TL;DR version: Mental illness is probably much worse than poverty or physical illness. Interventions which change how people think - i.e. reduce mental illness and increase happiness - may be more cost-effective ways of increasing happiness than AMF or Give Directly. I outline some new opportunities EA should look into.
Update (18/06/2017): this paper is now slightly out of date. It doesn't reflect my latest thinking and contains some errors. For one, it was unhelpful to compare saving lives (preventing premature deaths) to improving lives (increasing the average happiness within a life) as there are multiple views you can have about the value of saving lives. For more on that, see this post questioning anyone should want to give to AMF.
Longer version:
Hello EAs,
I don't really use the forum but I've been encouraged by others (Rob Wiblin of 80k, Joey Savoie of Charity Entrepreneurship) to share some of my research I've discussed with them here. I've written a 5,000 word draft paper "What should a billionaire do to maximise world happiness" as a potential draft chapter in my PhD where I argue EA currently overlooks human happiness and mental health.*
Here's the summary of main points which I discuss in greater depth in the draft:
1. Effective altruism has so far focused on "external happiness interventions" (EHIs) which aim make people happier by changing their external circumstances (e.g. poverty, malaria). It's neglected "internal happiness interventions" (IHIs) which try to increase happiness by changing how people think (e.g. mental health treatments).
2. It's very plausible, but as yet untested, that some Internal Happiness Interventions might turn out to be more effective than our current EHIs such as Give Directly or AMF.
3. Mental health and ordinary human unhappiness (e.g. 'normal' stress, worry, sadness) are big. The former affects 700m+ (depends how you count it) and everyone suffers from the latter.
4. Our intuitions about how happiness works are very misleading. We adapt to lots of changes (hedonic adaptation) are a very bad at correctly predicting how we'll feel in the future (affective forecasting). For instance, people in poverty are not as unhappy as you imagine they are.
5. It's questionable whether cash transfers to those in poverty will increase happiness. The only RCT into Give Directly showed their cash transfers had no long-term effect on life satisfaction scores. The trial showed GD's recipients did have increased life satisfaction in the short-term (6 months) but that the non-recipients had their life satisfaction go down by more than recipients' went up. This suggest Give Directly's work does not increase happiness (taking 'happiness' as 'life satisfaction'). More research is needed.
6. QALYs/DALYs very likely underrate the badness of mental health conditions on happiness. First, they are measures of health, not happiness. Second, their weighting are creating by asking people how bad they expect various conditions to be, rather than assessing asking people with those conditions to report their subjective well-being. As mental health conditions are hard to imagine and hard to adapt to, their are underrated. As a estimate that can definitely be argued about, they might be 10-18 times worse than we imagine them to be.
7. Putting all this together, I guess that mental health and ordinary human unhappiness cause 4-72 times more misery each year to living people than do poverty and malaria combined (see Annex B of my paper). This shouldn't be taken too seriously, it's just to get a sense of scale.
8. Mental health and ordinary unhappiness unhappiness might be surprisingly tractable. There are quite a few methods which seem to work for both of them: cognitive behavioural therapy, mindfulness-based stress reduction and positive psychology to name the most promising. They can be delivered in person and, also exciting, electronically. Anti-depressants also seem to be somewhat effective (although I don't discuss this in the paper).
9. No one has really tried to disseminate these widely. Therefore promising strategies for EA are: digital public health campaigns for happiness/mental health; setting up new charities in the developing world to deliver in-person therapy and drugs ("AMF for Prozac" H/T Rob Wiblin); lobbying developed world governments to do stuff.
10. Once you revise the $/DALY cost-effectiveness figures to take into account how DALYs underrates happiness, it's possible treatments for depression at around $1000/DALY (such as by Stronger Minds) are in roughly the same ball park as AMF, which is $100/DALY (this gets confusing when you account for the badness of death). Given that mental health treatment is new compared to physical health treatments, there is good reason to be optimistic that our mental health treatments will get much more effective in the future.
11. I discuss some of the objections. Most relevant: if you already think X-risk is the biggest problem, my argument probably won't bite. If you support AMF or Give Directly, I think you should reconsider those and join me. If you support animal welfare you may want to reconsider depending on how cost-effective you think animal interventions are at present.
Feedback and thoughts would be very welcome. Because of I think human happiness is neglected I've working on a happiness app, Hippo, that I'd also be delighted to talk to people about (note: you might think this makes my above argument biased, or you might think is me trying to be consistent. Up to you)
Michael
*I'd like to credit Konstantin Sietzy, with whom I co-wrote an earlier version of the paper. I'd like to thank Hilary Greaves (my PhD supervisor), Michelle Hutchinson, John Halstead, Hauke Hillebrandt, James Snowden and probably other people I've now forgotten for their helpful comments.
[Has anyone from GiveWell looked into mental health interventions? I couldn't find an intervention report on their website but I'd be interested to know whether they have any informal take on it.]
At first blush this is pretty intriguing, especially the following points:
Other things this makes me wonder:
BTW, one note on the paper: you remark that "[a billionaire] should also run randomised controlled trials to assess how much happiness is increased by anti-poverty and anti-malarial interventions"--in fact, you can achieve a lower bound on the happiness increase of anti-malarial interventions because the main mechanism by which they reduce DALY burden (at least in GiveWell's cost-effectiveness analysis) is by reducing mortality. Unlike severe pain, one cannot hedonically adapt to being dead, so anti-malarial interventions (and other mortality-reducing interventions) should have less of the 10x bias than e.g. cash transfers.
*I'm not incredibly confident in this argument; determining the actual quality of life burden here seems like a pretty subtle measurement problem of which I'd love to see a more thorough treatment than the paper provides, since it's really the crux of the quantitative argument.
From what I can tell, the problem is more with outreach and retention than with effectiveness. Most of what I've read shows that computer based cognitive behavioral therapy (cCBT) is as effective as in-person CBT for anxiety and depression in the context of a RCT. But "in the wild", rates of adherence drop considerably, with estimates of 0.5% and 1% completion in the only two published studies I could find [1,2].
