"The biopsychosocial approach: a note of caution" George Davey Smith (2005/2006)

[Art Vandelay]

Stumbled onto this thread earlier - what on earth happened to GDS in the years between, such that he became a BPS supporter?

He probably realised that BPS research in the UK is one hell of a gravy train.

 

[slysaint]

He also quit from being another publication editor last year:
The International Journal of Epidemiology (IJE)
"
the resignations of George Davey Smith and Shah Ebrahim late last year"
http://www.epimonitor.net/PrintVersion/Jan 2016/Jan-2016-The-Epidemiology-Monitor-final.pdf

he's part of this:
https://hceconomics.uchicago.edu/
https://hceconomics.uchicago.edu/people/george-davey-smith
"George Davey Smith is Professor of Clinical Epidemiology at the University of Bristol, Honorary Professor of Public Health at the University of Glasgow and Visiting Professor at the London School of Hygiene and Tropical Medicine. He is Scientific Director of the Avon Longitudinal Study of Parents and Children (ALSPAC) and Director of the MRC Centre for Causal Analyses in Translational Epidemiology."

I'm guessing he was 'recruited' to help the BPS gang to get into genetics via epidemiology.

Plus there is possibly use of the 'Sue Marsh' tactic. (ie get your biggest critic to work for you).

 

[Woolie]

Anyone interested in discussing Mendelian randomisation techniques (what GDS is famous for)? I've been reading some of the papers. What I concluded was:

1. The aim of Mendelian randomisation is not to study the role of genetic factors in disease. Its aim is to examine how biological or lifestyle factors affect health outcomes. The genetics is just a tool that can be used for this purpose - which might be less susceptible to common confounds.

2. Its based on the idea that if you can locate a gene that's reliably associated with a particular lifestyle practice (e.g. fat gene, which predicts obesity), then you can see if people with that gene are more prone to a particular health outcome (in this case, it might something like raised inflammatory markers). Of course, its less direct than just seeing if fatties have more inflammation, but the big plus is in (3) below.

3. The big plus of this is that genes vary independently of all the other psychological and social factors that are commonly correlated with the lifestyle factor you're interested in (so obesity is negatively correlated with social class, And we know social class affect health in all sorts of ways). So its possible in this way to control covariates in a way you can't by studying the actual observed factors.

4. Most importantly, you can test direction of causation. To do this identify a gene that's associated with your chosen inflammatory marker. You then look at whether this gene predicts obesity. If this analysis is non significant, then you can conclude that obesity causes inflammation, not the other way around.

Anyone want to discuss? I'll post my own worries later if people seem interested.

 

[Woolie]

Anyone want to discuss? I'll post my own worries later if people seem interested.

That went down like a cup of cold sick!

Maybe I was being way too nerdy...

 

[A.B.]

That went down like a cup of cold sick!

Maybe I was being way too nerdy...

Not sure about the others, but I lost interest in Davey-Smith after the MEGA and PACE special issue fiasco. He should just leave us alone. He has indeed turned into neoliberal hypocrite.

 

[Hutan]

That went down like a cup of cold sick!

Not at all, it's just the weekend. We're trying to do other stuff and getting too tired to think.

It's interesting. My brain's not co-operating, but I'll try.

Its based on the idea that if you can locate a gene that's reliably associated with a particular lifestyle practice (e.g. fat gene, which predicts obesity), then you can see if people with that gene are more prone to a particular health outcome

Isn't the finding of a gene that's reliably associated with a particular lifestyle practice usually a problem though? Even if you know that one gene increases the risk of something, there are probably other genes that might mitigate that risk and others that also increase it (and not all of these genes may be identified yet).

Seems to me that most of the time there would be so much noise that there wouldn't be a clear association between having a particular gene and a particular life style practice.

It might end up being a bit like the functional MRI's - you can fish around and prove whatever theory you want?



​

 

[user9876]

2. Its based on the idea that if you can locate a gene that's reliably associated with a particular lifestyle practice (e.g. fat gene, which predicts obesity), then you can see if people with that gene are more prone to a particular health outcome (in this case, it might something like raised inflammatory markers). Of course, its less direct than just seeing if fatties have more inflammation, but the big plus is in (3) below.

