The alarm, not the fire: David Clark's 1986 model of panic and how well it has held up

If you have ever had a panic attack, you know the moment: the heart thumps, the chest tightens, and a thought arrives that feels like a fact — I'm having a heart attack, I'm going to faint, I'm losing my mind. In 1986 the British psychologist David Clark proposed that this thought is not just a by-product of panic but the engine of it. His short paper changed how panic is understood and how people learn to handle it. Forty years on, the core idea has partial support, a serious rival, and a clear practical legacy.

The original study

"A cognitive approach to panic" was a ten-page theoretical paper, not an experiment. Its central claim, as later reviews summarise it, is that panic attacks come from misinterpreting harmless body sensations as signs of an immediate catastrophe — physical, mental or social. A pounding heart is read as a coming heart attack; light-headedness as fainting; a racing mind as "going mad".

The model is usually taught as a loop. Something — a staircase, coffee, a stressful thought — produces a sensation. The sensation is read as dangerous. That reading creates fear, fear produces more sensations, and the stronger sensations seem to confirm the danger. The loop can spin up to a peak within minutes, which is why, in this account, panic can seem to come "out of nowhere": the trigger is a body sensation.

Two further predictions followed. First, people with frequent panic attacks should have a lasting tendency to read body sensations catastrophically, measurable even when they are calm. Second — and most important for practice — changing the interpretation should stop the loop.

What happened next

A method built on the model. Clark and colleagues developed a cognitive approach to panic. A 1991 case series gave preliminary evidence that cognitive work aimed specifically at misinterpretations can reduce how often panic attacks happen, while cognitive work that did not target them may not (Salkovskis, Clark & Hackmann, 1991). In a 1999 randomised trial with 43 participants, both a full and a brief version (the brief one needed about 6.5 hours of specialist time) beat a waiting list, with very large effects (around 3.0) that held at 12 months (Clark et al., 1999). The numbers are striking, but the trial was small and run by the model's own authors — a setting where results can look more favourable than in independent replications.

Testing the core claim. A 2001 review concluded that the evidence for catastrophic misinterpretation being specific to panic, and for it being the process that drives panic, was "partial or inconclusive", and pointed to inconsistent ways of measuring it (Austin & Richards, 2001). A 2018 meta-analysis of seven studies was more favourable: people with recurrent panic read body sensations catastrophically more than healthy people and more than people with other kinds of anxiety, with medium-to-large differences (Ohst & Tuschen-Caffier, 2018). That supports the "distinctive feature" prediction — though with only seven studies, a single study can sway the result, as the authors themselves note.

A serious rival. Learning theorists offered a different account. In this view, early panic attacks condition fear to internal cues (sensations) and external ones (places), so later sensations can trigger panic directly — without a catastrophic thought being necessary; the frightening thoughts may be a side effect rather than the cause (Bouton, Mineka & Barlow, 2001, as summarised by Ohst & Tuschen-Caffier, 2018). One phenomenon often discussed here is nocturnal panic: waking from sleep already in a state of panic, reported at least once by 44–71 % of people with recurrent panic in the studies reviewed by Craske & Tsao (2005). Programmes for nocturnal panic, for example, target misreadings of sensations, over-breathing and learned reactions to body cues, drawing on both camps (Craske & Tsao, 2005).

Which ingredients matter. The most direct test of the model's practical legacy is to take programmes apart. A component network meta-analysis of 72 trials with 4,064 participants found that deliberately bringing on the feared sensations in a safe way (interoceptive exposure) and face-to-face delivery were linked to better results and fewer dropouts, while muscle relaxation and virtual-reality exposure were linked to worse results; breathing retraining made programmes more acceptable but added little to their effect (Pompoli et al., 2018). The difference between the best and the worst combination that could both be called "evidence-based" was large (odds ratio for remission 7.69, though with a very wide credible interval of 1.75 to 33.33).

Where it stands today

Evidence: Moderate evidence. That people with recurrent panic read body sensations more catastrophically than others is reasonably supported. Whether that reading is the cause of panic, or one strand alongside learned fear of sensations, is still debated. What has held up best is the practical consequence: learning, through experience, that the sensations are uncomfortable but not dangerous — above all by meeting them on purpose rather than avoiding them.

What this means for you

The most useful idea from Clark's model fits in one sentence: the sensations are the alarm, not the fire.

• Start with what the alarm is and why it fires at sensations: What anxiety actually is.
• Notice the thought that turns a sensation into a catastrophe, and see it as a prediction to test: Thoughts are not facts.
• When a wave comes, stop adding a second wave of fighting: Riding the wave, and in the moment, SOS · The Wave.
• Learn why meeting the sensations on purpose helps — and do the health check first: Befriending body sensations on purpose. It is not for you without a doctor's okay if you have a heart, blood-pressure or breathing condition, epilepsy, a history of fainting, or are pregnant.

If you have never discussed intense physical surges with a doctor — especially chest pain, fainting or sensations that are new or different — get a medical check first. If panic attacks are frequent or you have started avoiding places because of them, a licensed mental health professional can help you work through this properly.

What this does NOT mean

• It does not mean your sensations are "imaginary": they are real physical changes, just not dangerous ones.
• It does not mean every chest pain is panic — new or unusual sensations deserve a medical check.
• It does not mean you can simply argue yourself out of panic in the moment; the evidence favours learning by experience, not debate.
• It does not mean relaxation or slow breathing are the core of overcoming panic; in the component analysis they added little or less.

Sources

• Clark DM. A cognitive approach to panic. Behav Res Ther, 1986;24(4):461–470. DOI: 10.1016/0005-7967(86)90011-2
• Salkovskis PM, Clark DM, Hackmann A. Treatment of panic attacks using cognitive therapy without exposure or breathing retraining. Behav Res Ther, 1991. DOI: 10.1016/0005-7967(91)90044-4
• Clark DM, Salkovskis PM, Hackmann A, et al. Brief cognitive therapy for panic disorder: a randomized controlled trial. J Consult Clin Psychol, 1999. DOI: 10.1037//0022-006x.67.4.583
• Austin DW, Richards JC. The catastrophic misinterpretation model of panic disorder. Behav Res Ther, 2001. DOI: 10.1016/s0005-7967(00)00095-4
• Bouton ME, Mineka S, Barlow DH. A modern learning theory perspective on the etiology of panic disorder. Psychol Rev, 2001. DOI: 10.1037/0033-295x.108.1.4
• Craske MG, Tsao JC. Assessment and treatment of nocturnal panic attacks. Sleep Med Rev, 2005. DOI: 10.1016/j.smrv.2004.11.003
• Pompoli A, Furukawa TA, Efthimiou O, et al. Dismantling cognitive-behaviour therapy for panic disorder: a systematic review and component network meta-analysis. Psychol Med, 2018. DOI: 10.1017/S0033291717003919
• Ohst B, Tuschen-Caffier B. Catastrophic misinterpretation of bodily sensations and external events in panic disorder, other anxiety disorders, and healthy subjects: a systematic review and meta-analysis. PLoS One, 2018. DOI: 10.1371/journal.pone.0194493