If your server time and ongoing development costs are low enough, though, cCBT could still be a cost effective approach despite poor retention. This assumes that those that fail to complete the training aren’t harmed, but evidence seems to suggest that even partial completion is helpful [1,2]. Note that in study [1], about 15.6% completed 2 or more of the 5 modules, so a larger portion of people at least partially complete the training. I haven’t done a $/DALY estimate, but it would be fairly easy to come up with one with the results from study [1].
[1] A Comparison of Changes in Anxiety and Depression Symptoms of Spontaneous Users and Trial Participants of a Cognitive Behavior Therapy Website. http://www.jmir.org/2004/4/e46/
[2] Usage and Longitudinal Effectiveness of a Web-Based Self-Help Cognitive Behavioral Therapy Program for Panic Disorder. http://www.jmir.org/2005/1/e7/
Yes, but not formally. I'll ask Howie if he'd like to comment on this post.
Hello Ben.
A couple of comments:
I wouldn't expect people to be able to adapt to severe pain, not when you consider the evolutionary advantages of always taking your hand out of the fire. I'd expect people to die before they got used to pain.
What is going on is that mental pain may have a bigger impact on your happiness then physical pain and more than we imagine it does. I.e. chronic depression is worse than chronic lower back pain.
(You might reply that this is unfair because mental pain and happiness are basically the same thing: i.e. it's obvious being unhappy has a bigger impact on happiness than just being in pain, so you've just measured the same thing twice. What you'd really want is data which showed the impact different health states have on people's emotional experience/moods (which is what I take happiness to be). Nevertheless given that depression/anxiety seems to be lots of negative mental states, whereas chronic pain isn't, that's still a point in favour of depression/anxiety being where the unhappiness is.)
And yes, so I think depression, which already looks bad on DALYs, is much worse even than that.
Also, it seems that mental health issues are all over the world in a way that, say, malaria is quite concentrated. That's why I say it's possible mental health interventions may be more effective in developed rather than developing countries - people have more technology are greater familiar with mental health.
I can't tell you what the 'in the wild' effect size is because I don't know it and I don't think it's been tried. That's why I suggest a billionaire tests it to find out! The evidence is the CBT works (remedies about 50% of cases of depression) so I'd say the challenge is more getting it to people and getting them to use it.
Developed world happiness interventions? I'm not sure what you mean. Some people in some governments are beginning to think explicitly in terms of happiness, but it hasn't really caught on.
On the death thing, we have different intuitions. In your parlance, I'd say you adapt totally to being dead: there's no you after death for anything to be good or bad for! So all this analysis is very sensitive to philosophical issues.
Sorry. Severe pain may have been a bad example. Other high-DALY-weight conditions do seem to show hedonic adaptation though, e.g. paraplegia (see my response to Lila for sources).
"hedonic adaptation applies to severe pain"
I find this implausible. Where's the citation?
Sorry. Severe pain may have been a bad example. However, for instance, paraplegia does exhibit hedonic adaptation (source) despite having a disability weight of 0.57 (source).
A meta-analysis seems to contradict that (as well as claims in the OP): http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3289759/
Good find. Should have known better than to trust well-established pysch findings. (sob) Thanks for the correction, I'll edit the OP.
I'm not sure what you think this meta-analysis contradicts. Could you please be more precise?
Card on the table, I'm more interested in 'affective well-being' than 'cognitive well-being' as they call it - i.e. 'happiness' rather than 'life satisfaction - and I take the meta-analysis as being broadly in my favour.
I’m glad to see some discussion of this topic here, I think it could be a pretty effective area for EAs to work. I have a few comments specifically related to electronic delivery of therapy. I’ve been following the area for awhile, although most of what I’ve read is in the context of anxiety and depression treatment so it might not be applicable to interventions focused on general happiness.
cCBT is as effective as in-person CBT for anxiety and depression in the context of a RCT. But when you change over to open access therapies, rates of adherence drop considerably, with estimates of 0.5% and 1% completion in the only two published studies I could find [1,2]. If your server time and ongoing development costs are low enough, though, cCBT could still be a cost effective approach despite poor retention. This assumes that those that fail to complete the training aren’t harmed, but evidence seems to suggest that even partial completion is helpful [1,2]. Note that in study [1], about 15.6% completed 2 or more of the 5 modules, so a larger portion of people at least partially complete the training. I haven’t done a $/DALY estimate, but it would be fairly easy to come up with one with the results from study [1].
One promising approach to improve retention is to offer health coaches, which interact with cCBT users and help them stay on track to completion. This would be more expensive, but could be a middle ground between cCBT and in-person therapy. Ginger.io is one startup using this approach, and I’m excited to see how things go for them. They offer cCBT, health coaches, and psychologists via video conferencing if needed. This approach could make it pretty seamless for those with mental illness to seek help. There are a few clinical trials testing their technology here, but I can’t find any results yet.
For a good overview of some of the other emerging startups in this space, see this article. It’s especially encouraging to see people with very strong academic credentials founding or on the boards of these startups, which suggests there is fairly good scientific support for the approach. If you want to read more of the literature, the faculty profiles at Australia National University e-health group and cbits at Northwestern are good place to start. ANU’s moodGYM has been around since 2001, so it has been tested in a number of RCTs.
How could effective altruists help in this area?
Now that a number of promising cCBT companies exist, their outcomes might be inevitable. But EAs could still help the therapies spread more quickly, fund RCTs to verify or improve effectiveness, or work directly for these research groups/companies. On the regulatory side, each state in the US has different licensing processes for mental health professionals, which prevents them from video conferencing with patients in other states. Relaxing this barrier would be especially helpful for rural patients. Getting a cCBT approved for Medicare/Medicaid in the US would also be a step forward, but I would think that stronger randomized evidence would be needed before that would happen. One interesting side note is that the UK, Australia, Denmark and Sweden found the evidence strong enough to approve cCBT years ago, so maybe the problem is that nobody has lobbied hard enough in the US?
