Exposure therapy is a treatment used by psychologists to help people overcome their fears. It is rooted in the common sense idea that facing your fears can be therapeutic, and that avoiding them can be counterproductive. There is compelling research that exposure therapy is an effective treatment for generalized anxiety, specific phobias, and post-traumatic stress disorder.
The success of exposure therapy can be explained scientifically with concepts from behaviorism (such as operant conditioning and inhibitory learning), or predictive processing (which is about how expectation affects perception). Under either model, the basic mechanism is the same: perception of threat is reduced when bad expectations are violated and negative associations are unlearned.
The science of exposure therapy for anxiety should be interesting to people who treat chronic pain for at least two reasons. First, people with pain might have unreasonable fears of movement, and this might create disability or prevent healthy exercise. To use a very simple example, someone with back pain might develop an extreme fear of forward bending, and they will need to face their fears at some point if they want to recover normal movement.
Second, fear itself may be contributing to pain, and therefore working to reduce the fear may also help treat the pain directly. Thus, we can view exercise as a form of “graded exposure” to the threat of feared movements. Exposure might help with pain by reducing kinesiophobia and catastrophizing, increasing self-efficacy and confidence, and inhibiting learned associations between movement and pain.
With these ideas in mind, I was interested to read this recent paper by Andre Pittig and colleagues called Change of threat expectancy as mechanism of exposure-based psychotherapy for anxiety disorders: Evidence from 8484 exposure exercises of 605 patients. (Thanks to Derek Griffin for pointing it out on Twitter.) It has a good deal of information that can be applied to the treatment of pain. Following are some key ideas supported by quotes.
Exposure therapy is generally effective but individual results vary:
Exposure-based cognitive-behavioral therapy (CBT) consistently yields large within-group treatment effect sizes for symptom reduction …[but] individual responses to exposure-based CBT vary substantially. While some patients show full remission of symptoms, others do not fully benefit. Moreover, average treatment success tends to stabilize or improve in the long-run but some individuals show a return of symptoms after successful treatment.
Exposure therapy inhibits negative associations by creating new positive associations:
Fear extinction refers to the process of learning that the anticipated threat does not occur or no longer occurs and is thus initiated by prediction error. Inhibitory learning theory assumes that this learning process does not erase the original threat association but prompts a novel association that the anticipated threat does not occur under specific circumstances. This novel learning actively inhibits the original threat association and thereby down-regulates fear and anxiety responses.
The most common strategy used in exposure therapy is to maximize expectancy violation:
The most frequently proposed strategy is the maximization of threat expectancy violation. … Expectancy violation refers to the mismatch between this threat expectancy and the actual occurrence. This mismatch is believed to be the clinical indicator for the occurrence of a prediction error and is thus assumed to create novel inhibitory learning. … it is thus assumed that the more threat expectancies are violated during exposure, the better the treatment outcome.
However, expectancy violation is insufficient for change if it is not accompanied by an actual change in expectancy:
Expectancy violation does not necessarily result in expectancy change. …
During exposure, a patient may, for example, test the specific threat expectancy “When I give a presentation, I will make a mistake and the audience will laugh at me.” by giving a presentation in front of an audience and experience no laughing (i.e., expectancy violation). However, if the threat expectancy remained unchanged for a repeated presentation, the learning rate would be 0.
In this paper, data analysis showed that expectancy change was the key variable predicting treatment success:
Not expectancy violation itself, but higher learning rate and expectancy change predicted better treatment outcome.
…
Thus, successful exposure not only requires a mismatch between threat expectancy and actual occurrence, but most importantly this mismatch needs to trigger an actual change of threat expectancies.
After reading the paper, I thought of the following practical takeaways.
First, we should make it a priority to identify all of the fears that a client might have about the connection between movement and pain. People tend not to talk about their fears, especially when they are unreasonable, so you probably won’t learn them without specific questions.
Some fears about movement will be rational and well-justified. For example, if past experience has shown repeatedly that heavy deadlifts cause back pain, it is entirely reasonable to expect something bad to happen after heavy deadlifts. But if the fear of picking things up also extends to suitcases and even pencils on the floor, perhaps it has some irrational components that could be easily tested with some simple experiments. But you might never learn about this fear without asking specific questions.
