Guys, I love that you do this work but y’all are really messing up how you talk about depression these last two shows. You never define it explicitly and then you talk about it implicitly like it’s just excessive bad moods. At one point you even speculated that exercise might lift some people’s moods but discourage others. (FWIW depressed people are pretty much frustrated and discouraged trying anything new — that’s a big part of how it maintains itself.)
Asking whether exercise treats depression is a bit like asking whether radiation to your head treats cancer. You need more information. Depression isn’t one thing any more than cancer is. But the standard measures we use for depression are handed down from a time when we mostly thought of depression symptoms as reflecting a common underlying disease process. The upshot is it’s impossible to publish a depression study without using a noisy standard measure that’s always a step removed from what you actually predict the benefits of your treatment will be. It’s a wonder we find effects of anything with measures this insensitive.
Maybe it’s beyond the scope of your show to review the evidence for a bunch of mechanisms. But I think you owe it to your listeners to explain depression as a construct and talk about the associated measurement limitations. That’s normally what you’re good at. Instead I feel like what you’ve done is effectively tell people their doctor doesn’t have good evidence to suggest exercise for their depression. It doesn’t sound to me like you’ve explored this topic enough to justify leaving that impression.
Just my two cents. Hope you’re doing well over there.
I'd bet the Cochrane review had better selection criteria, but that won't save it from a poorly framed question. This paper is Cochrane's standard medical way to look at something -- syndrome, treatment, and straightforward way to measure efficacy. You can analyze it just like a vaccine. No attention to mental processes required. What's all that got to do with depression anyway.
I had some thoughts after listening to your episode on exercise, and specifically the bit about exercise as a treatment for depression. I agree with you that the evidence on this is disappointingly weak.
You seemed to lampoon Simon Hattenstone’s view (1) that exercise clearly helps him, and that we should respect his self-knowledge on this. I certainly agree with you that he should not dismiss negative results (or describe them as “churlish” or “patronising”) merely on the basis of a strong contrary belief. But I would argue that there is an evidence base that supports respecting self-knowledge of what works in individuals.
I’m referring to the evidence on “behavioural activation”. Behavioural activation is a treatment for depression that involves working with people to change their behaviour so that they do more of the things that make them feel better. These may be things that give pleasure, or things that afford a sense of accomplishment or meaning. The precise activities will vary between people, because therapy is individualised – i.e. self-knowledge of what helps is at the heart of behavioural activation.
This relatively simple treatment may be similar in effectiveness to CBT (2), though like a lot of non-pharmacological interventions, there are limitations and uncertainties in its evidence base (3). Personally, (as a GP) I am comfortable that it is a reasonable thing to try with my patients who are open to it, especially when psychological treatments are hard to access and antidepressants aren’t appropriate (e.g. in milder or subsyndromal depression).
I would argue that for people like Hattenstone, who know they feel better when they exercise, what they are doing when they exercise is behavioural activation. His view amounts to more than merely anecdotal evidence, because it fits the description of an evidence-based behavioural intervention.
I don’t think this reasoning would be valid for all self-knowledge (e.g. “I’m coming off my antidepressants now because after the first couple of days they're making me feel worse” or “drinking alcohol makes me happier so I’ll open another bottle”). But for exercise in people like Hattenstone, I think it’s reasonable.
Perhaps exercise’s relatively small and uncertain effect size is because (as you suggest later in the episode) there are individual differences in response – i.e. only a proportion of people stand to benefit. If exercise only gives a sense of enjoyment or accomplishment to some people, then perhaps it is only in that subset of people that it functions as proper behavioural activation. Tom seemed to have a strong feeling that this would be the case, and I think the behavioural activation evidence base lends some rigour to his intuition. It’d be good to see research that explores this in more detail, not least because it might help predict responders.
