Avoidance, Limits and Protection
What am I moving away from, and why?
A quick note on this page. In CBT, avoidance is one of the main things that keeps anxiety going, and exposure is one of its best-supported treatments. Both ideas are sound. But avoidance cannot be identified just because someone does less. For a disabled person, “not going” may be fear, protection from a real risk, a decision about energy, a refusal, or several of these at once.
The short version
In CBT, avoidance can keep fear going. But doing less can also mean protecting yourself from a real risk, conserving energy, or simply choosing not to. One action can do several of these at once. Working out what a behaviour does, over time and together, comes before deciding whether to change it.
You might recognise this
Illustrative statements drawn together from common experiences. They are not quotes from individual people.
- “They called it avoidance. I called it needing two days to recover.”
- “I stopped going because the last three times I couldn’t get in the door.”
- “I don’t go to that clinic because of how they spoke to me. That isn’t a phobia.”
- “I tell people it’s my back. If I’m honest, I’m also scared.”
- “I wanted someone to help me face it, not to talk me out of trying.”
What CBT means by avoidance
In the CBT model of anxiety, avoiding a feared situation brings immediate relief. That relief makes avoidance more likely next time. Behavioural theory calls this negative reinforcement: a behaviour becomes more likely because it takes something unpleasant away. It also prevents the person from learning that the situation may be safer, or more manageable, than it feels. Over time the feared area grows and life narrows.
Paul Salkovskis and others added the idea of safety behaviours: subtler actions, like gripping a rail, rehearsing every sentence or always sitting by the door, that let a person stay in a situation while still preventing the same new learning. Exposure therapy works by approaching what is feared, in planned steps, without those protections, so that fear can be relearned.
All of this rests on one condition: the feared situation must be safer than it feels. When it is not, the model does not apply.
Four ways of moving away
Different reasons, different responses
- Fear that outruns the risk. The classic target of exposure. The situation is reasonably safe; the fear says otherwise.
- Protection from a real risk. The place is inaccessible, unsafe, or has been hostile before. Moving away is sensible.
- Conservation. Taking part costs energy, pain or recovery time that the person has decided not to spend. That is a judgement about resources, not a symptom.
- Refusal. The person simply does not want to. Therapy does not get to decide that every activity is one they should value.
These functions can also overlap in a single action. Leaving a party early may protect energy and bring relief from anxiety at the same time. Immediate relief, on its own, does not prove that something is problematic avoidance.
Most real situations mix these. Someone may avoid dating partly because many venues are inaccessible and partly because they expect rejection. Therapy can work with the expectation while practical planning deals with the venues. It does not need to pick one explanation and discard the other.
The fear-avoidance model, and its limits
In chronic pain, CBT draws on the fear-avoidance model described by Johan Vlaeyen and Steven Linton. It proposes that when pain is interpreted as a sign of serious harm, fear of movement can lead to avoidance, and avoidance can lead to greater disability and distress. The authors revisited the model in 2012. Vlaeyen & Linton, 2012
For many people with musculoskeletal pain, the model describes something real and useful. But it was built around conditions where movement, though painful, is generally not damaging. It says much less about conditions where activity carries a genuine physiological cost. In ME/CFS, for example, NICE describes post-exertional malaise: symptoms that can worsen after even minimal physical, cognitive, emotional or social activity. Reducing activity there is not a fear response to be extinguished. NICE NG206
A researcher writing for the International Association for the Study of Pain has also noted that far more research has looked at pain-related fear than at avoidance itself, and the links between them are still not fully understood. IASP Relief, 2021
Look at what the behaviour does
CBT has a practical tool for this: asking what a behaviour achieves, rather than judging what it looks like. Take cancelling an appointment. It might:
- reduce fear;
- prevent a symptom crash;
- avoid a professional who has been dismissive;
- save energy for essential care;
- express a refusal the person hasn’t been able to say out loud;
- reflect a transport failure;
- or several of these together.