The notion of one gene one thing and no interaction seems strange. I assume genes really control quite low level processes making it hard to find a 'fat gene' but there maybe genes that effect metabolism, appetite signals, absorbsion, etc but then put them all together. The search for a single gene with a high-level effect always seemed strange as I would have thought effects are a function of many genes + environment. With the function being complex, possible feedback loops leading to possible chaotic (i.e. unpredictable) behaviours.

But then I know nothing about biology or genetics.

 

[user9876]

So its possible in this way to control covariates in a way you can't by studying the actual observed factors.

I think even quite a simple mathematical model would start to show this is hard once you include several non-linear functions in the equations. Perhaps you would still get some correlations but I suspect little certainty.

 

[user9876]

That went down like a cup of cold sick!

Maybe I was being way too nerdy...

No not too nerdy just didn't see.

I think its interesting and something that has worried me in the past but to understand it would require a lot of work (given my complete lack of knowledge of the area).

 

[Woolie]

Isn't the finding of a gene that's reliably associated with a particular lifestyle practice usually a problem though? Even if you know that one gene increases the risk of something, there are probably other genes that might mitigate that risk and others that also increase it (and not all of these genes may be identified yet).

Yea. Not everyone with the gene will exhibit the feature, because of other genes or environment or both. That's assumed to be the case. Its just a probabilistic association. The idea is that you use the gene to divide the population into two groups, and one of the groups has a reliably higher probability of having your behaviour or marker of interest than the other.

Of course the size of the probability difference matters. If its marked, then you have more of a chance of capturing the thing you're interested in. If its small, then it will capture it poorly. But to be fair, these papers always report the strength of the initial association - in the Abstract itself - so the reader can decide for themselves how much weight to put on the results.

The logic part of all this I sort of follow. The genetics bit, not so much. I'm still not sure whether they isolate a single gene, a set of genes or what. And I have no clue how they actually do that part.

It might end up being a bit like the functional MRI's - you can fish around and prove whatever theory you want?

It seems to be mainly used to address questions that have already been asked. You don't just fish around will nilly. Most of the published stuff reports positive results, so there might be a fair bit of publication bias going on there. But there are the odd negative findings which make it into print. And there's a fair few studies that use the technique to find out the direction of an already known association (e.g. does smoking cause anxiety, or does anxiety make you more likely to be a smoker? The answer seems to be the second one).

think even quite a simple mathematical model would start to show this is hard once you include several non-linear functions in the equations. Perhaps you would still get some correlations but I suspect little certainty.

You may be way ahead of me here. But if you're asking whether they control for covariates statistically, then no, they don't. The stats are extremely simple. The method itself is designed to take care of covariates. That's the whole point.

So for example, if you find a reliable fat gene, you assume this gene will vary independently of all other genes across your population. So your two comparison groups should not differ systematically on any other gene other than your fat gene. They should be equally likely to be rich or poor, smokers or nonsmokers. This won't always be entirely true but at least there won't be the same sort of multiple huge confounds between variables you see in conventional epidemiology studies.

It sounds like I'm really in favour of the approach, but I do have some worries (which I'll talk about some other time).

 

[Valentijn]

2. Its based on the idea that if you can locate a gene that's reliably associated with a particular lifestyle practice ....

The entire idea is pretty much a non-starter due to the bit quoted above. It reflects either a complete lack of understanding regarding genetics, or a fundamental misrepresentation of genetics.

I suspect it's actually a ploy for undercutting physiological research, treatment, and classifications by inserting a "behavioral" factor - which can be overcome with a little CBT of course. It's similar to claiming that a SNP occurring in fibromyalgia patients with worse symptoms must be the cause for catastrophization about their normal levels of pain, rather than possibly contributing to a physiological disease.

 

[user9876]

You may be way ahead of me here. But if you're asking whether they control for covariates statistically, then no, they don't. The stats are extremely simple. The method itself is designed to take care of covariates. That's the whole point.