I think this is an excellent point to raise. I use behavioral activation approaches all the time with therapy clients. I'd even take your point a step further and note that exercise doesn't even have to be a primary reinforcer for them -- in other words, they don't have to like it when the start. The aim is to help a client change their relationship to meaningful experiences they're avoiding. Exercise is initially an aversive experience for most depressives, so they're negatively reinforced (i.e., relieved) when they give up on it. When clients want to improve their health, you help them find positive reinforcement in pursuing their goal, set goals that are easier to pursue, and work through their aversiveness of doing something new. So it's not even a matter of the exercise functioning as an anti-depressant because it feels good to them. For most depressives, it won't feel good.
The big picture is that the effects of exercise for these clients depends a lot on how they understand their goals, the maintenance factors that sustain their depression, what type of depression they're experiencing, etc. It's fine to look for a main effect of exercise on client's symptoms as measured by a clunky depression scale. But you also have to recognize it's a much more complicated question. These are psychological disorders we're talking about, so we should assume going in that how clients understand their behaviors is going to moderate the effects.
In regards to why people are prickly about saying exercise treats depression, my experience has been that doctors and the public in general are very bad at talking about causal levels of treatment. If you have major depression, I think it’s likely that exercising may make a 5-10% difference in how you feel. But people hear “evidence-based treatment for depression” and think it means “first-order cure” in the same way antibiotics are a treatment for an infection. Depressed people don’t want to spend the time and energy to go to the doctor and be told advice that is a 5-10% effective treatment for a debilitating condition: they want a first-order cure/treatment.
Also, being told to exercise is sort of ridiculous without any follow-up or understanding of why people are depressed. Making time to exercise takes logistical planning. For instance, telling a new mom with postnatal depression to exercise when she has been awake with a sick baby for days is not at all helpful. This sounds like an exceptional example but it’s actually incredibly common for depressed people to have co-occurring conditions like chronic pain or other diagnoses that make even small lifestyle changes very difficult. And yet, many depression studies exclude these people with comorbid conditions up front to avoid “complicating the data” even though they may represent the standard patient! From what I understand, it’s actually pretty rare to have major depression and absolutely no other potentially related health conditions.
I think people get upset at these studies because there is a misconception that so-called “lifestyle” changes are somehow more socially acceptable and doable than pills when the exact opposite is true. Lifestyle changes—in my experience as someone who has been poor and mentally ill in the past—are the most burdensome and inaccessible to the patient and have the least medical benefit. I suspect the only reason we are still studying nonsense like mindfulness, diet, and exercise is for the sake of the researchers and not the patients. When briefly did psychology research at a clinic it was much easier to get IRBs for studies that involved behavioral interventions rather than experimental drugs (of course). And there is such a stigma around scientific solutions for anything (see adderall and Ozempic) that there is a huge pressure to produce these “exercise is good” studies even though there is little practical benefit to the very to patients who need treatment the most.
But that’s why I like this show; let’s stop telling everyone to do mindful vegan yoga and actually just get on with making real medicine with real science that people can actually use!
Thanks for the mention guys, plenty coming on the sports science replication center/project. Expecting at least one big pub on the project this year https://osf.io/3vufg/
As someone who is paying hefty co-payments for my twice-weekly physiotherapy to address a frozen shoulder, I was dismayed to hear about the poor quality studies in this discipline. Should I just save my money?
If it were me, I wouldn't make any big changes just off this conversation. Just because there are quality problems within a field overall definitely doesn't mean the evidence for any particular treatment is poor. An experienced physiotherapist might have really good intuitions for what will help that aren't reflected yet in the research literature.
Guys, I love that you do this work but y’all are really messing up how you talk about depression these last two shows. You never define it explicitly and then you talk about it implicitly like it’s just excessive bad moods. At one point you even speculated that exercise might lift some people’s moods but discourage others. (FWIW depressed people are pretty much frustrated and discouraged trying anything new — that’s a big part of how it maintains itself.)