Second, it is easier to revise expectancies if they are testable, precise, and objective, as opposed to vague and subjective. Here’s some advice about this from the paper:
therapists were trained to define testable beliefs about observable feared outcomes (e.g., “When I stutter during the presentation, people will notice and make depreciating comments.”) but not … untestable outcomes (e.g., “Something bad will happen.”)
… when a testable outcome was unclear or ambiguous for a specific threat belief (e.g., “I will go crazy” or “I will lose control”), therapists were trained to determine concrete and testable outcomes or behavioral responses associated with these beliefs (e.g., by asking patients what would happen or how it would look like if they go crazy).
In the context of pain, consider the following expectation, which is designed to be highly testable: “if I run for three days in a row, my knees will have high levels of knee pain for at least a week.” This is a testable prediction, and there are many different results that might lead to a change in expectancy: if the pain was only mild, or lasted only two days, or only arose after five days of running. But an ambiguous prediction is not easily testable and is less likely to cause a change in expectancy. For example: “if I go running, my knees will hurt.” This kind of vague prediction will get confirmed by almost any result.
Third, we need to respect the fact many people will not be that good at updating their expectancies based on experience, because this kind of disposition is far more common in people with anxiety, depression and chronic pain.
The bottom line: seek to identify your clients’ unreasonable fears through specific conversations, and then devise experiments to test whether they are justified.
"Under either model, the basic mechanism is the same: perception of threat is reduced when bad expectations are violated and negative associations are unlearned."
Under predictive processing, though, discouraged people don't sense encouraging signals as strongly, and lack confidence that good news is a salient, motivating signal.
"Depression, traditionally viewed as a disorder characterized by negative cognitive biases, is associated with disrupted reward prediction error encoding and signaling. Accumulating evidence also suggests that depression is characterized by impaired local and long-range prediction signaling across multiple sensory domains." Depression, then, is sensory. If people typically can't unsee optical illusions despite rationally understanding how the illusions work, directly fixing depressed sensory processing with reasoning seems unlikely to work. Depressed people's vision sees less contrast. Their motor activity is less confident. It's as if they sense the world through a fog of uncertainty, and that may be so. "Generative models help an individual formulate predictions about incoming sensory information that are tested against incoming sensory inputs and produce prediction errors. Prediction errors, in turn, are used by the brain to revise its model of the world by updating predictions in order to minimize prediction errors (Friston, 2010)." The less confident predictions about sensory input are, the more inconclusive the results when they're tested against incoming input, and the less reason to register prediction error to revise the predictions. In depression, it seems, a fog of uncertainty unnecessarily stifles motivation.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8927302/
In some circumstances, though, a fog of uncertainty is a discouraging but accurate reflection of reality. These circumstances (say, a string of disorienting misdiagnoses, especially needlessly victim-blaming ones) are ones people with chronic pain are more likely to have been through. According to predictive processing, seeing what we expect to see is largely efficient, not necessarily confirmation bias. But because we do see what we expect to see, discouraged people are prone to oversampling discouraging information.
How do you get a person whose very sensory system lacks confidence in good news to believe it?
Specificity is good, but perception of distress is holistic, and it can be tough to trust good news (like "my knee hurts less than expected") when it comes embedded in overall "ooginess" (like "but now I've got a migraine and my ankle hurts"). People with chronic pain often juggle several problems at once, and their brains may not be particularly impressed by one problem subsiding if others flare.
If specific tests help convince a hesitant brain, it may also help if these tests require minimal self-monitoring. Checking whether knee pain lasts at least a week means *looking* for knee pain a whole week. That's awfully close to the rumination "this kind of disposition" is counseled to avoid. Easy measurements of what you can do without a significant increase in discomfort, like blocks walked or weight lifted, are less overwhelming to keep track of, and harder to psych yourself out over. Speaking just for myself, if I were set the assignment of checking whether a distressing level of knee pain lasted for an entire week and I found that it did, I'd wrack myself with guilt, doubt, and despair over whether I *really* felt that much pain, or just *thought* I did because I'm some kind of gloomy Gus too unmotivated to succeed.