition and behavioral exposure | What the client does to reduce uncertainty: asking others, postponing, avoiding | Graded behavioral experiments |
| 5 | Re-evaluation of the usefulness of worry | What the client believes each specific worry is for | Personal examples of those beliefs |
| 6–8 | Problem-solving training | Worries about current problems, with the emphasis on problem orientation | A visible list of recurring problems |
| 9–11 | Imaginal exposure | Worries about hypothetical situations: the core fear, written as a scene | Daily exposure to the scene |
| 12 | Relapse prevention | Skills review; a setback versus a relapse | An action plan for after therapy |
Uncertainty is worked with experiments, not reassurance. Bados asks for «mínimo de uno a la semana, pero mejor tres o más» (at least one a week, better three or more), logged in five fields (what I’ll do, what I fear, what happened, how I coped, what I conclude) and starting small: «enviar un correo electrónico de poca importancia sin releerlo» (sending an unimportant email without rereading it). When the client is unsure how to classify a worry, «el terapeuta no proporciona tranquilización» (the therapist does not reassure). In 2022 the same group tested a version stripped down to this component, 12 weekly one-hour sessions, against a wait-list; it was not compared with the full treatments.
Each kind of worry has its own tool. Worries about current problems go to problem solving; hypothetical ones go to imaginal exposure, and according to Dugas and Robichaud (2007), as Bados reports, most clients need work on only one or two core fears. The rule the manual uses to end each exposure, once anxiety comes down, is the one examined in the guide to graded exposure.
What stays out: thought stopping. It appears in none of the six phases, and the imaginal exposure module opens by showing the client, for example with the white-bear experiment, that «intentar bloquear las preocupaciones puede, de hecho, facilitar su persistencia» (trying to block worries can in fact help them persist).
The ending is planned. The version of the protocol that Bados describes runs 14 to 16 sessions with follow-ups at 3, 6 and 12 months, and Spain’s 2024 guideline suggests ending by «distanciando las sesiones progresivamente» (spacing sessions out gradually). The three booster sessions NICE assumed when costing the treatment fit there; what to record at the end is in the guide to documenting termination.
When a client asks for a technique to stop the thoughts, this script holds the model without dismissing the request:
“I understand you want to switch those worries off. We are going to do almost the opposite, because trying to block worries can help keep them going. We’ll sort your worries into two kinds: the ones about a problem that already exists, which we’ll solve step by step, and the ones about something that might happen, which we’ll face on paper instead of dodging. I won’t promise you nothing bad will happen; I will help you live with not knowing.”
The same dose, a different treatment: applied relaxation
1.2.18 is explicit: if the person chooses a high-intensity psychological intervention, “offer either CBT or applied relaxation”, and 1.2.20 gives applied relaxation the same dose and the same requirement to follow the trial manuals. Lars-Göran Öst developed it in the 1980s, and the full guideline describes five steps: progressive muscle relaxation, release-only relaxation, cue-controlled relaxation, rapid relaxation and application in anxiety-provoking situations. The last “is critical and distinguishes applied relaxation from other forms of relaxation training and practice that do not have the applied component.” In Dugas 2010 it took 1 session of psychoeducation, 4 of tension-release (from 16 muscle groups down to 4), 3 of relaxation without tensing (by recall and by counting), 3 of applying it in daily life along a graded hierarchy and 1 of relapse prevention.
Öst and Breitholtz (2000) treated 36 patients for 12 weekly sessions and found “no difference between AR and CT on any measure.” In Dugas 2010, CBT was marginally superior at post-treatment; over follow-up the two were equivalent, “but only CBT led to continued improvement.” And the full guideline adds that “clinicians should note the less robust evidence base.” Applied relaxation is a legitimate choice, especially when the client prefers it; progressive muscle relaxation without the applied step, as in both examples above, is not the intervention NICE recommends.
Inside a session
Session structure comes from cognitive therapy generally. Here is the one the Cognitive Therapy Scale rating manual by Young and Beck describes, adapted to this plan:
- Measure and check-in: the measure due (the GAD-7, every two weeks), completed before the session, and a brief account of the week.
- Agenda: “When done properly, the agenda can usually be set within five minutes”, with one or two target problems, rarely more.
- Homework review, near the start: if it is not reviewed, the manual warns, the client may come to believe there was no need to do it carefully.
- The phase’s component, with its rationale and a summary of conclusions.
- New homework, with its rationale and in a fixed format: Dugas 2010 used “standard forms for the between-session exercises.”
- Summary and feedback.
What to measure and when to review
1.2.21 asks practitioners to “use routine outcome measures and ensure that the person with GAD is involved in reviewing the efficacy of the treatment”, but names no instrument. A workable scheme:
- GAD-7 at baseline, then every two weeks. What its two-week window means for cadence is in the GAD-7 guide; how much change counts as reliable is in measuring therapy progress.
