Is trauma a potential EA cause area?

post by nonzerosum · 2019-05-31T22:04:17.690Z · score: 2 (8 votes) · EA · GW · 14 comments

This is a question post.


    3 Jemma
    3 Milan_Griffes
    2 aarongertler
    2 Khorton
1 comment

I'm referring to general traumatic experiences – both the ones that might more typically come to mind like witnessing a violent event and things like this, but also more "microtraumas" like getting bullied at school.

My hypothesis might be that reducing trauma could improve human co-operation and coordination. The logic might be that trauma means that when people communicate they accidentally "trigger" something below the surface in the other person, and so then the conversation becomes about something that it's not ostensibly about, and cooperation becomes harder.

You could imagine an experiment where there are X "trauma-free" people and a varying number of people with some level of trauma doing a task that requires cooperation and coordination of all participants to solve, and then comparing the "time-to-solve" and success rates, to see how much coordination and cooperation might be improved by solving trauma.


answer by Jemma · 2019-06-04T10:20:55.594Z · score: 3 (3 votes) · EA · GW

To put a brighter spin on what other people have said about tractability, many EA-backed cause areas reduce trauma already, too. If a child doesn't die of malaria, their siblings and parents are also spared the huge trauma of experiencing the death of a family member.

answer by Milan_Griffes · 2019-05-31T22:26:44.670Z · score: 3 (2 votes) · EA · GW

MDMA therapy is showing a lot of promise as a treatment for post-traumatic stress disorder.

Here's an excerpt from my recent "prizes for arguments against psychedelics being an EA cause area" post:


3(d). Trauma alleviation

Childhood trauma is plausibly upstream of several burdensome problems. See this excerpt from The Body Keeps Score, a pop-sci review of academic trauma research (on p. 150):

The first time I heard Robert Anda present the results of the ACE study, he could not hold back his tears. In his career at the CDC he had previously worked in several major risk areas, including tobacco research and cardiovascular health.
But when the ACE study data started to appear on his computer screen, he realized that they had stumbled upon the gravest and most costly public health issue in the United States: child abuse.
[Anda] had calculated that its overall costs exceeded those of cancer or heart disease and that eradicating child abuse in America would reduce the overall rate of depression by more than half, alcoholism by two-thirds, and suicide, IV drug use, and domestic violence by three-quarters. It would also have a dramatic effect on workplace performance and vastly decrease the need for incarceration.

Psychedelic therapy seems very promising for resolving PTSD, which could plausibly break the cycle of abuse that creates new traumatic experiences. (Trauma appears to transfer from generation to generation via multiple pathways.)

In particular, MDMA-assisted psychotherapy for PTSD is yielding extremely promising results in recent randomized controlled trials (see Mithoefer et al. 2012, Mithoefer et al. 2018, Ot’alora et al. 2018). From the abstract of Mithoefer et al. 2018:

At the primary endpoint, the 75 mg and 125 mg groups had significantly greater decreases in PTSD symptom severity (mean change CAPS-IV total scores of −58·3 [SD 9·8] and −44·3 [28·7]; p=0·001) than the 30 mg group (−11·4 [12·7]). Compared with the 30 mg group, Cohen's d effect sizes were large: 2·8 (95% CI 1·19–4·39) for the 75 mg group and 1·1 (0·04–2·08) for the 125 mg group.
PTSD symptoms were significantly reduced at the 12-month follow-up compared with baseline after all groups had full-dose MDMA (mean CAPS-IV total score of 38·8 [SD 28·1] vs 87·1 [16·1]; p<0·0001).

A Cohen’s d of 2.8 is extremely large (“Cohen suggested that d = 0.2 be considered a 'small' effect size, 0.5 represents a 'medium' effect size and 0.8 a 'large' effect size” source). Here’s a good resource for interpreting Cohen’s d.

In this study, 30 mg of MDMA was used as an active placebo, and the intervention groups were given 75 mg or 125 mg of MDMA.

From Mithoefer et al. 2012, a long-term follow-up of the first MDMA RCT:

We found the majority of these subjects with previously severe PTSD who were unresponsive to existing treatments had symptomatic relief provided by MDMA-assisted psychotherapy that persisted over time...