Asking whether exercise treats depression is a bit like asking whether radiation to your head treats cancer. You need more information. Depression isn’t one thing any more than cancer is. But the standard measures we use for depression are handed down from a time when we mostly thought of depression symptoms as reflecting a common underlying disease process. The upshot is it’s impossible to publish a depression study without using a noisy standard measure that’s always a step removed from what you actually predict the benefits of your treatment will be. It’s a wonder we find effects of anything with measures this insensitive.
One way scientists try to get around this problem is to evaluate the evidence for hypothesized treatment mechanisms. For example, some people think SSRIs and aerobic exercise both increase BDNF, which in turn alleviates depression. (https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2023.1102526/full)
Maybe it’s beyond the scope of your show to review the evidence for a bunch of mechanisms. But I think you owe it to your listeners to explain depression as a construct and talk about the associated measurement limitations. That’s normally what you’re good at. Instead I feel like what you’ve done is effectively tell people their doctor doesn’t have good evidence to suggest exercise for their depression. It doesn’t sound to me like you’ve explored this topic enough to justify leaving that impression.
Just my two cents. Hope you’re doing well over there.
Matt
This meta-analysis on exercise and depression is very recent: https://www.bmj.com/content/384/bmj-2023-075847
Did you get a chance to cover it?
I'd bet the Cochrane review had better selection criteria, but that won't save it from a poorly framed question. This paper is Cochrane's standard medical way to look at something -- syndrome, treatment, and straightforward way to measure efficacy. You can analyze it just like a vaccine. No attention to mental processes required. What's all that got to do with depression anyway.
Hi Stuart and Tom,
Thanks for the podcast – I enjoy it a lot.
I had some thoughts after listening to your episode on exercise, and specifically the bit about exercise as a treatment for depression. I agree with you that the evidence on this is disappointingly weak.
You seemed to lampoon Simon Hattenstone’s view (1) that exercise clearly helps him, and that we should respect his self-knowledge on this. I certainly agree with you that he should not dismiss negative results (or describe them as “churlish” or “patronising”) merely on the basis of a strong contrary belief. But I would argue that there is an evidence base that supports respecting self-knowledge of what works in individuals.
I’m referring to the evidence on “behavioural activation”. Behavioural activation is a treatment for depression that involves working with people to change their behaviour so that they do more of the things that make them feel better. These may be things that give pleasure, or things that afford a sense of accomplishment or meaning. The precise activities will vary between people, because therapy is individualised – i.e. self-knowledge of what helps is at the heart of behavioural activation.
This relatively simple treatment may be similar in effectiveness to CBT (2), though like a lot of non-pharmacological interventions, there are limitations and uncertainties in its evidence base (3). Personally, (as a GP) I am comfortable that it is a reasonable thing to try with my patients who are open to it, especially when psychological treatments are hard to access and antidepressants aren’t appropriate (e.g. in milder or subsyndromal depression).
I would argue that for people like Hattenstone, who know they feel better when they exercise, what they are doing when they exercise is behavioural activation. His view amounts to more than merely anecdotal evidence, because it fits the description of an evidence-based behavioural intervention.
I don’t think this reasoning would be valid for all self-knowledge (e.g. “I’m coming off my antidepressants now because after the first couple of days they're making me feel worse” or “drinking alcohol makes me happier so I’ll open another bottle”). But for exercise in people like Hattenstone, I think it’s reasonable.
Perhaps exercise’s relatively small and uncertain effect size is because (as you suggest later in the episode) there are individual differences in response – i.e. only a proportion of people stand to benefit. If exercise only gives a sense of enjoyment or accomplishment to some people, then perhaps it is only in that subset of people that it functions as proper behavioural activation. Tom seemed to have a strong feeling that this would be the case, and I think the behavioural activation evidence base lends some rigour to his intuition. It’d be good to see research that explores this in more detail, not least because it might help predict responders.
What do you think?
