Is it true that more than half of medical consultations are for MUS? A look at the evidence.

[Inara]

People here are interacting on eye-level.

I feel I have to say it, since it's claimed several times in this thread: No, I don't feel like that. Certainly, there are people that interact with me on eye-level, but there also people who don't. Just with regards to me, because I can't speak for others, I don't feel this is generally true.

 

[Inara]

I think there's a big misunderstanding here about what I'm saying when I say that MUS professionals generally accept a 50% prevalence figure. I am not saying that 50% of symptoms are psychosomatic

What I'm saying is that half of symptoms are undiagnosed in outpatient settings. Then I say (in the article) that research typically equates undiagnosed symptoms with psychosomatic symptoms. The article is then devoted to showing how it's unscientific, and unethical, to assume that undiagnosed symptoms are psychosomatic. See?

That's what I've just wanted to write: That's how I understood it.

On top of that, it's quite realistic the MUS numbers aren't correct - but it is assumed they are by some people, including institutional settings.

 

[duncan]

If you write about any industry long enough, whether or not it be medical, you will use rule-of-thumb numbers. It's a short hand. It doesn't make those numbers right. I use them all the time with Lyme when I reference 20% of those infected do not get well despite abx. I am fully cognizant that there are those who dispute those numbers - both ways. But those numbers ARE widely acknowledged.

Nevertheless, there are vehement protests against those numbers, and maybe with good cause.

I think clearly that is all that is going on here. I do not blame @Diane O'Leary for indulging in that tool.

 

[Trish]

What I'm saying is that half of symptoms are undiagnosed in outpatient settings. Then I say (in the article) that research typically equates undiagnosed symptoms with psychosomatic symptoms. The article is then devoted to showing how it's unscientific, and unethical, to assume that undiagnosed symptoms are psychosomatic. See?

Maybe this misunderstanding explains why people have felt such a need to dispute that figure?

Yes, I do get that. My point is not related to the undiagnosed versus psychosomatic controversy or the ethics involved in that. My point is much simpler - where did the 50% figure come from?

It may not be important to your central ethical argument, but it is important if it means our government in the UK is spending billions on MUS 'services' on the basis of a vastly inflated figure. That matters to me.

 

[Amw66]

Yes, much like the 250,000 with ME that seems to have been the " go - to " figure for a long, long time in UK.

 

[Jonathan Edwards]

What I'm saying is that half of symptoms are undiagnosed in outpatient settings.

I think the problem is that we don't have any reason to say that. I don't think even any of the papers counting proportions of 'MUS' (in capitals implying the author assumes a psychological origin) say this.

I am not saying that you are writing as if 50% were psychosomatic. I am puzzled that you use these figures for unexplained symptoms, assuming that are not psychosomatic.

You are describing what I would call 'mus' (in small letters meaning that it is read literally, not as a term of psychosomatic art). But there is no reason to think the numbers for MUS equate to numbers for mus because all the papers ascertaining proportions for MUS build criteria relating to psychosomatic attribution into the counting. MUS does not mean any mus. It means symptoms in people who repeatedly attend doctors for multiple different complaints and because of this behaviour are specifically deduced to have a psychological problem. The situation is complicated by the fact that the papers quoted may well disingenuously add up all examples of mus they can find despite claiming that they are only counted if they fit criteria for MUS. But these figures are specifically intended to imply psychological cause. Their authors would not want them to be used for non-psychological mus.

I guess the point is that if we all agree that these figures are not credible for MUS then we have already assumed that they are pretty meaningless and so we would not want to make use of them as numbers for mus. As indicated above, mus occur in 100% of us. So what? The problem is not the mishandling of mus. The problem is the mishandling of people who for whatever reason attend doctors repeatedly with multiple complaints.

 

[Trish]

It's a good question, Trish. It might help to know that the higher estimate in Kroenke and Mangelsdorf (in the body of the paper) was incredibly influential.

OK, I have not been able to get hold of the full paper, can anyone help me here, I'm not very good at getting Sci hub to work.

Alternatively perhaps you could help out Diane, and quote me the section in the Kroenke and Mangelsdorf paper that gives this highly influential figure, since they have not put it in their abstract as far as I understand it. (See posts #2 and #18 in this thread).