[1] A Comparison of Changes in Anxiety and Depression Symptoms of Spontaneous Users and Trial Participants of a Cognitive Behavior Therapy Website. http://www.jmir.org/2004/4/e46/
[2] Usage and Longitudinal Effectiveness of a Web-Based Self-Help Cognitive Behavioral Therapy Program for Panic Disorder. http://www.jmir.org/2005/1/e7/
[3] The Law of Attrition. http://www.jmir.org/2005/1/e11/
[4] Adherence in Internet Interventions for Anxiety and Depression: Systematic Review. http://www.jmir.org/2009/2/e13/
Phil,
this is all really helpful. thanks for the links to the other pieces of research and to a couple of eCBT companies I hadn't heard of yet. Also interesting to see that the challenge is substantially behavioural.
I do hope this outcome is inevitable too, and I hope I can speed it along, rather than uselessly duplicating the work of others.
Yeah I think “inevitable” might be an overstatement, but there do seem to be some pretty promising companies in the area of cCBT for depression right now.
I know less about the apps focused on happiness. Their completion rates might be closer to those of open online courses (~7% on average) because the users might be more motivated. I think building a support community around the app could be important, maybe with users coaching each other? Duplication of effort isn’t necessarily a bad thing at this point because a lot of different approaches are needed to find the right combination of technology/content/support.
I was in contact with Michael before, and let me first say I'm happy he promotes the focus on IHI vs EHI in the EA community.
However, I disagree on how to think of IHI's. I've been struggling with how to think of human suffering since I learned about EA, and it seems to be caused by different views on human suffering between philosophers and what I've learned from clinical psychology, mainly by more pragmatic contextual behavioral sciences (not as an authority argument, but FYI I'm a clinical psychologist/CBT-therapist/PhD-student).
My argument boils down to these 2 points: 1) The premisse and constructs of the EA movement already causes a bias towards Positive Psychology and 2) there are better potential alternatives out there.
So I believe this is a false distinction, and there is great added value in using a more pragmatic paradigm that can be considered as a third option. And that is: building contexts wherein people can live value-based lives, and preventing avoidable psychological suffering.
Yes it is an arbitrary distinction, and contributes to the bias towards positive psychology. I don't see a reason why to make a distinction: humans are humans, and with the knowledge of how high prevalence rates of psychopathology are, it's more logic to assume that underlying mechanisms are present in each of us. The danger of making this distinction is that you end up with interventions targeting the 'ordinary human unhappiness' and not taking into account what these interventions do with people higher up the continuum of suffering. And that's exactly one of the criticism positive psychology receives.
This bias becomes more explicit in the article when Michael describes branches of psychotherapy.
If it is to a much lesser extent, then why acknowledge Positive Psychotherapy? NICE guidelines (UK) and APA guidelines (US) don't regard positive psychology interventions as evidence based. (And by the way, classical CBT and MBCT can be regarded as just being part of the happy family of CBT). Another problem is the word 'methods'. Before we start thinking of methods, we need a theory on human suffering, so that when we think of interventions we don't just start from constructs like happiness. This discussion is very alive in the CBT-family, because of the rise of another branch: Acceptance and Commitment Therapy (ACT).
So EA'ers interested in this topic, please read on ACT and the underlying theory of it (Relational Frame Theory, RFT). Before we use numbers, we need a decent theory on suffering to frame them.
If this topic is still alive, I'll try to write another post on how effective altruism based on contextual behavioral sciences might look like. To say it very briefly, a distinction between two sorts of interventions is needed:
By the way, I e-mailed this before to CEA after attending the ABCT-conference in New York. ABCT= Association for Behavioral and Cognitive Therapies (US). Maybe interesting for some of you:
*Given the fact I heard a lot of ambitious attempts to reduce human suffering the last couple of days, I realized I haven’t encountered these voices in the EA movement yet. Maybe these suggestions have been made before, but I’ll give it a try anyhow.
I make these suggestions as speakers, because I think none of these CBT-interventions are ready to compete with the most efficient ways of reducing human suffering. But still, I guess evidence-based talks on what human suffering is and how to reduce it, are still interesting for EA conferences.
1/ Steven Hayes Stubborn bigshot in CBT, founder of Relational Frame Theory (RFT) & Acceptance and Commitment Therapy (ACT). https://en.wikipedia.org/wiki/Relational_frame_theory Had an enormous impact on how CBT-therapists and researches view human pathology and suffering. He’s a fantastic speaker. And a nerd, I guess he’ll love EA and is able to make a great talk, for example on why human suffering differences from non-human-animals (language!), and what to do about it.
2/ Michelle Craske President of ABCT. Presented this very ambitious project today: http://grandchallenges.ucla.edu/depression/ “Understanding, preventing and treating the world’s greatest health problem’
3/ Vikram Patel https://www.ted.com/talks/vikram_patel_mental_health_for_all_by_involving_all?language=nl But apparently this link already exists a bit https://www.givingwhatwecan.org/post/2015/12/mental-health-interventions-may-be-more-cost-effective/
4/ David Clarke https://www.psy.ox.ac.uk/team/david-clark His work on IAPT is great: implementing evidence based care in UK to reach out to a lot of people. And changing the culture into one where data-collection is a great part of it. https://www.penguin.co.uk/books/184573/thrive/*
I think IAPT is the best way to go for systemic change on human psychological suffering right now. It probably has the biggest added value. Having those meta-organizations like NICE in the UK, the implementation of more mechanism-focused therapies and intervention for prevention will follow automatically.
And yes, I'd love to hear any of all of these people talk at EA events.
Hello Kris, good to e-hear from you again. I haven't checked this thread in a couple of months so have only just seen this.
I'm not totally sure what it is you're suggesting we do (instead). You seem to be objecting to positive psych, but I never said it was all about positive psych, just that it was one of a number of tools that might allow us to increase happiness. My main point was that we've neglected internal happiness interventions and we should explore those alongside the external happiness interventions we're already working on.