I wasn't thinking of the stats directly rather if you build an abstract of the genes and the way they interact with environmental stimuli then I assume this would include a lot of non-linearity which makes predictions hard and unreliable. The test of the stats then comes in terms of whether you can find correlations that you have build into the model given the complexities of other interactions. (I am a believer in testing properties of stats models via simulation under different known stochastic conditions)

you assume this gene will vary independently

I always worry about assumptions of independence in stats. Does a single gene really vary independently (or approach that) given the way humans (or animals) breed? Not sure about the social status stuff being independent since social mechanisms seem to keep money and status in families (at least for a while). I guess there is a question as to whether this has been checked?

sounds like I'm really in favour of the approach, but I do have some worries (which I'll talk about some other time).

I'm not for or against just cynical.

 

[Barry53]

But they're very good at observing what they want to see.

It's a natural human weakness, and it takes perception, integrity and moral fibre to get past it.

 

[trishrhymes]

That went down like a cup of cold sick!

Maybe I was being way too nerdy...

Just discovered this thread by chance. Fascinating discussion which I will now read and follow. My problem is this arose on an old thread I wasn't following, and I tend to only check out new thread, not new posts on old threads. Maybe worth starting a new thread?

 

[Jonathan Edwards]

No not too nerdy just didn't see.

I think its interesting and something that has worried me in the past but to understand it would require a lot of work (given my complete lack of knowledge of the area).

Indeed. When interesting new thoughts get posted on old threads it is easy to miss them too.

Let me suggest an illustration of this effect that might be relevant and hen ask Woolie to voice the concerns. Would like to hear what they are.

Around 40 years ago Rodney Bluestone's lab found almost everyone with ankylosing spondylitis carried the HLA B27 gene. They were not looking for it (it was a control) and did not know what to make of it. Derrick Brewerton saw the interest and published the findings. It is probably the most important thing we know about AS.

However, what HLA-B27 is doing causing AS remains a puzzle. HA genes determine whether you reject a transplant or not but clearly that is not why they are there are what they do in people without transplants. The individual gene alleles or types (like B27) probably protect against specific infections. So B27 is about the best at stopping HIV infection leading to AIDS. We also know that HLA-B genes are involved in activating CD8 T cells. More recently we have discovered that they interact with NK cell receptors.

So we have a gene association. The gene is also linked to various biological traits (as for an 'obesity gene') but turns out to cause a disease that nobody thought was particularly to do with those traits. IN fact now that we know cytokine receptor genes are also associated with AS we can see that AS must have something to do with immune cell activation but we are still in the dark as to exactly why B27 is important.

What this seems to show to me is that finding a gene association with a disease can be very powerful in terms of gathering once's thoughts about how the disease works and providing a fixed anchor point in any theory of cause. Any theory you have has to explain the gene association. For AS we are still not quite there. But for RA we are much closer. The relevant gene for RA is HLA -DR4. That told us to look at CD4 cells and eventually the penny dropped that what CD4 cells are mostly for is to help B cells make antibody and the DR4 was telling us that the antibodies causing the disease needed T cell help. Again we still do not quite know why DR4 but there are sensible suggestions.

The danger that I see is that you may find a gene and if you have a woolly (with two l's and a y) hypothesis about how a disease may be caused you may well come to think the gene is doing one thing when it is doing another. That was precisely the problem for RA and T cells. So if your thinking is vague enough you can use genetics to support almost any idea. The fact that the gene has to be causally first remains a great strength, but it may still lead you up the garden path.

The only real potential weakness of picking out one gene as far as I know is that you may have linkage disequilibrium. That means that people with HLA-B27 might turn out almost always to have a gene for poor posture by some quirk of DNA structure or evolutionary pressures. We worried a lot about that in the 1990s but it looks as if it hardly ever causes a problem. Modern techniques can home in on a DNA sequence and define precisely where the causal factor is likely to be. That can usually be traced to one gene through one trick or another.

So what worries @Woolie?

 

[user9876]

The danger that I see is that you may find a gene and if you have a woolly (with two l's and a y) hypothesis about how a disease may be caused you may well come to think the gene is doing one thing when it is doing another. That was precisely the problem for RA and T cells. So if your thinking is vague enough you can use genetics to support almost any idea. The fact that the gene has to be causally first remains a great strength, but it may still lead you up the garden path.

