How many sessions of CBT? There is no single number: each guideline sets its own, disorder by disorder

There is no session count for “CBT”. There is one per guideline, per disorder and sometimes per severity, often with an adverb attached. England’s depression guideline, NICE NG222, puts individual CBT at “usually” 8 sessions for less severe depression and 16 for more severe. Its anxiety guideline, CG113, gives usually 12 to 15 weekly one-hour sessions for generalized anxiety with marked functional impairment or that has not improved after low-intensity treatment; for moderate to severe panic it counts hours, not sessions: 7 to 14 in total. The PTSD guideline gives typically 8 to 12 sessions; the OCD guideline counts therapist hours per patient: up to 10, or more than 10. For chronic insomnia, a 2026 Spanish practical guide and the American Academy of Sleep Medicine describe a usual course of 4 to 8 sessions; for depression, Mexico’s IMSS guideline gives 6 to 8 or 16 to 20 depending on severity. What circulates, in English and in Spanish, is something else: one range for all of CBT. Every figure below is dated, so you can set it beside the guidance where you practice.

What circulates: one range for all of CBT

In English, Psychology Today’s CBT explainer, last modified in May 2026 according to its page metadata, puts CBT at “generally 5 to 20 sessions, though there is no set time frame.” The hedge is honest, but the range names no disorder, so it cannot say whether your client needs 8 or 16.

In Spanish, the range is older and no more precise. A 2007 piece from the Argentine newspaper La Nación, still published by IntraMed, says “Un tratamiento de terapia cognitivo conductual consta de doce a quince sesiones y no dura más de 6 meses” (a course of CBT consists of twelve to fifteen sessions and lasts no more than six months), citing no study or guideline. The version circulating in Mexico today, on Doctoralia Mexico’s CBT page (undated, consulted 7 October 2026), widens it to “entre 5 y 20 sesiones” (between 5 and 20 sessions). Neither names a disorder.

The problem is not the range but that there is only one. The 2007 figure of 12 to 15 matches the dose NICE would write in 2011 for a single disorder at a single step: generalized anxiety at step 3. It falls short for more severe depression (16 in NG222) and overshoots for chronic insomnia (a usual ceiling of 8). A range without a disorder is not a data point: it is an anchor.

The table: each guideline, its disorder, its dose

NICE and IMSS rows are for adults; the insomnia guide indicates CBT-I “independientemente de la edad” (regardless of age). Every row carries its year; NICE’s own year tag is in square brackets where NICE gives one.

Sessions, hours and months are not the same unit

Panic is the clearest case: CG113 sets no session count for it. It asks for CBT “in the optimal range of duration (7 to 14 hours in total)” (1.3.15) and, “For most people”, weekly sessions of 1 to 2 hours completed “within a maximum of 4 months of commencement” (1.3.16). Fourteen hours can be fourteen one-hour sessions or seven two-hour ones: the guideline fixes the total, not the count.

In social anxiety, counting sessions misleads. Clark and Wells allows up to 14 sessions of 90 minutes: 21 hours at most. Heimberg has 16 sessions, 15 of 60 minutes and one of 90 for exposure: 16.5 hours. At full length, the protocol with more sessions is the one with fewer hours.

In the current guideline, CG31, OCD is not measured in sessions at all: it defines intensity as “the hours of therapist input per patient”. Low intensity is up to 10; more intensive CBT, more than 10, has no ceiling.

“Usually”, “typically”, “habitualmente”: the adverb is part of the dose

NG222 puts “usually” in front of every session count in its treatment tables, and in the individual CBT and behavioral activation rows it adds that “additional sessions may be needed” for comorbid mental or physical health problems, complex social needs or residual symptoms. No NG222 row or recommendation mentions fewer.

The guideline that provides for both directions is CG113, for generalized anxiety: “fewer if the person recovers sooner; more if clinically required”. NG116 names only one: “more if clinically indicated, for example if they have experienced multiple traumas”. The Spanish insomnia guide says CBT-I “se aplica habitualmente en 4 a 8 sesiones” (is usually delivered in 4 to 8 sessions) and that its application “no responde a un protocolo rígido por sesiones, sino a una secuencia estructurada y flexible” (does not follow a rigid session-by-session protocol but a structured, flexible sequence); the 2021 American Academy of Sleep Medicine guideline describes the same course: “typically 4–8 sessions”.

Where do the numbers come from? NG222 does not say where each table figure comes from, but 1.4.5 asks you to “Use psychological and psychosocial treatment manuals to guide the form, duration and ending of interventions”. CG113 is more explicit: 1.2.19 bases CBT for generalized anxiety on “the treatment manuals used in the clinical trials of CBT for GAD”. Change the protocol and the duration may change with it.

NICE and the IMSS: what each guideline is, and what it is not

NICE, the National Institute for Health and Care Excellence, writes guidelines for health and social care in England. They are in this table not as rules for your practice but because they write each dose down with a recommendation number and a year.