What sort of theory of suffering might you be referring to? Is that about the nature of suffering, or about what makes suffering bad, or something else?
I'd be happy for you to facebook/email me so we can chat through this if you think that's interesting.
GiveWell recently funded a RCT to test a mindset intervention: http://www.givewell.org/international/charities/ipa/may-2016-grant
There is evidence linking diet to mental health described on http://nutritionfacts.org/topics/mental-health/
"[The intervention] will attempt to encourage cash transfer recipients to reconsider fundamental aspects of themselves, their surroundings, and their future by showing them a brief video and providing interaction with a coach."
Ouch, hope that's not as patronizing as it sounds, though it might be worth testing at a small scale.
Hi Michael,
You mention the Easterlin paradox a few times in these comments and in you draft paper. I briefly looked into the literature a while ago and I place less weight on the Easterlin paradox than you. Here’s a quick summary of what I found.
Easterlin claimed (p.113-118) that average satisfaction in a country doesn’t increase as a country grows wealthier. Since then there has been a back and forth in the literature but there is now a growing body of correlational evidence that strongly conflicts with Easterlin’s initial claim (for instance, p.3, p.4, p.10, p.12). It seems that the literature now suggests that the relationship between income and life satisfaction is one of diminishing returns but that an increase in income is correlated with an increase in life satisfaction. A nice heuristic to use is that a doubling in income increases subjective well-being by 0.34 standard deviations (p.7).
There has been only limited research into the effect of income on emotional well-being or the affective aspect of subjective well-being. The research there has been suggests that particularly at low levels of income an increase in income correlates with an increase in emotional well-being (p.3, p.8, p.8). A meta analysis on subjective well-being noted that there is a weaker association between income and emotional well-being than income and cognitive well-being (p.3).
hello Keiran and thanks for your comment.
I don't discuss the Easterlin Paradox in any depth in the paper because it was largely tangential to the point I was making. It's really interesting and something I've thought about a lot.
Whether you think the Easterlin Paradox is correct or not somewhat depends on what you think it shows in the first place. There are different formulations of this. My reconstruction is that the Easterlin Paradox makes three claims:
Richer people within a country more satisfied than poorer people.
Richer countries more satisfied than poorer countries
As countries have got richer, life satisfaction has remained broadly flat.
What makes the Easterlin Paradox interesting is that the lack of evidence for 3 seems weird given the truth of 1 and 2. That's the paradoxical part: if being richer than other people at particular moments makes us more satisfied, why don't countries get more satisfied if they get richer? Isn't more money always better?
Now, I should point out that no one doubts the truth of 1 or 2. To my eyes, the battle ground is about the 3rd point: does growth increase life satisfaction? There's some dispute over whether it does, but at best economists only find there is a tiny difference e.g see Stevenson and Wolfers. Perspectives can disagree, but I take that a victory for the Easterlin side: if growth does matter, it seems pretty trivial, so lets focus on increasing satisfaction by other means.
(You can also get into a tedious quagmire of how to best assess 3 given the data available. Stevenson and Wolfers look at growth, but it's not really a surprise that growing/shrinking GDP increase/decrease satisfaction. What you really want is long-run growth, which Beja 2015 looks at and concludes that, if long run growth increased by 1%[per year], then life satisfaction would rise from 3 to 3.0027" (life satisfaction was measured on a 1-4 scale) which I mention in reply to Ben Todd below)
So I'm not doubting
But that refers to point 1, not point 3. What seems to be going on is that income has a large relative effect and a small absolute effect on life satisfaction. The fact that richer people are more satisfied than poorer people doesn't allow you to infer than making those poor people richer would increase their happiness. Analogously, if I change who wins a 100m sprint so yesterday's loser becomes today's winner then it's not obvious (but still possible) I've increase overall satisfaction with the result.
Anyway, this is all analysis in terms of cognitive, not affective well-being. I think the latter is what really matters. The Kahneman and Deaton 2010 paper is really interesting because it shows life satisfaction (cognitive well-being) goes up with income but 'happiness' (affective well-being) plateaus at $40,000ish for household (not individual) income. Given that's a survey conducted in America we might suppose this figure would be far lower elsewhere. It seems unlikely to be the case that $40,000/year/household is the figure beyond which incomes stop affecting happiness for all times and all places. If you conceptualise the relevant figure as "the ability to shameless participate in society" then you'd expect that to change.
As a result, I think more work is required to find out what the level of absolute income is that people require. As the Give Directly study shows, it may be incredibly low and so low that cash transfers do surprisingly little.
When you mention the $1000/DALY by "Stronger Minds," are you referring to strongminds.org? I asked them if they had a cost estimate for DALYs but never received a reply. If it does refer to StrongMinds, do you know if they have predicted a cost per DALY once there viral group therapy model of treating depression grows significantly? Mayberry says in his TED talk that he expects it to become cheaper as it grows.
Also, does your discussion of DALYs for mental health interventions only include YLD, or also YLL? I would think there could be a large difference between the two considering the huge impact depression has on morbidity and mortality (comparable to obesity).
Are you saying you didn't check footnote 45 of my paper? Outrageous! The $1000/DALY figure comes from
"Patel, V et al. (2015). Addressing the burden of mental, neurological, and substance use disorders: key messages from Disease Control Priorities. The Lancet. These figures should be treated with caution. As the authors note, p1681 “hardly any published evidence exists on the cost-effectiveness of population-based or community-level strategies in or for low-income and middle-income settings" "
I'm afraid I don't know exactly who conducted the studies they refer to.
My criticism of MH intervention is more in terms of the way QALYs don't seem to capture YLD in the same way a happiness-based approach would. I'm afraid I don't, but probably should, know how Givewell's $100 malaria figure is split between YLD and YLL and how the $1000 depression figure is split between YLD and YLL. I actually found it incredibly hard to find that sort of information, maybe because I'm a health economist and don't know where to find it. If you know, please tell me!