Reference links:
1. Hattenstone's opinion piece in the Guardian: https://www.theguardian.com/commentisfree/2012/jun/06/exercise-depression-research-misses-point
2. Behavioural activation vs CBT trial in the Lancet: https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(16)31140-0/fulltext
3. Cochrane review re behavioural activation: https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD013305.pub2/full
I think this is an excellent point to raise. I use behavioral activation approaches all the time with therapy clients. I'd even take your point a step further and note that exercise doesn't even have to be a primary reinforcer for them -- in other words, they don't have to like it when the start. The aim is to help a client change their relationship to meaningful experiences they're avoiding. Exercise is initially an aversive experience for most depressives, so they're negatively reinforced (i.e., relieved) when they give up on it. When clients want to improve their health, you help them find positive reinforcement in pursuing their goal, set goals that are easier to pursue, and work through their aversiveness of doing something new. So it's not even a matter of the exercise functioning as an anti-depressant because it feels good to them. For most depressives, it won't feel good.
The big picture is that the effects of exercise for these clients depends a lot on how they understand their goals, the maintenance factors that sustain their depression, what type of depression they're experiencing, etc. It's fine to look for a main effect of exercise on client's symptoms as measured by a clunky depression scale. But you also have to recognize it's a much more complicated question. These are psychological disorders we're talking about, so we should assume going in that how clients understand their behaviors is going to moderate the effects.
In regards to why people are prickly about saying exercise treats depression, my experience has been that doctors and the public in general are very bad at talking about causal levels of treatment. If you have major depression, I think it’s likely that exercising may make a 5-10% difference in how you feel. But people hear “evidence-based treatment for depression” and think it means “first-order cure” in the same way antibiotics are a treatment for an infection. Depressed people don’t want to spend the time and energy to go to the doctor and be told advice that is a 5-10% effective treatment for a debilitating condition: they want a first-order cure/treatment.
Also, being told to exercise is sort of ridiculous without any follow-up or understanding of why people are depressed. Making time to exercise takes logistical planning. For instance, telling a new mom with postnatal depression to exercise when she has been awake with a sick baby for days is not at all helpful. This sounds like an exceptional example but it’s actually incredibly common for depressed people to have co-occurring conditions like chronic pain or other diagnoses that make even small lifestyle changes very difficult. And yet, many depression studies exclude these people with comorbid conditions up front to avoid “complicating the data” even though they may represent the standard patient! From what I understand, it’s actually pretty rare to have major depression and absolutely no other potentially related health conditions.
I think people get upset at these studies because there is a misconception that so-called “lifestyle” changes are somehow more socially acceptable and doable than pills when the exact opposite is true. Lifestyle changes—in my experience as someone who has been poor and mentally ill in the past—are the most burdensome and inaccessible to the patient and have the least medical benefit. I suspect the only reason we are still studying nonsense like mindfulness, diet, and exercise is for the sake of the researchers and not the patients. When briefly did psychology research at a clinic it was much easier to get IRBs for studies that involved behavioral interventions rather than experimental drugs (of course). And there is such a stigma around scientific solutions for anything (see adderall and Ozempic) that there is a huge pressure to produce these “exercise is good” studies even though there is little practical benefit to the very to patients who need treatment the most.
But that’s why I like this show; let’s stop telling everyone to do mindful vegan yoga and actually just get on with making real medicine with real science that people can actually use!
Thanks for the mention guys, plenty coming on the sports science replication center/project. Expecting at least one big pub on the project this year https://osf.io/3vufg/
Prior to Kinesio tape there was the fad for sports people to wear those nasal strips that supposedly helped breathing. What happened to them?
As someone who is paying hefty co-payments for my twice-weekly physiotherapy to address a frozen shoulder, I was dismayed to hear about the poor quality studies in this discipline. Should I just save my money?
If it were me, I wouldn't make any big changes just off this conversation. Just because there are quality problems within a field overall definitely doesn't mean the evidence for any particular treatment is poor. An experienced physiotherapist might have really good intuitions for what will help that aren't reflected yet in the research literature.