I can see your argument in that you relate genes to quite detailed processes in the body and hence possible disease mechanisms. That makes complete sense to me. It is when generalizations are made about that to a general condition in the body (such as a fat gene) that it worries me as I would see this being causes by a potential interaction between many different processes and environment.

Even with your argument does this suggest a single gene which has the potential to cause something like AS or is it more like a lot of people with ES have this specific gene and we know it controls this so look here. But other people may have the issue due to the same mechanism but with other genes?

Woolly thinking around data is always a problem and the data should lead to a hypothesis that can then be tested. There is a nice paper by a statistician David Freedman called statistics and shoe leather where he basically says don't rely on statistics to tell you the answer they are just a tool to think about the problem. You still need to put in the work (hence the shoe leather) to understand and draw conclusions.

 

[trishrhymes]

Thank you @Jonathan Edwards for that fascinating explanation. Are you saying that everyone with RA has the HLA-DR4 gene and everyone with AS has the HLA-27 gene?

If so this seems rather different from what I've gathered about GDS's approach. For example for the MEGA study he wants 12000 participants to do GWAS studies. This suggests he is looking for genes with a much weaker association with ME/CFS, and therefore likely to be much less useful in both understanding ME/CFS and seeking treatments.

 

[Snow Leopard]

Anyone want to discuss? I'll post my own worries later if people seem interested.

It'll all get lost in the noise. There are no "gay genes", there are no "obesity genes" and so on... There might be risk factors, but when they are not the dominant factor, there is usually too much noise to derive anything meaningful.

If someone wants to prove me wrong mathematically, be my guest, otherwise...

 

[Woolie]

The entire idea is pretty much a non-starter due to the bit quoted above. It reflects either a complete lack of understanding regarding genetics, or a fundamental misrepresentation of genetics.

I suspect it's actually a ploy for undercutting physiological research, treatment, and classifications by inserting a "behavioral" factor - which can be overcome with a little CBT of course. It's similar to claiming that a SNP occurring in fibromyalgia patients with worse symptoms must be the cause for catastrophization about their normal levels of pain, rather than possibly contributing to a physiological disease.

No, I made it sound like it was all about health practices and behaviours, but in truth its applied to any question of an epidemological nature. Some examples (click on the studies for links):

- Does vitamin D deficiency increase one's risk for MS? (the answer from this method seems to be yes)
- Is high C-reactive protein a risk factor for depression? (the answer from this is yes).
- Are fat people more prone to depression? (the answer seems to be no. This result is surprising, because conventional studies find that fat people are more prone to depression. This is a case where conventional methods maybe drive by the social class confound - obesity is more prevalent in low SES).

Your second concern - how it would be applied in MECFS research - is a very real one. But it certainly wouldn't be things like 'catastrophisation' they would study. It has to be things where a single gene can be found that controls the feature or behaviour of interest. So I'd expect they'd be interested in things like inflammatory markers (CRP, IL6), but maybe also anxiety, depression, obesity. The plus would be they would also look at the reverse -- whether have a CFS 'gene' if they identify one raises your likelihood of being anxious, depressed, etc.

In other words, it could possibly work in our favour.

Off to bed now, look forward to reading you other thoughts tomorrow.

 

[Jonathan Edwards]

Thank you @Jonathan Edwards for that fascinating explanation. Are you saying that everyone with RA has the HLA-DR4 gene and everyone with AS has the HLA-27 gene?

If so this seems rather different from what I've gathered about GDS's approach. For example for the MEGA study he wants 12000 participants to do GWAS studies. This suggests he is looking for genes with a much weaker association with ME/CFS, and therefore likely to be much less useful in both understanding ME/CFS and seeking treatments.

It is surprising how many people you need topic down even a gene like B27 in AS. But yes, the situation varies widely. Most people with RA have DR4 but only a slight majority. For lupus there are causal genes that are only relevant in a tiny number of patients - like complement C1q deficiency or C2 deficiency. But because these cause lupus very reliably that tells us that the muddle of C4 gene variants are probably relevant too.

There might be a certain gene that you really have to have to develop ME. There are some genes like that for MS and AS. You still need huge cohorts to pick them out if you are trawling blind because of the statistical problems of false positives.