In retrospect, I'm probably not careful enough in the paper and I'll have to re-write it for the next draft. I say that QALYs underrate mental health by quite a bit, and on DALY estimates malaria is 10x cheaper, so maybe mental health treatments are in the right sort of ballpark already. Really I'd want to know how much depression and malaria each reduce happiness ('happiness' to be more carefully specified) whilst sufferers as still alive, and how much they increase mortality. Then you can do a more principled cost-effective analysis where you plug in how bad you think being dead is for someone. Unfortunately no one seems to be in the business of measuring health in terms of experienced happiness, so we may not be able to try and answer this question for some time.
Hi Michael! As I said before, congrats on an interesting paper.
A few points on this comment:
1) DCP3 didn't have any cost-effectiveness figures for the StrongMinds intervention (interpersonal group therapy). Is the $1,000/DALY figure you mention related to primary care advice on alcohol use?
2) I'm currently writing a piece on mental health for a HNW donor and tried to model c-e of StrongMinds. I got c.$650/DALY reducing to $400/DALY as intervention scales. The biggest uncertainty in this estimate is the long term effects of psychosocial treatment as hardly any evidence exists. (I will post the calcs later - they're on another computer)
3) Givewell's estimate ignores YLD and is only based on U5 child mortality. So it's entirely YLL. You can find the calculations here: http://www.givewell.org/international/technical/criteria/cost-effectiveness/cost-effectiveness-models
Hello James.
in reply:
1) check figure 3 on p1681 of DCP3. The box says "Depression: episodic treatment in primary care with (generic) antidepressant medication and psychosocial treatment" at the box implies it's between $1000/DALY and $10,000/DALY. I'm not sure exactly where those numbers came from. Have I misread that?
2) I'd be really keen to see your calculations and how you're putting it all together!
3) Can you point out where in their spreadsheet they specify that? I've just spent 20 minutes looking through it and am a bit lost.
My estimate is then likely a miscalculation - thanks for pointing that out! What I should have done is use YLD figures for malaria and depression, and then transformed the disability rating using subjective well-being numbers.
Then I'd need to say something about the badness of losing a life. I think this indicates the perils of using DALY numbers...
1) Ah yes - thanks for pointing out. Probably has limited external validity for the strongminds model though (which is psychosocial treatment alone for most patients delivered by community health workers, with only the most serious cases referred to clinics for medication). The numbers come from the Chisholm (2015) WHO-CHOICE model. http://www.bmj.com/content/344/bmj.e609
2) Analysis is here https://docs.google.com/spreadsheets/d/1-lCC1zQHVZlJS8f9OfqhzcZTetHMxuMkW7nT75QDGhk/edit#gid=960072536
[This is quick and dirty but gives a rough indication of cost-effectiveness. Most uncertain assumption is the long term impact of interpersonal group therapy on treated individuals 1-10 years down the line]
3) On the 'bednets' sheet you can see that the output measure is cost per under 5 child death averted. DALYs are then backcalculated from this to get c.$100 [not in sheet] . Something like $3,500 / 50 years of life for each death averted = c.$70/DALY. Because they're only looking at deaths, it's YLL not YLD. I haven't seen a quantitative estimate of the total morbidity burden of malaria. One important consequence of surviving severe (cerebral) malaria is a much higher chance of getting epilepsy later in life http://www.ncbi.nlm.nih.gov/pubmed/25631856 although I suspect there are many others. Child health is really important!
Also - could you specify what you mean by mental health being 10-18 times worse than we think. Does this mean: a) DALY weighting of severe depression is 0.65. Actually it should be 6.5 (so 6.5x worse than death. seems implausible) or b) Life with severe depression is worth 0.35 of healthy life. Actually it should be 0.035 (so 1 year of healthy life is worth c.30 years of life with severe depression. maybe but this seems like a lot)
Hi James!
Your comments to this blog post are very much appreciated - thank you for contributing.
We are a group of students at Oxford University doing a research project where we are trying to find the most effective charity to donate to (see https://oxpr.io/). We are currently looking into StrongMinds, and found your helpful cost-effectiveness model. If we may ask, we were wondering if possibly you have a more up-to-date / complete version of the same? And possibly also ask which inputs you received from StrongMinds vs. inputs estimated (and if so, how those estimates were made)?
Completely understandable if you don't have time to answer all questions, but truly any feedback from you at all would be greatly appreciated.
Please do let us know. Many many thanks!
1) Yeah, I never thought the numbers were that robust. More good measurement needed!
2) Thanks for this, will check this out soon.
3) Thanks
4). By being 10-18 times worse than we think, I mean anxiety/depression may cause about 10-18 more suffering than people expect them compared to other health conditions.
This is from the the Dolan and Metcalfe paper: they show people are prepared to trade off 15% of life to remove 'some difficulty walking' and 'moderate anxiety or depression', but that people with 'moderate anxiety or depression' report 10x the reduction in life satisfaction that those with 'some difficulty walking' do, and 18x reduction in terms of daily effect (their measure for what we might call 'happiness').
In other words, the average person imagines walking with a limp would be bad as moderate depression (as inferred from trade offs), but actually the depression would be much worse for their happiness than the limp. This is explained by the focusing illusion and the non-adaptation stuff.
If you look at the other numbers in the Dolan and Metcalfe paper, they show 'self care' and 'usual activities' are equally over-rated when people the trade-offs compared to how much they effect happiness.
In terms of weighting, my thought is that if we constructed well-being adjusted life years (WELBYs) depression would be 10-ish times worse than walking with a limb, but this would be rescaled. So if depression has a WELBY weight of 0.8 (e.g. 1 year with depression is worth 0.2 years of happy life), then minor mobility issues have a WELBY weight of 0.08 or something. I would not suggest all cases of mental health should be understood as being many times worse than death! I don't think I implied that anyway, but I would clarify that in future.
