TL;DR: CBT-AR is the leading evidence-based protocol for ARFID in adults, developed by Jennifer Thomas and Kamryn Eddy at Massachusetts General Hospital. It runs 20 to 30 sessions across four stages and targets three maintaining mechanisms: sensory sensitivity, fear of aversive consequences, and low interest in food. The evidence in adults is promising and still developing. The protocol is honest about what it does and does not address, and adult patients usually benefit from clarity about both.

A patient calls to ask whether CBT-AR will work for her. She is thirty-six. She has eaten roughly twelve foods for as long as she can remember. Her bloodwork is starting to show the consequences. She has read a Substack post about the Thomas and Eddy manual and wants to know whether the protocol will get her to eat a salad by Christmas. She is asking a clinical question and a hope question at the same time, and she deserves a careful answer to both.

This is what a careful answer looks like.

What CBT-AR is

CBT-AR (Cognitive Behavioral Therapy for Avoidant/Restrictive Food Intake Disorder) is a manualized 20-to-30 session protocol developed by Jennifer Thomas and Kamryn Eddy at the Massachusetts General Hospital eating disorders program and published in their 2019 treatment manual Cognitive-Behavioral Therapy for Avoidant/Restrictive Food Intake Disorder. It is the most studied psychosocial intervention specifically built for ARFID, and it has been adapted across age ranges from 10 through adulthood.

The protocol is targeted, structured, and limited by design. It is not a comprehensive treatment for every concern an adult with ARFID may bring to therapy. It is a focused intervention on the eating pattern itself.

The three maintaining mechanisms

The Thomas and Eddy formulation identifies three mechanisms that maintain ARFID in any given patient. Each patient typically presents with one primary mechanism and often a secondary one.

Sensory sensitivity. The patient experiences specific food properties (texture, smell, color, temperature) as intensely aversive in a way that is not under conscious control. The gag reflex on a texture is not behavioral resistance; it is a neurological response. The sensory-sensitive patient is the one who can describe in precise detail what is wrong with the surface of a tomato. Treatment focuses on sensory exposure and gradual habituation.

Fear of aversive consequences. The patient avoids food because of an anticipated bad outcome: choking, vomiting, allergic reaction, or pain. The fear is often, though not always, traceable to a specific past event (a choking incident, a serious gastrointestinal illness). Treatment focuses on interoceptive exposure to the feared bodily sensations and graduated exposure to the avoided foods.

Lack of interest in food or low appetite. The patient does not experience hunger reliably, finds eating effortful or boring, and tends to forget meals. Interoceptive awareness of hunger and satiety is often diminished. Treatment focuses on structured eating, behavioral activation, and rebuilding the felt signal of appetite.

The clinical value of the formulation is that it specifies what the work is targeting. A treatment that addresses sensory sensitivity in a fear-based patient will not produce the expected results, and the manual is explicit about this.

The four stages

Stage 1 (Sessions 1 to 4): Psychoeducation, motivation, and assessment. The patient learns the ARFID model, identifies which maintaining mechanisms apply, completes the structured eating record, and engages in motivation-enhancement work. For adults who have been restricted for decades, this stage often surfaces grief and shame that need clinical attention before the active treatment begins. A clinician who rushes Stage 1 in adults often regrets it.

Stage 2 (Sessions 5 to 10): Building the hierarchy and beginning in-session exposure. The patient and clinician build a hierarchy of avoided foods graded by approximate difficulty. New foods are introduced in session, starting with foods that share properties with existing safe foods and moving outward. The in-session structure matters because it builds the patient’s confidence that the work can be done and gives the clinician a direct view of what is happening at the moment of exposure.

Stage 3 (variable, typically Sessions 11 to 25): Focused mechanism work. This is the longest and most variable stage. Treatment proceeds according to which maintaining mechanism is primary. Sensory-sensitive patients work through structured sensory exposure, often with food-chaining techniques that move from familiar to less familiar via small graded changes. Fear-based patients work through interoceptive exposure (deliberately producing the feared sensations to reduce their threat value) followed by exposure to the foods themselves. Low-interest patients work on structured eating schedules, behavioral activation around mealtimes, and rebuilding the interoceptive signal of hunger.

Stage 4 (final 2 to 4 sessions): Relapse prevention and consolidation. The patient and clinician identify high-risk situations, build a written maintenance plan, and conduct a structured ending. Booster sessions are often scheduled at three and six months.

What the adult evidence actually says

The 2021 open-trial study of CBT-AR in adults, conducted at Massachusetts General Hospital, reported significant improvements in food variety, weight when weight gain was indicated, and overall ARFID symptoms across 20 to 30 sessions. The effect sizes were comparable to those reported in the adolescent literature. A larger randomized controlled trial in adults is in progress at the time of writing.

The evidence base is real and growing. It is also still developing. A clinician who claims definitive RCT-level evidence in adult populations is overstating the literature. A clinician who dismisses CBT-AR as unproven is also overstating, because the available evidence converges on positive outcomes and the protocol is methodologically careful in a literature that historically was not.

What CBT-AR does not address

The protocol is honest about its scope, which is one of its clinical virtues. CBT-AR does not address:

  • The decades of accumulated social shame an adult often carries from a lifetime of being labeled a picky eater
  • Family-system patterns that have organized around the restriction (the spouse who has done all the cooking, the parent who still negotiates around the patient’s foods)
  • Autistic sensory profiles, which often require accommodation rather than exposure and which the standard CBT-AR sensory protocol can mishandle if applied without adaptation
  • Complex trauma that may underlie fear-based ARFID, particularly when the original aversive event was itself traumatic
  • The somatic dysregulation that often accompanies long-standing ARFID and that may require body-based adjunctive work

These usually call for adjunctive treatment, sequential or concurrent, and a clinician should be explicit with the patient about which parts of the work CBT-AR will and will not deliver.

Somatic and other adjuncts

The somatic-treatment literature for ARFID is thin. Some patients benefit from somatic experiencing, polyvagal-informed work, or body-based regulation strategies as adjuncts to CBT-AR, particularly when the underlying autonomic activation is high. The evidence is preliminary; clinicians offering somatic work for ARFID should be honest about that and should not present body-based interventions as a substitute for the symptom-targeted protocol that has the actual evidence behind it.

What to ask a CBT-AR clinician

A patient evaluating a clinician for adult CBT-AR should ask: which of the three maintaining mechanisms do you think is primary in my case, and what evidence are you basing that on; how long do you anticipate Stage 3 will take given my history; what adjunctive supports do you recommend, and why; and what does relapse prevention look like in your practice. A clinician who can answer these directly is offering the work the manual was built to deliver.

The patient who called to ask whether she will eat a salad by Christmas deserves the same kind of answer. CBT-AR can substantially expand the food repertoire of an adult with long-standing restriction across 20 to 30 sessions of careful work. Whether a salad is on that list depends on which mechanisms are driving her restriction, what hierarchy makes sense for her, and what she chooses to put on it. The protocol is not a promise. It is a method. The method works.

Brian Nuckols, MA, LPC-A, treats ARFID in adolescents and adults at his Pittsburgh practice. To discuss CBT-AR or whether your situation calls for a different sequence of care, see the contact page.