As an added level of confusion, the high figure you quote from these authors is, you say, the proportion of symptoms that are unexplained, not the proportion of patients.

That can vastly inflate the story. I may have 10 symptoms which may all relate to an as yet undiagnosed condition when I visit the doctor, and she may be able to explain the single symptom each of the next 10 patients reports. So half the symptoms she is faced with that morning are unexplained, but less than 10% of the patients have unexplained symptoms.

Can you clarify?

 

[Diane O'Leary]

Yes, I do get that. My point is not related to the undiagnosed versus psychosomatic controversy or the ethics involved in that. My point is much simpler - where did the 50% figure come from?

It may not be important to your central ethical argument, but it is important if it means our government in the UK is spending billions on MUS 'services' on the basis of a vastly inflated figure. That matters to me.

Ah, I see now. The reason that Up to Date states "over 50% of patients are not suffering from a medical condition" is that this has been received opinion for decades, particularly in health policy. There is a ton of research over many decades that established this picture. Try looking up Barsky in the 1980s. (That stuff will really make your hair stand on end.)

In my opinion it's not possible to dispute this picture by quoting research. It's too entrenched, and there are literally thousands of papers that support it. In any case, it's actually useful for the ME cause to accept that half of symptoms are undiagnosed. It shows that half of the time doctors don't actually have answers. Then it's easy to show that there's never been a safe, scientific approach to handling that reality.

Unexplained symptoms actually are extremely expensive for the health system. It's good for the ME cause to acknowledge that too - because it helps us show that ME management is not based on what's good for patients. It's based on what's good for the budget.

 

[Diane O'Leary]

OK, I have not been able to get hold of the full paper, can anyone help me here, I'm not very good at getting Sci hub to work.

Alternatively perhaps you could help out Diane, and quote me the section in the Kroenke and Mangelsdorf paper that gives this highly influential figure, since they have not put it in their abstract as far as I understand it. (See posts #2 and #18 in this thread).

As an added level of confusion, the high figure you quote from these authors is, you say, the proportion of symptoms that are unexplained, not the proportion of patients.

That can vastly inflate the story. I may have 10 symptoms which may all relate to an as yet undiagnosed condition when I visit the doctor, and she may be able to explain the single symptom each of the next 10 patients reports. So half the symptoms she is faced with that morning are unexplained, but less than 10% of the patients have unexplained symptoms.

Can you clarify?

I have to go to work now, but quickly - you are so right that there's conflation of proportion of symptoms and proportion of patients. And you're right about how important this is to the statistical picture! The research on this stuff is so bad that they have not reckoned with this problem. That's what I'm saying. You're not going to find an answer to this question in research because they're not that careful. It's appalling. Very deeply confused on so many levels. That's a big part of what I do - I show all the ways that this stuff is unscientific.

Sorry, it takes time to hunt down that paper and find the passage you're looking for, and I have to get back to work. Anyway, suppose we could show that paper to be deeply faulty, what then? Nothing, because there are thousands more where that came from. Up to Date doesn't pick this stuff out of the sky, and they don't base it on one paper. Research has supported this picture for decades.

 

[Adrian]

The best way to assess the validity of the 50% figure (that's the global figure I use, after noting that 52% is generally what they use in health policy in the UK), is to look and see how those 12 commentary and criticism papers respond to it. It's helpful to note that not one of those 12 critical evaluations of my article saw a need to debate the grounding figure of 50%. Again, that's not to say that it cannot be debated - but it is to say that according to the expertise of reviewers, and the 12 teams of professional experts who commented on the paper, it is correct to say that a figure of roughly 50% represents typical understanding.

Perhaps an important question is exactly what is being measured and whether the measurements are sufficient. To look, for example, at health records that would classify as MUS then that may indicate that the doctors seeing the patients didn't find a physical cause. But then we also need to look at the quality of the data. For example, was there later an identified cause (I've known people who have raised issues with doctors for several years and multiple visits before finding a (physical cause).