The overall thought is more like: daily life with depression is quite bad in terms of happiness, health conditions which don't cause depression (or pain) at all (or for very long) are probably not nearly as bad as we imagine they are, and we should re-prioritise bearing this in mind. Non-depression mental health disorders may also turn out to be much worse than we expect, and maybe also worse than most physical health conditions. This is all a bit broad ("what does 'most physical health conditions mean?'") but I hope you get the point.
My gut reaction to this is pretty negative, and I think it would be bad for EA PR if this were shared widely. How it reads to me is "Extreme poverty isn't that bad. Those people get too much attention - first-world people have problems too!"
The somewhat more defensible version is "Extreme poverty is pretty bad and causes mental health problems. It's hard to fix the causes of these problems, so let's just try to treat the effects, by giving people Prozac and teaching them meditation." From a strictly utilitarian perspective, we should be fine with, for example, slaves who have been drugged to enjoy slavery. But most people are kind of squeamish about this.
Many people underestimate how bad global poverty is, because they don't understand the violence and exploitation that the poor experience. For example:
There's evidence that these all take a huge toll on mental health. Poverty is arguably one of the most significant causes of mental health problems.
I recommend reading Behind the Beautiful Forevers to get a better understanding of daily life in a slum.
I realize that my versions of your argument are straw men, but I'm explaining why this comes off so negatively. It's fine to say you think that mental health is important, but glibly comparing it to poverty (while promoting a happiness app) reinforces the worst stereotypes of EA as out-of-touch Silicon Valley people.
I agree with the worry about framing, but I downvoted due to negative tone of the first paragraph – it makes the forum unwelcoming to a new user.
Thanks for the comments. I was wondering how long it would be before someone said this! I'm slightly sympathetic to the 'this is bad for PR' argument. But two points on that:
this is mostly a question of framing. It's not 'poverty is great' it's 'mental health is really bad, even worse than poverty'.
I think there are virtues to honesty. If we self-censor and don't try to work out how to do the most good just because we think somehow, somewhere, somebody is going to disagree, then we're almost certainly not going to work out how to do the most good.
More generally, I don't think your "Extreme poverty is pretty bad and causes mental health problems" version is true. As you can see from this: (http://cep.lse.ac.uk/layard/thriveannex.pdf) there's only slightly more depression in poor countries in low-income countries (7%) than high-income countries (6.5%). Layard calculates the averages in his book Thrive (p.41). So that undermines the idea that poverty causes mental health issues and removing people from poverty would fix them. Mental health seems to be a problem of being a human anywhere.
I think a more promising argument, which I make in the paper, is that it might just be a lot cheaper to treat mental health in poorer countries. I also take the utilitarian view that misery, not injustice or anything is, just is the problem. There's some discussion about whether teaching people to be happy is the right long-term solution even in terms of happiness, but that's an empirical question that needs investigation.
You also seemed to have merged mental health, poverty and failed states. I'm not sure how to respond to that one.
In response to Behind the Beautiful Forever, you might want to check out this paper - Making the Best of a Bad Situation: Satisfaction in the Slums of Calcutta (http://www.jstor.org/stable/27526958?seq=1#page_scan_tab_contents) which shows those in slums "experience a lower sense of life satisfaction than the more affluent groups, but are more satisfied than one might expect". I'm somewhat disinclined to take anecdotes or stories as serious evidence: they tend to focus on the interesting parts of life, whereas the upshot from the happiness literature I've seen in that most people's lives are, well, pretty ordinary. That is, unless you've got a mental health disorder. Then your life sucks.
First of all, let me say that I agree that mental health is very important. But I think you're approaching this in a culturally insensitive way that excludes the experiences of many people.
"'mental health is really bad, even worse than poverty" Yeah that's the type of statement that sounds super cringey (though I assume you mean mental illness, not mental health). When I saw your post, I was like, "Oh I'll bet anything this is based on his own experiences with depression." The link to your app confirmed this. While your experience was unfortunate, it can blind you to the reality of other people's suffering.
I'm very skeptical of the cross-cultural validity of depression diagnosis, particularly when it involves self-reporting.
Violence and disempowerment is not only a problem of failed states - it's a nearly universal experience for the poor. You might also want to read A Plague of Locusts which describes the everyday violent crime experienced by the poor, even in "middle-income" countries such as Peru, which has one of the highest rates of sexual violence in the world. The book also presents more statistical evidence for how widespread mental health problems are among crime victims in the developing world.
The idea that people in the developing world are mentally unaffected by everything from police abuse to unprosecuted rape and murder (when we know that these are traumatizing to people in the developed world) has unfortunate echoes of old racist stereotypes. Nineteenth century explorers were often surprised to see indigenous people crying: whites had assumed that non-white races were more resilient and less emotional. (Think of the "stoic Native American" stereotype.)
Great point about the cross-cultural validity of depression diagnosis.
For that matter, I'd be awfully concerned about the cross-cultural (or cross-socioeconomic-group!) validity of life-satisfaction measures. Often they are asked something like so:
There are obvious ways in which this question might cause someone to give, say, their life satisfaction as a percentile compared to people around them, rather than an absolutely comparable number, which would bias it up a lot for poor countries.
I'm not sure what an absolutely comparable number would be: people would have to be comparing themselves to the same unchanging criteria over time. The evidence, from the Easterlin Paradox etc. is that people do change their standards over time and largely seem work out how they are doing my comparing themselves against others. As such it looks like increasingly worldwide life satisfaction would be very hard.
I take these sorts of argument as reasons to move away from life satisfaction towards direct measures of people's experience in the moment. I want to know how good or bad the person actually feels, not how well they they are doing against an arbitrary and changing standard.
Just a small note: the book you refer to, "A Plague of Locusts", is actually called "The Locust Effect" (Haugen and Boutros) - took me a while to find it, so figured that others might appreciate the correction :) Thanks for the recommendation.
I don't think my own experiences are terribly important and I'm much more interested in arguing about the points themselves.
