Safety Behaviors in Anxiety: What Research Says About the Things We Do to Cope

A person with a fear of flying may board the plane but grip the armrest throughout, convinced that their vigilance is what keeps it in the air. Someone with social anxiety may attend the meeting but rehearse every sentence before speaking, leave early, or avoid eye contact. These are not full avoidance — the person enters the feared situation — but something subtler is happening. Cognitive models call these safety behaviors, and they may be part of what keeps anxiety in place.

What safety behaviors are

The concept was formalized by Salkovskis in 1991 as part of his cognitive model of panic and anxiety. Safety behaviors are actions taken within a feared situation to prevent a dreaded outcome [1]. The person believes, explicitly or implicitly, that without the behavior something catastrophic would happen: they would faint, be humiliated, lose control or cause harm.

Common examples by anxiety type:

• Panic disorder: carrying medication, staying near exits, monitoring heart rate, sitting down to prevent fainting
• Social anxiety: rehearsing sentences, avoiding eye contact, speaking quietly, wearing concealing clothing
• Health anxiety: checking symptoms repeatedly, seeking reassurance from doctors, avoiding reading about illness
• Generalized anxiety: making excessive lists, calling family members to check they are safe, over-preparing for unlikely events

The defining feature is that the behavior is performed to prevent a feared outcome, and the person believes it is necessary for safety.

How they are thought to maintain anxiety

The cognitive account is straightforward. When a person enters a feared situation while using a safety behavior, and the feared outcome does not happen, they attribute their survival to the behavior rather than revising their belief about the danger. “I didn’t faint because I sat near the exit” instead of “I didn’t faint because fainting was never likely.” The threat belief survives intact, and the anxiety persists.

This is a problem for exposure therapy, which works by allowing the person to learn that feared outcomes are less likely or less catastrophic than believed. If safety behaviors prevent that learning, exposure may not produce lasting change.

The experimental evidence

Salkovskis and colleagues tested this directly in 1999 with people who had panic disorder with agoraphobia. Participants did exposure in two conditions: one where they were explicitly asked to drop their safety behaviors, and one where they used them freely. Dropping safety behaviors produced greater reductions in anxiety beliefs and symptoms [2].

A similar experiment in social anxiety had participants give speeches either using or not using their typical safety behaviors (reading from notes, avoiding eye contact). Those who dropped safety behaviors received more favorable ratings from observers and reported lower anxiety afterwards, consistent with the idea that the behaviors were preventing disconfirmation [3].

A systematic review of 11 experimental and clinical studies on safety behaviors in panic treatment found that the evidence generally supports the cognitive model: dropping safety behaviors during exposure tends to produce better outcomes. However, the review noted that the evidence was mixed, with some studies finding no difference, and that the quality of many studies was limited [4].

A counterargument: judicious use

Not everyone agrees that safety behaviors should always be eliminated. Rachman, Radomsky and Shafran argued in an influential 2008 paper that the recommendation to eliminate safety behaviors was based on limited evidence and that a more nuanced approach was warranted [5]. They proposed that safety behaviors could sometimes be used strategically — for example, to help someone enter a feared situation they would otherwise completely avoid.

The argument is that total refusal of any safety behavior may deter people from starting exposure at all. If a person with severe agoraphobia will only leave the house if they carry their phone, insisting they leave it behind might mean they never leave. Used as a stepping stone rather than a permanent crutch, some safety behaviors might facilitate rather than hinder treatment.

This position remains controversial. The inhibitory learning framework for exposure therapy, proposed by Craske and colleagues, emphasizes expectancy violation — maximizing the mismatch between what the person fears and what actually happens [6]. Under this framework, safety behaviors that reduce expectancy violation should impair learning. But a safety behavior that allows a more intense exposure (one the person would otherwise avoid) might produce a larger violation, and thus more learning, than a milder exposure without the behavior.

Where the evidence is thin

Several questions remain open:

• Which safety behaviors matter most? Some may be more maintenance-relevant than others. Carrying medication in a pocket may have a different effect from actively monitoring one’s heart rate.
• Fading versus elimination. Whether gradually reducing safety behaviors produces equivalent outcomes to immediate elimination has not been well tested.
• Individual differences. Whether some people are more affected by safety behaviors than others is largely unstudied.
• Self-identification. People often do not recognize their own safety behaviors, which makes assessment difficult. What a therapist identifies as a safety behavior the patient may experience as a reasonable precaution.

The distinction problem

Perhaps the most intellectually honest difficulty is that the line between a safety behavior and a coping strategy is blurred. Deep breathing during a panic attack could be a safety behavior (done to prevent fainting) or a coping strategy (done to manage discomfort). The difference lies in the function and the belief behind it, which are not always easy to assess.

This matters because therapists need to know which behaviors to target. Getting it wrong in either direction is costly: targeting a helpful coping strategy may undermine the person’s confidence, while leaving a maintaining safety behavior unchallenged may limit treatment gains.

For how exposure therapy’s effects are measured more broadly, see what exposure therapy measures.

When to look for more support. If anxiety is limiting your life and you recognize yourself in these patterns, a therapist trained in CBT or exposure therapy can help you identify and test your safety behaviors in a structured way. Talk to your GP for a referral. If you are in crisis, contact emergency services or call/text 988 (US) or 116 123 (UK and Ireland, Samaritans).