So it feels really important to define what is measured - that of the doctors perception that a patient is suffering from MUS. That needs to be caveated by the data collection methodology - opinion given during 5 minute conversation with a doctor who has a p(diagnosing MUS | ¬MUS) of x (i.e some information about the likely accuracy of the diagnosis. How does counting happen when a diagnosis is later achieved which could account for the symptoms.

I also wonder if there is a higher frequency of diagnosis of MUS after training/advertising campaigns etc.

By pulling together some information about the measurement frameworks, room for error perhaps interesting questions could be asked around the likely validity of the approach around counting % of patients with MUS and therefore whether the data used for forming public policy has any validity.

 

[Lisa108]

I feel I have to say it, since it's claimed several times in this thread: No, I don't feel like that. Certainly, there are people that interact with me on eye-level, but there also people who don't. Just with regards to me, because I can't speak for others, I don't feel this is generally true.

Yep, point accepted, @Inara. I can only say so for myself.

_________
This statement I find worrysome:

You good people are very trusting, and I think you don't know what an attack looks like in the language of professional discourse.

Could you please explain, @Diane O'Leary, what you mean with "you good people" [in contrast to whom?] to "are very trusting" [into what?] to "I think you don't know what xyz in professional discourse looks like" [why not?].

Thanks.

________

 

[Trish]

I have to go to work now, but quickly - you are so right that there's conflation of proportion of symptoms and proportion of patients. And you're right about how important this is to the statistical picture! The research on this stuff is so bad that they have not reckoned with this problem. That's what I'm saying. You're not going to find an answer to this question in research because they're not that careful. It's appalling. Very deeply confused on so many levels. That's a big part of what I do - I show all the ways that this stuff is unscientific.

Thank you for clarifying that.

I agree it's probably impossible for little old me lying in bed to influence anything and I'm certainly not proposing to spend the rest of my life reading 1000's of papers.

However, I started this thread with a much more focused target. I looked at the 52% claim in the UK MUS NHS document linked in my opening post, and tried to follow the evidence trail the writers of that document had used to justify their choice of that figure. I was also interested to see whether there were other papers that contradicted that figure.

So quite a narrow purpose, with, perhaps, a chance to challenge that particular document. I have no illusions that I would succeed in changing it, but there's no harm in trying if that's how I choose to spend my time and energy.
...............

This thread was not intended to be about the wider topic of your article on ethics.

Of course I have no control over where a thread goes, but since you have joined in the discussion and highlighted your ethics paper which repeats the 50% figure, I am interested also in your choice of papers to justify your quoting of the 50% figure, and your decision not to challenge the validity in statistical terms of that figure, but only on ethical terms. Does that make sense?

 

[Inara]

"I think you don't know what xyz in professional discourse looks like" [why not?]

Maybe because most people here haven't presented their findings on conferences and witnessed the shark tanks there? There are some "buzz words" with a certain meaning. Also, my experiences were questions weren't asked out of interest and the desire to understand, but to show how incompetent someone was. It was pure competition.

 

[Wonko]

I'm not sure if the data informing policy being valid is relevant.

As a completely uninformed and unqualified outsider it looks to me as if policy is decided, then data generated and/or manipulated to fit the policy. For PR purposes only. The intent is never to help people, to solve a problem, the intent is to achieve whatever political objective the policy was designed to, whilst also, if possible, looking like they 'care'.

So expecting any data to fit with the stated objectives, as opposed to the actual objectives, seems to miss the point, at least to me.

I suspect this is why it's so hard to refute some studies, some people, because nothing we say, no objections we raise, are, in any way, relevant to them. Because the studies, the people, weren't ever intended to help anyone, and pretty much everyone involved, from academics, to government, knows that.

 

[Diane O'Leary]

I mean, genuinely, that the people on s4me are good, well meaning people. You're also highly intelligent, articulate and very well informed.

 

[Arnie Pye]

But then we also need to look at the quality of the data.

Medical records in the NHS are dreadful. Doctors will lie if telling the truth will put them in a bad light. If they get something wrong they tend to make it disappear. As a 13-year-old I had a very serious medical issue that is (nowadays) described as a medical emergency. I was in extreme pain for a month before I was finally referred to hospital. My medical notes record that I was in pain for 24 hours. The explanation for the problem I had has been hidden away and watered down.