MDMA helped resolve severe PTSD symptoms in patients who had not responded to other treatment regimens. For 86% of patients, this benefit persisted 17+ months after the MDMA session.

comment by nonzerosum · 2019-06-04T19:09:50.300Z · score: 1 (1 votes) · EA · GW

Yes, agreed. In particular though I'm wondering about the "impact" piece and separate of possible interventions/tractability, how trauma might rate on the "impact" and "neglectedness" pieces.

comment by Milan_Griffes · 2019-06-04T19:11:54.559Z · score: 2 (1 votes) · EA · GW

Don't the studies I point to suggest a large impact?

From my answer:

A Cohen’s d of 2.8 is extremely large (“Cohen suggested that d = 0.2 be considered a 'small' effect size, 0.5 represents a 'medium' effect size and 0.8 a 'large' effect size” source). Here’s a good resource for interpreting Cohen’s d.
comment by nonzerosum · 2019-06-05T23:29:06.086Z · score: 1 (1 votes) · EA · GW

I mean EA impact of reducing trauma, not impact of MDMA therapy on trauma (which I agree seems large).

Similar to how 80000hours gives a ranking of the 'impact' of different causes, I wonder how "reducing trauma" would compare on their impact assessment.

answer by aarongertler · 2019-06-01T06:55:58.466Z · score: 2 (1 votes) · EA · GW

Reducing trauma seems good as something that improves human flourishing
// reduces suffering, more than as something which "improves coordination". But it doesn't stand out to me as something that seems better than any of the major established cause areas.

EA cause areas aren't just about the scale of an issue -- to be plausible candidates, they require methods to address the issue that are proven or at least promising. Are you aware of any extremely efficient ways to reduce trauma? Is trauma something that can easily be measured (maybe secondarily through stress hormones)?

People interested in this topic may find value in Michael Plant's look at mental illness, an area where we have some evidence for the existence of a reasonably cheap and effective treatment.

comment by Halffull · 2019-06-05T13:12:55.455Z · score: 2 (2 votes) · EA · GW
Are you aware of any extremely efficient ways to reduce trauma ?

There are several promising canidates that show high enough efficacy to do more research. Drugs therapies such as MDMA show promise, as do therepeutic techniques like RTM. (RTM is particularly promising because it appears to be quick, cheap, and highly effective).

Is trauma something that can easily be measured.

Of course. Like most established constructs in psychology, there are both diagnostic criteria for assesment by trained professionals and self-report indexes. Most of these tend to be fairly high on agreement between different measures as well as test-retest reliability.

comment by Milan_Griffes · 2019-06-01T09:08:30.095Z · score: 2 (1 votes) · EA · GW
Are you aware of any extremely efficient ways to reduce trauma? Is trauma something that can easily be measured (maybe secondarily through stress hormones)?

There are well-validated instruments for measuring post-traumatic stress disorder. MDMA therapy is proving highly effective at treating PTSD (see my answer [EA · GW] for evidential support).

comment by nonzerosum · 2019-06-04T19:08:16.439Z · score: 1 (1 votes) · EA · GW
more than as something which "improves coordination"

What makes you say that? I have the sense that the less trauma people have, the easier they'll find it, and the more desire they'll have, to co-operate and coordinate.

comment by aarongertler · 2019-06-05T08:20:13.747Z · score: 2 (1 votes) · EA · GW

What makes you think that this secondary effect, which requires trauma reduction in an enormous number of people (to generate network effects) or a tightly-knit group of people, would have a greater impact than the primary effect of "people have less trauma and feel better"?

comment by nonzerosum · 2019-06-11T19:15:30.667Z · score: 1 (1 votes) · EA · GW

Thanks for that question! Weakly held. Some sense that we're under-invested in "improving coordination" (see:

But it's a good point that it would be hard! And I agree that tightly knit groups may be a better approach for this.

e.g. trauma reduction for a group of AI safety researchers to help them better coordinate, or something like that.

comment by nonzerosum · 2019-06-11T19:15:55.228Z · score: 1 (1 votes) · EA · GW

And I'm also very interested in the direct impact, too.

answer by Khorton · 2019-05-31T23:28:54.918Z · score: 2 (3 votes) · EA · GW

"You could imagine an experiment where there are X "trauma-free" people"

I cannot imagine that experiment. I've never met someone who hasn't experienced something at least mildly traumatic.


Comments sorted by top scores.