I'm not sure how you can disagree with the statement that mental health disorders are worse than poverty in terms of happiness; that's true simply in virtue of what they are. Being depressed means you experience lots of negative mental states (i.e. are unhappy). Being poor doesn't mean that. Actually being unhappy has to be worse than being a state that is likely to make you unhappy.
Putting this the other way round, you could say "depression is bad for your income, but being in poverty is worse". Depression might cause you to earn less, but if you're looking at income, then the state of having a low income has to be worse than having something liable to cause you to have low income. It's also not true that all people in poverty are depressed. So I think it makes more sense to target misery, not poverty.
Do you think I've misunderstood your point? Sorry if I have.
I'm also very much in favour in targeting violence, crime, etc. as those seem obviously bad for happiness. However I'm not sure how bad, nor do I have a potential solution. On that note, the problem of adaptive preferences, as discussed by Sen, etc. is quite interesting. He argues that the poor adapt to their terrible condition and thus they seem surprisingly happy - they are making the best of a bad lot - and that is an additional reason to do something.
I take the other line: if I could save person X who will be made happier, or person Y who won't be because person Y will adapt to the new condition, I'd want to treat X, not Y.
I feel like this is a motte-bailey argument. Sure, in some trivial sense "depression" is what everyone cares about, in that everyone wants to prevent suffering. A chicken on a factory farm is "depressed". An acid attack victim is "depressed". But it seems like when you say depression, what you really mean is "feelings of lethargy, apathy, and discouragement that afflict many people (particularly Westerners) at some point during their lives" - this is what most people understand depression to mean. It's certainly arguable whether this is worse than many other forms of suffering, and it feels a bit arrogant to glibly dismiss other experiences.
Furthermore, you continue to use "mental health disorder" as synonymous with "depression". As I asked before, where's your solution for schizophrenia?
I'm not sure how you've reached your conclusion on the basis of what I've said. I'm taking 'depression' to be a mental health disorder with certain symptoms, including substantially lowered mood. I'm not using 'depression' as a synonym for 'feeling a bit bad'.
I'm making the claim that depression, which is constituted by lots of negative emotional states (i.e. states that feel bad to the person) is probably worse that many, if not all, other forms of suffering in terms of happiness (your 'happiness' = the sum total of momentary positive mental states less your negative ones). In part this is because depression induces intense negative states, in part because we don't seem to be able to hedonically adapt to the condition in a way we can to, say, becoming paraplegic.
I'm not glibly dismissing other cases of suffering. I think you seem to be objecting to doing all things considered evaluations of how bad various things are. If you don't want to do all things considered evaluations it's quite hard, if not impossible, to make important moral choices.
Depression is just one of a range of mental health disorders, that's true. Where I've used them synonymously that was simply careless and definitely my mistake.
And I don't have solution for schizophrenia, nor did I realise I was supposed to! I'm a philosopher who researches happiness, not a psychiatrist.
I think the claim that depression "is probably worse that many, if not all, other forms of suffering in terms of happiness" is far too strong and you haven't provided sufficient evidence for it. As you said, you're not a psychiatrist, and I think you're relying too much on a priori (and somewhat tautological) reasoning, evidence from self-reporting of questionable validity, and outdated research on adaptation (see the meta-analysis that I linked in another comment). There's a case to be made for your position, but it needs to be strengthened to justify your sweeping statements.
The reason I brought up schizophrenia is because it felt like you were moving the goalposts, not because I expected you to have a solution. You brought up aggregate mental illness statistics but then only discussed depression. But thank you for admitting the mistake.
Just to add to this. Acute schizophrenia is one of the worst health conditions on GBD13 DALY weightings (c.0.8). Severe depression is also one of the worst (c.0.65).
See http://www.thelancet.com/action/showFullTableImage?tableId=tbl2&pii=S2214109X15000698
So Michael - I agree it's very possible that mental health disorders are underweighted by DALY weightings because of the focusing illusion. But they are actually weighted quite highly at the moment. 10 years with severe depression is worth approximately 3.5 years of healthy life.
yes, it's interesting that schizophrenia tops the list anyway, although this could be the case in virtue of the face it stops you from leading a normal life part of the disability (this being based on the EQ-5D), rather than because severe schiozophrenics are less happy than severely depressed people.
although they are weighted highly, that doesn't stop them from being underweighted. Given the way DALYs are constructed - measuring health, not happiness; using preferences, not adaptation - it's conceptually very hard to see how mental health conditions can't be underweighted in terms of happiness. Unless and until we measure people's experiences of various diseases we really won't know.
My guess is that depression could be the most comparatively underweighted health state: if you can function normally, but you're just really sad, you might do pretty well on an EQ-5D metric because you're only doing badly in 1 of 5 criteria, and that's the criteria most linked to happiness.
As an aside, do you know how the GBD project takes episode duration into account? Or is it just a measure of intensity?
1.5./6.5 = 23% difference. That is statistically significant under any reasonable assumptions.
I agree that the way it is presented can cause an instant dislike.
When I've talked about effective altruism to people that haven't heard about it, I talk about global poverty and then also mention that there isn't much focus on mental health in any country in the world and that we need to focus on both physical and mental health at the same time.
This usually gets a reasonable amount of agreement and I think if it is presented side by side with increasing physical health/wealth it doesn't sound out of touch.
Good suggestion. I've edited my original post slightly.
Gosh! I'm very late to this party, but have just been sent here by Benjamin Todd via an 80k hours LinkedIn post! I am a psychologist - with feet in both applied (therapeutic) practice and research. Being the age I am, and already in a second career, I have been looking for ways to have more impact given my existing skills/knowledge/experience - which would be in psychological well-being (however that is determined within a given cultural context), and I could niche down more into behaviour change, substance misuse, and addictions. Broadly, I practice elements of cognitive-behavioural therapies.