- PSWQ (Penn State Worry Questionnaire) at baseline, mid-point and end. It measures worry in more detail than the GAD-7, but as a «rasgo de preocupabilidad» (a trait-like tendency to worry), in Bados’s words, so weekly use makes little sense. Dugas 2010 used the 16-item version; in Mexico, Padrós-Blázquez and colleagues (2018) studied it in 2,267 participants and recommend the short 11- and 8-item versions with directly worded items.
- The phase’s own logs: experiments, exposures, problems worked.
Two formal reviews anchor the plan, at sessions 6 and 12; Spain’s 2024 guideline suggests reviewing response «después de 4-6 semanas» (after 4 to 6 weeks) of weekly CBT. At each, decide with the client among the outcomes 1.2.19 itself allows: fewer sessions if they recover sooner, the planned number, or more if clinically required.
If at a review the GAD-7 has not moved or has worsened, extending the same protocol is not the only option. When GAD has not responded to a full course of a high-intensity psychological intervention, CG113 says to offer drug treatment (1.2.34), which for you is a medical referral, not a decision of your own, and to consider referral to step 4, specialist care, when severe anxiety with marked functional impairment comes with a risk of self-harm or suicide, significant comorbidity, self-neglect or an inadequate response to step 3 (1.2.37). If risk of self-harm or suicide appears, assess and document it that day, without waiting for the review; verify your local crisis line (in Mexico, Línea de la Vida, 800 911 2000, and 911).
Older adults, and the manuals in Spanish
For older adults, the IMSS has a newer guideline, IMSS-499-19 (2019). It recommends CBT to reduce GAD-related anxiety in older adults and, in a separate recommendation, makes it conditional: «si puede aprender y practicar la intervención de manera adecuada» (if the person can learn and practise the intervention adequately). It also cites a meta-analysis in which «el número medio de sesiones de TCC fue 12» (the mean number of CBT sessions was 12).
For bilingual clinicians, Bados’s 2017 review, open access from the University of Barcelona, describes the Borkovec, Barlow, Dugas and Wells treatments in Spanish and refers to Dugas and Robichaud (2007) for the full Dugas protocol.
Writing the plan down
An example for the record, not for the client:
Presenting problem: excessive, hard-to-control worry about her children’s health, work and money; muscle tension and initial insomnia; checks every decision with her partner. GAD-7: 15 (severe). PSWQ-11 baseline recorded.
Model and dose: intolerance-of-uncertainty CBT (Dugas), 12 weekly 60-minute sessions plus up to 3 spaced booster sessions, within the 12 to 15 range of NICE CG113, 1.2.19.
Goal 1: GAD-7 below 10 by session 12, administered every two weeks. Goal 2: at least 1 logged behavioral experiment a week (target: 3) between sessions 2 and 4. Goal 3: decide purchases and schedules without checking in 4 of 5 logged occasions within 8 weeks.
Phases: as in the table. Reviews: sessions 6 and 12.
The goals follow the rules in a treatment plan with SMART goals. If you use Gesell, the draft starts from your intake: goals and, if you ask for the full plan, a session arc with a review roughly every 6 to 8 sessions and a closing session; techniques and homework are suggested as you prepare each session, and the GAD-7 can be re-sent every N days. It is a draft to check against the manual you chose, not a manual.
References
- NICE — Generalised anxiety disorder and panic disorder in adults: management (CG113), 2011, last updated 2020: recommendations
- NCCMH / NICE — Generalised anxiety disorder in adults: the NICE guideline on management in primary, secondary and community care (full guideline, 2011, PDF)
- IMSS — Diagnóstico y tratamiento de los trastornos de ansiedad en el adulto, GPC IMSS-392-10, 2010, updated 2011 (PDF)
- IMSS — Diagnóstico y tratamiento del trastorno de ansiedad generalizada en la persona mayor, GPC-IMSS-499-19, 2019 (PDF)
- GuíaSalud — Guía de Práctica Clínica para el Manejo de Pacientes con Trastornos de Ansiedad en Atención Primaria, Agencia Laín Entralgo, 2008 (PDF)
- GuíaSalud — Guía de Práctica Clínica para el Tratamiento del Trastorno de Ansiedad Generalizada en Atención Primaria, 2024
- Dugas et al. — A randomized clinical trial of cognitive-behavioral therapy and applied relaxation for adults with generalized anxiety disorder (Behav Ther, 2010)
- Bados — Trastorno de ansiedad generalizada: naturaleza, evaluación y tratamiento (University of Barcelona, 2017)
From case to treatment plan
Gesell drafts the plan from your intake interview: SMART goals, the focus of each session, and review sessions. You review and edit it.
Try gesell.aiHave a question about your practice?
How to document a case, which format suits you, or how Gesell fits your practice: write to us and we will answer.
About the author
Gesell Team
Clinical and product content written by the gesell.ai team together with certified clinical psychologists.