 

[Adrian]

In my opinion it's not possible to dispute this picture by quoting research. It's too entrenched, and there are literally thousands of papers that support it. In any case, it's actually useful for the ME cause to accept that half of symptoms are undiagnosed. It shows that half of the time doctors don't actually have answers. Then it's easy to show that there's never been a safe, scientific approach to handling that reality.

Its not a case of quoting research rather we should be looking at the quality of the data by assessing the recording and reporting systems. Its also not clear to me what is measured is it by a per visit basis? So when someone I knew visited 10 GPs whilst suffering from aplasitic anemia and was told to go away because she wasn't ill and the 10th formed a diagnosis of aplastic anemia (a lack of blood should be pretty easy to spot!). Does this form a 90% MUS to 10% not MUS ratio in recording.

I've known so many patients who need to see a doctor on multiple occasions prior to getting a diagnosis. How are these visits counted in the stats. Without a really detailed knowledge of the data collection and an assessment of the accuracy of recording we can't make any real judgements about numbers.

A lesson we should learn from data science is the issues in getting meaningful data, the need to spend a lot of time understanding and cleaning up the data. Then we can start to derive trends otherwise inferences are often meaningless or misleading,

 

[rvallee]

Note too that the Up to Date research review system - which you won't be able to access - tells doctors that "more than 50% of patients are not suffering from a medical condition", or something like that. This is the system that summarizes current consensus on everything in a way that doctors can access while they're in the exam room or hospital hallways.

I had seen that opinion a few times and thought it was a fringe idea. Seems it's a common belief that influences practice. That really explains a lot.

I really can't follow the thought process, it does not follow any rational logic beyond the god of the gaps or some other logical fallacy. It makes intuitive, superficial, sense but that's not something a professional should be fooled by. That kind of thinking basically amounts to recording unsolved crimes as "no crime occured". It's not just an inability to differentiate, it distorts the data entirely.

Plus as has been discussed it does not differentiate between patients and symptoms. Most patients with difficult to diagnose conditions will have seen 10-20 physicians, all of whom will have concluded "no medical condition" as a temporary status. But there is no back-propagation to mark those consults, they remain false negatives that leave the impression that 9-19 or so patients were incorrectly found to not have a medical condition.

This is really like getting old enough to understand that not a single adult actually knows what they're doing. That's usually not so bad, but when it's coupled with a mentality that refuses to acknowledges mistakes or fallibility in a life and death profession, well, people suffer and die. Serious, major, reforms are badly needed.

 

[Amw66]

I had seen that opinion a few times and thought it was a fringe idea. Seems it's a common belief that influences practice. That really explains a lot.

I really can't follow the thought process, it does not follow any rational logic beyond the god of the gaps or some other logical fallacy. It makes intuitive, superficial, sense but that's not something a professional should be fooled by. That kind of thinking basically amounts to recording unsolved crimes as "no crime occured". It's not just an inability to differentiate, it distorts the data entirely.

Plus as has been discussed it does not differentiate between patients and symptoms. Most patients with difficult to diagnose conditions will have seen 10-20 physicians, all of whom will have concluded "no medical condition" as a temporary status. But there is no back-propagation to mark those consults, they remain false negatives that leave the impression that 9-19 or so patients were incorrectly found to not have a medical condition.

This is really like getting old enough to understand that not a single adult actually knows what they're doing. That's usually not so bad, but when it's coupled with a mentality that refuses to acknowledges mistakes or fallibility in a life and death profession, well, people suffer and die. Serious, major, reforms are badly needed.

Perhaps a sign of getting older is the realisation as to just how much of life is determined by people " winging it" .

 

[John Mac]

Yes, much like the 250,000 with ME that seems to have been the " go - to " figure for a long, long time in UK.

Except for when the BPS crowd tried to inflate that figure to 600,000 in 2013.
It was only when it was repeatedly questioned where that figure came from that they dropped it.
Always empire building.
Here's Bristol University's press release quoting the figure on behalf of Crawley.

https://www.bristol.ac.uk/news/2013/9741.html