The year matters. CG113’s recommendations on how to deliver CBT for panic (1.3.13 to 1.3.19), duration included, carry the tag [2004]. On reviewing the guideline on 11 September 2026, NICE concluded “that an update is needed” but that it cannot begin the work for now. NG222 was reviewed on 30 January 2026 with a different conclusion: “We reviewed this guideline and will not update the recommendations”. And CG31 is being updated: the September 2026 draft, open for consultation until 21 October, proposes “Usually 12 to 14 weekly sessions, but more sessions may be needed” of individual CBT for adults with OCD, with publication expected on 16 February 2027. Until then, CG31 applies.

Mexico shows what to check in any national guideline. IMSS clinical practice guidelines (CPGs) are institutional guidance from the Mexican Social Security Institute: a clinical reference, not a binding Norma Oficial Mexicana. The most recent IMSS-161-09 we could find is the update of 2 December 2015, and it cites both session counts to the 2014 depression guideline of Spain’s national health system.

Wherever you practice, ask your own guideline three questions before quoting it: when the recommendation dates from, whether it has been reviewed since, and where its figure came from.

From the table to your estimate

1. Diagnosis and severity first. NG222 does not use the original PHQ-9 bands: it defines less severe depression as scoring “less than 16 on the PHQ-9 scale”. A PHQ-9 of 15 is “moderately severe” in the original bands and, for NG222, less severe depression.
2. Check where to start. For less severe depression every option in table 1 is first-line, but NG222 asks you to “consider the least intrusive and least resource intensive treatment first (guided self-help)”, which usually consists of “6 to 8 structured regular sessions”.
3. Adjust with the reasons the guidelines name — comorbidity, complex social needs, residual symptoms, multiple traumas, early recovery — and if you use another, write it down.
4. State the unit and what the guideline leaves open. “14 sessions” and “14 hours” coincide only if your sessions last an hour. NG222 says only “regular sessions”: no minutes, no frequency. If you set weekly 50-minute sessions, record that as your decision.
5. Take it to the plan. The estimated number goes into the treatment plan, next to the date of the first review.

How to say it in the consent conversation

The informed consent guide lists “estimated session length and frequency” and the expected number of sessions among the essentials, and APA Standard 10.01 asks you to inform clients about the “anticipated course of therapy”; neither gives you the figure. This script, for generalized anxiety, assumes NICE’s step 3 (1.2.17): marked functional impairment, or not enough response to the low-intensity options of step 2.

“What you’re describing fits what the guidelines call generalized anxiety, and it is already weighing on your daily life. The reference guideline I work from, England’s NICE guideline, puts this treatment at usually 12 to 15 weekly one-hour sessions: fewer if you improve sooner, more if needed. I suggest we plan for 12. Each week you’ll fill in a short questionnaire; at session 6 we’ll go over the numbers and the plan together, and before the last session we’ll talk about how to keep what you’ve gained. If money or time limits how many sessions you can have, tell me today and we’ll adjust.”

If the client can only afford a certain number of sessions, APA Standard 6.04(d) asks you to discuss it as early as feasible; the table tells you whether that number reaches any dose with evidence behind it.

A scheduled review, not an open-ended course

The guidelines set no cadence for reviewing the plan, but they do set checkpoints. NG222 asks you to review how the treatment is working “between 2 and 4 weeks after starting treatment” and to consider sessional outcome measures, “for example PHQ-9” (1.4.3). If after 4 to 6 weeks of psychological therapy the depression “has not responded at all”, 1.9.1 does not order a change of treatment: it asks you to discuss with the person what might explain it, missed sessions included.

That requires measuring every session: without data, CG113’s “fewer if the person recovers sooner” is an impression.

The ending is scheduled too: near the end of a course, NG222 asks you to “discuss ways in which they can maintain the benefits of treatment” (1.4.9), and NG116 includes “planning booster sessions if needed, particularly in relation to significant dates (for example trauma anniversaries)” (1.6.17). What goes in that final note is in the therapy termination guide.

If you use Gesell, the plan it proposes includes an estimated number of sessions and, when you generate the full plan, a review session roughly every 6 to 8 sessions and a closing one; check that estimate against this table.

References

• NICE — Depression in adults: treatment and management (NG222), 2022, recommendations
• NICE — Generalised anxiety disorder and panic disorder in adults: management (CG113), 2011, updated 2020
• NICE — Social anxiety disorder: recognition, assessment and treatment (CG159), 2013
• NICE — Post-traumatic stress disorder (NG116), 2018
• NICE — Obsessive-compulsive disorder and body dysmorphic disorder: treatment (CG31), 2005
• Alianza por el Sueño — Guía práctica de Terapia Cognitivo-Conductual para el insomnio crónico (practical guide to CBT for chronic insomnia), 2026 (PDF hosted on infocop.es)
• IMSS — Diagnóstico y tratamiento del trastorno depresivo en el adulto (depressive disorder in adults), CPG IMSS-161-09, 2015 update (PDF)
• IMSS — Diagnóstico y tratamiento de los trastornos de ansiedad en el adulto (anxiety disorders in adults), CPG IMSS-392-10, 2010, updated 2011 (PDF)