I certainly think how one measures/tracks well-being/'happiness'/life satisfaction/'the absence of symptoms/the presence of good-health/ etc. is key here. For instance, rather than a scale measuring presence/absence of, say, "depression" (however that is defined within a culture) - I would know more about how the presence/absence of this 'depression' is affecting someone if I knew, for example, how many hours they spent in bed, hours they spent crying, how many times they bathed or did the dishes in the past week, completed tasks they had set themselves etc. Apps I think need to have some way of tracking the behaviours and signifiers for the individual that show they are 'more' or 'less' symptomatic of the 'problem'. This would mean that the user input and tracked more meaningful data - along a more quantified self type of strategy - which would hopefully keep them more engaged with the intervention.
Anyhoo - I shall follow this thread and other discussions around mental health and well-being closely!
kind regards EA community! Helen
Hello Helen,
Thanks for the comment.
If you're interesting in chatting about this further - either for research purposes or if you think you can help me with Hippo, my app - do get in touch on michaelplantuk@gmail.com
I was glad to see this article -- I think it's a very interesting issue, and generally want to encourage people to bring up this kind of thing so that we can continue to look for more effective causes and beneficiary groups. Nice work!
I didn't find the presentation unpleasant, personally, but I have a high tolerance for being opinionated, and it's been helpful to see others' reactions in the comments.
Giving What We Can published a report on mental health in the developing world: https://www.givingwhatwecan.org/cause/mental-health-neurological-disorders/
Thanks for this. I'm already aware of this stuff and, in fact that was written my Konstantin Sietzy, who has worked with me on this and helped co-write an earlier unpublished report into this problem.
I think these are good points, and agree that if you mainly cared about the happiness of presently existing people in the short-term, then mental health seems like one of the most promising causes.
However, I think most in the EA community – even those who don't particularly care about xrisk –support GiveWell recommended charities partly for their short-term boosts to happiness, and partly for their medium-term (1-30 years) benefits to the community the people are based in.
For instance, (this is not my area of expertise, but understanding from reading the GiveWell blog) large scale net distribution or deworming can lead to permanent reductions in disease, producing a long-term of reduced illness (which is not usually included in cost-effectiveness estimates).
Moreover, not having malaria or worms makes it much easier to get education and work, which should make people wealthier. Then that benefits your children and people you buy good from, making them wealthier, and so on, contributing to economic growth.
GiveWell often mention the benefits on income and education as a bigger reason for funding deworming than the QALY gains.
Mental health interventions should have positive flow through effects too but they seem less well understood. Though I could easily see myself being persuaded here - I would guess depression has a pretty terrible effect on your income and so on. On the other hand, it seems much harder to eradicate once and for all, which is possible with health.
There are other reasons for supporting health interventions interventions too – such as the strong degree of evidence behind them, measurable results, ability to mobilise large numbers of people to support them, expert support, clear room for funding and so on. They look promising from many perspectives.
Hello Ben and thanks for your comments!
I think I'm substantially more pessimistic than you are about the role economic growth has in making people happier ('happiness' conceived either as life satisfaction or emotional experiences). So I buy the story that Give Well charities make people healthier, this allows them to get an education, this helps them become richer and so help their children, but I'm sceptical that any of this increases happiness.
For instance I came across this analysis of the Easterlin Paradox yesterday (http://bit.ly/28XXisi) which looked at GDP and life satisfaction scores of European countries over the last 40 years. I emailed the author for some clarification and he explained "if long run growth increased by 1%[per year], then life satisfaction would rise from 3 to 3.0027" (life satisfaction was measured on a 1-4 scale). So he thought economic growth was statistically significant but not interesting as far as life satisfaction is concerned.
All this seems quite surprising. As far as I know there is lots of evidence suggesting economic growth doesn't do much/anything for happiness - including the Haushofer (2015) paper into Give Directly, but almost none indicating growth would increase happiness. If you know of any, please send it to me, because this seems strange.
As a debunking explanation for why we think money matters when it doesn't, I think it helps to remember we're not very good at affective forecasting. When we think about wealth it seems important ("I'd be so much happier if I were on a yacht right now") but we forget we'll adapt to whatever it is, focus on other things, and that our being wealthier will make other people feel comparatively less wealthy.
I agree there's more work to be done in terms of comparing the flow-through effects (I'm still not really sure what these are - is this just about economic growth? Could you provide a link to where I can read about them?) of mental health to physical health. A quick scan suggest the former is also bad, but I'm not sure how bad. This is probably the most relevant point for those concerned about the very long run. On this one, I think I'm probably more optimistic than you about how low the low-hanging mental health fruit are. AFAIK, no one has really tried to pick them and find out.
Regarding your last paragraph I'm not sure physical health interventions do a whole lot better than mental health ones. I think physical health is a lot easier to imagine and show progress on because you can see it. One of my general concerns is I don't think we currently do a very good job of measuring human suffering (i.e. via QALYs) and we're relying too heavily on our intuitions about the imagined badness of various things. I'm also hopeful that, with digital technology, we may find it's much easier to treat mental health because those often receive cognitive rather than physical interventions. Maybe at some point in the future we're consider physical health problems, which require a real human to help you, as the complicated ones.
Somewhat related - what about finding biochemical/psychological methods of inducing erotic love? Isn't that a potentially very cheap way of improving quality of life?
I've discussed this briefly with Julian Savelescu, a philosophy professor here at Oxford, and I think he said he was working on it. I don't know much about the topic and it's sort of a different problem from the one I'm looking at here (dealing with one's own emotions). i'm sceptical but happy to be proved wrong.
Where do non-depressive mental illnesses such as schizophrenia fit in here?
In terms of badness: my guess is they are also bad, although I don't know what research has been done with happiness measures. Speaking to some psychiatrists, I gather it's possible that some cases of schizophrenia don't have to be bad (e.g. you hear voices but it doesn't bother you) but often they are.
In terms of treatment: different. I gather schizophrenia requires at least different types of CBT from those used to treat depression, and typically anti-psychotics, so it might be quite a lot less susceptible to the approach I suggest, but I'm not really sure.