How to Untangle Anxiety, Step by Step

An overlooked idea from the 1980s yields a promising new treatment approach that anyone can use

I’m a psychiatrist who struggles with anxiety.

I had my first full-blown panic attack when I was in residency training. It woke me from a sound sleep like a freight train suddenly blowing its whistle in my ear. Heart pounding and short of breath, I felt like I was going to die.

Instead of calling 911, I went through the psychiatric diagnostic checklist in my head. Check, check, check. Yup, those were all the symptoms of a panic (rather than heart) attack, I reassured myself. And with a hefty sleep deficit as my sleeping pill (thank you, residency), I nodded off again.

I had a few more panic attacks during those years, but had learned something in medical school that helped me not develop a full-blown panic disorder (which I might have otherwise been prone to). For someone to be diagnosed with panic disorder, according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), panic attacks “must be associated with longer than one month of subsequent persistent worry about: (1) having another attack or consequences of the attack, or (2) significant maladaptive behavioral changes related to the attack.”

Basically, worrying about future worry, or panicking about getting panicked, is what pulls you from just having panic attacks into the black hole of panic disorder.

In medical school and residency, I learned what medications to prescribe for patients with panic disorder and generalized anxiety. Selective Serotonin Reuptake Inhibitors (SSRIs) such as Prozac and Zoloft are the first line of treatment given their safety and generally minimal side effect profile. Benzodiazepines (“benzos”) such as Xanax, Klonopin, and Valium are to be used sparingly (and for short-term treatment) because of their addictive potential, which presents an additional danger when mixed with alcohol and/or opioids.

As a newly minted MD, I went off into the world to start treating patients. When I diagnosed someone with anxiety, if appropriate (and it generally was), I’d prescribe a pill. Yet, many of my patients came back saying the medication didn’t help. I’d increase the dose. They’d come back a month later saying the same thing. What was I missing?

There’s a medical term for the odds of a medication working: the number needed to treat (NNT). Similar to playing a game of craps, where you can calculate the likelihood of getting a seven when you roll the dice because you know all of the combinations that add up to seven, medical researchers can aggregate all of the research studies and calculate the likelihood of any one person benefitting from a given medication. For the family of SSRIs, the NNT is 5.15. In other words, I have to prescribe an SSRI to just over five patients before I can expect to see a significant benefit in one of them. Not the worst odds, but not great either. No wonder my patients were asking for something else. This disconnect got me struggling with their anxiety.

Anxiety triggers worrying as a mental behavior, which results in feeling like we are in control. So even if we can’t immediately fix our anxiety, at least we’re doing something, and worrying can feel like problem-solving.

As a researcher, I started wondering what we (the field of psychiatry) were missing. My main area of research at the time was addiction. I had some success in clinical trials of mindfulness training in helping people quit smoking: Our first randomized controlled trial showed a whopping five times better quit rate than the gold standard treatment of cognitive therapy. And a study at the University of California, San Francisco (UCSF) of an app-based mindfulness training I had developed showed a 40% reduction in craving-related eating. But it seemed anxiety didn’t really fit into the category of habitual behavior — or did it?

I dug back through the literature and found a pretty impressive array of articles from the 1980s (ironically, the same decade Prozac went to market) that suggested that anxiety could be perpetuated through a process called negative reinforcement. We learn certain behaviors — indeed habits — through this process. All we need is a trigger, a behavior, and a reward. For example, if we walk down the street with our shoe untied and we trip, tripping can trigger us to learn to make sure our shoes are tied (the behavior), and because not tripping feels better than face-planting on the sidewalk (the reward), we learn to make sure our shoes are tied on a regular basis. This is called negative reinforcement because learning the behavior prevents a negative outcome from happening in the future.

It turns out that anxiety triggers worrying as a mental behavior, which results in feeling like we are in control. So even if we can’t immediately fix our anxiety, at least we’re doing something, and worrying can feel like problem-solving — like we’re taking control of the situation. Essentially, worrying helps us avoid the feeling of anxiety. So because feeling like we’re in control and/or avoiding anxiety feels better than wallowing in the feeling of anxiety, it can be negatively reinforced.

What serendipity.

In medical school, in addition to memorizing drug dosing and side effect profiles, I had started learning mindfulness practice (which had led to my interest in studying it in my lab post-residency). In particular, I had learned a particular practice called “noting,” in which I would silently note to myself what was happening in my experience. For example, if I was walking down the street (shoes tied!), and I noticed a tree, I might note “leaf” or “green” if the green color was predominant in my experience. If seeing the tree triggered a certain emotion or body sensation, I might note “joy” or “heart beating.”

Noting has a particular psychological effect on us. Often, we are so identified with our thoughts, emotions, and body sensations that we don’t realize that these are momentary phenomena — they come and go. We get so caught up in them that they become habitual, taking on a life of their own, and owning us in the process. Observing our thoughts and feelings distances us enough that we can see them for what they are: thoughts and feelings.

In residency, I had been doing a lot of noting practice. I could do it while walking down the hospital halls, or before entering a patient’s room. It had become somewhat habitual. So when I had my first (and second, and third) panic attack, I woke up and noted — heart racing, tunnel vision, shortness of breath. This helped me keep perspective and observe what was happening: “Oh, I’m having a panic attack.” Critically, noting the experience helped me not build the attack into panic disorder because I wasn’t identified with it. I could see my thoughts and emotions as momentary experiences that came and went (typical panic attacks are surprisingly short) and didn’t start worrying about having a future panic attack.

Years later, I encountered a few research papers suggesting that mindfulness training, such as Mindfulness-Based Stress Reduction (MBSR) could help people with anxiety. Connecting the dots between anxiety, negative reinforcement, and my own success with noting practice, the team in my lab put together an app-based mindfulness training for anxiety. The idea was to help people learn that anxiety is negatively reinforced so that they could work with their own minds (by learning noting and other practices).

We performed several clinical trials with remarkable results — starting with a population that is particularly challenging to work with: physicians. Not only do we physicians learn to “armor up” in medical school (suppress our own emotions), we also carry the mentality that we don’t have time to take care of ourselves because we must focus on our patients (which ironically leads to us burning out and becoming less effective).

After three months of app-based mindfulness training, anxious physicians showed reductions in their own anxiety by 57%. In another randomized controlled trial, people with Generalized Anxiety Disorder (GAD) showed a drop of 67% in anxiety after two months of the same training. With these data in hand, we could calculate the NNT to see how effective the treatment was: 1.6. That’s right, just under two people needed to try this type of mindfulness training for one to show significant benefit.

This finding was astounding — and showed that it wasn’t just me. I came to understand, through a mix of personal experience and extensive research, that mindfulness tools could help others, like my patients. And, I’m happy to report that my own anxiety about being able to help my patients’ anxiety dropped significantly.

Addiction Psychiatrist. Neuroscientist. Habit Change Expert. Brown U. professor. Founder of MindSciences. Author: Unwinding Anxiety. www.drjud.com. @judbrewer

Sign up for Inside Your Head 🧠

By Elemental

A weekly newsletter exploring why your brain makes you think, feel, and act the way you do, by Elemental senior writer Dana Smith. Take a look.

By signing up, you will create a Medium account if you don’t already have one. Review our Privacy Policy for more information about our privacy practices.

Check your inbox
Medium sent you an email at to complete your subscription.