A simple sequence helps: what happened before → what I did → what changed straight away → what happened later. The “later” matters. Fear-driven avoidance tends to bring relief now and a narrower life later. Rest that protects capacity tends to cost something now and make participation possible later.
Working this out usually takes more than one conversation. A first guess about what a behaviour does is a hypothesis, refined as you both notice more.
Questions before encouraging approach
- Is the situation genuinely accessible?
- Is there a meaningful risk, physical, social or emotional?
- What does taking part cost, and when is that cost paid?
- What support would change the situation?
- Does the person actually want to take part?
- Is fear adding restriction beyond the real constraint?
- How would we know if the plan was helping, or causing harm?
A moment in the session
An illustrative exchange, written for teaching. It is not a real session.
When exposure does fit
Exposure can be valuable for disabled people when the fear really does outrun the risk. Someone may want to ask a trusted colleague for an adjustment, and expect a cold reaction, despite good evidence that the colleague is responsive. A planned request can test that prediction, and the client may learn something important.
That is different from repeatedly placing someone in a hostile or inaccessible setting so they learn to tolerate it. Two rules help:
- Medical limits are not exposure targets. Pain that signals harm, post-exertional symptoms and real physical risk are not fears to be extinguished.
- The experiment must allow reality to answer. If the colleague does react badly, that is information. It may change the formulation, or point to a need for boundaries rather than more exposure.
Access supports are not safety behaviours
Captions, mobility aids, communication devices, an assistant, necessary medication, access checks and planned rest are not automatically safety behaviours. They are often what makes taking part possible. An access check can be essential; checking again and again after it stops giving new information can also become part of an anxiety cycle. The question is always about function, and it is worked out together. The CBT hub sets this out in more detail.
Reviewing it together
- What happened?
- What did it cost, on the day and afterwards?
- What did we learn?
- What remains uncertain?
- Do you want to repeat it, change it, or stop?
Signs of change
- Being able to say which kind of moving away this is, and when it is more than one.
- Going to something that matters, with the support that makes it possible.
- Facing a fear that turned out to be bigger than the risk.
- Declining something without guilt, because the cost was real.
- Less shame about rest.
For therapists
- Have I checked access, risk and cost before calling this avoidance?
- Am I treating rest as a behaviour to reduce?
- Is my exposure hierarchy aimed at a fear, or at a barrier?
- Would I be comfortable if the client concluded the situation really is unsafe?
A different experience
Sometimes a client tells the therapist that fear is part of the picture alongside a physical limit: “It’s my back, and if I’m honest I’m also scared.” When the person raises this, there is room to explore it, and to choose exposure work if they want it. It is not a reason for therapists to suspect that physical limits hide fear.
Evidence and status
- Sources checked in October 2026, using targeted searches rather than a systematic review. “We have not located” means our searches did not find it, not that it does not exist.
- Documented theory: the CBT models of anxiety maintenance, safety behaviours and exposure; the fear-avoidance model of chronic pain.
- Research: exposure-based CBT has strong support for anxiety disorders. NICE guidance on ME/CFS sets out why activity cannot be treated the same way there.
- Our synthesis: the four-way distinction between fear, protection, conservation and refusal. We have not located a standard assessment tool that separates fear-driven from barrier-driven avoidance for disabled clients.
A question worth carrying: if I stopped calling it avoidance for a moment, what would I call it instead?
Where this connects
Approach what is feared. Respect what is real. Ask which is which.
Within the CBT section
- CBT Through a Disability Lens – the hub, including safety behaviours and access supports.
- When “Cognitive Distortions” Are Actually Reality – the same question, asked about thoughts.
- Pain, Fatigue and Fluctuating Capacity – when energy is the limit.
Across the site
- Safety, Agency & Choice – choice as more than a token.
- Acceptance and Commitment Therapy – a different view of approach and avoidance.
- When Disability Gets Misread as Psychology – the wider pattern.
← Cognitive Restructuring · CBT section map · Next: Behavioural Activation →