Can we know anything?

RE: Can we know anything?

John Flynn — Fri, 06 May 2016 00:22:06 GMT

I've read a few threads on this homopathy since joining this forum, but they seem to generate more heat than light.

As far as I can tell, this is one of those threads at heart.

I remain quite unclear what homoapthy is (and have never met anyone outside of this forum with an interest in the topic, whatever it may actually be, or heard of any connection between this and animals before I googled it), but as far as I can tell it is, in the words of its fan-page:

WHAT IS HOMEOPATHY

Homeopathy (homoeopathy / homœopathy) is a system of medicine that bases its therapeutics on the principle of ‘let like be cured by likes’ – ‘similia similibus curentur’. The medicines used may be derived from animal, vegetable or mineral sources and, in latter times, remedies have also been derived from man-made substances. The initial requirement for treatment is knowledge of what effect a particular remedy or substance will have on a healthy body (i.e. what signs and symptoms it can provoke in a healthy body). The signs and symptoms presented by a sick animal or person are then compared to this ‘symptom picture’ of the various medicines, choosing that medicine which is the closest ‘match’."

This is obviously a very silly approach to choosing a medicine and doesn't make any sense.

It is also not something that lends itself to studies involving frequentist statistics on an individual treatment chosen in this way in an attempt to infer either that this is a reliable method of choosing a treatment or an unreliable method of choosing a treatment.

Publications are littered with plenty of meaningless infernal p-values without adding more!

It is a requirement of frequentist statistics that the proposition sounds plausible. This proposition is no more scientifically refutable by doing isolated "studies" on individual treaments picked by this mantra than it is prove-able - it is simply implausible.

I see very little that is "proven" in veterinary treatments for obvious reasons.

I see plenty of bad maths however.

I think a campaign for rational veterinary medicine could focus on proper use of maths and the application of rational thinking to medicine selection. For instance, a common topic in almost any clinical thread to medicine choice seems to be (perceived - none of the posters seem to agree on what these "rules" actually are) regulatory restrictions to UK vets making rational medicine choices - this appears to have nothing to do with homopathy, but is what I thought a campaign for rational veterinary medicine in UK might be about before I read on.

On the bad maths front, I just pulled up 2 classics on elbow dysplasia as an exemplar, but actually it is decidedly harder to find examples of good maths than bad maths:

Paper 1 (from 2008) provides a "meta-analysis" of 4 studies that don't answer the question posed before wielding Bayesian statistics to improbably conclude: "The probability that arthroscopy is superior to medical management is 1.0, with a 17% ([10%,25%], a 95% probability interval) increase in improvement if arthroscopy is used instead of medical management. A probability of 1.0 is inferentially similar to a very small p-value (e.g. p<0.0001). It means that it is
certain that arthroscopy is superior to medical management and not due to chance."

Paper 2 (2011) uses extremely sophisticated (probably £xxxxxx) technology to study a vanishingly small number of dogs that couldn't possibly give results that would demosntrate their hypothesis that arthroscopy is of no benefit [they obviously were unaware of the conclusions of the 2008 paper or they wouldn't have needed to perform a study at all...] to be likely correct to any reasonable confidence with standard equivalence testing, before concluding (after failing to disprove their hypothesis with a reasonable degree of certainty):

"We concluded that there is no therapeutic benefit to arthroscopic removal of fragmented coronoid processes or chondroplasty of the MCP"

Both these papers appear to be from UK universities. I would start my efforts for a campaign for rational veterinary medicine there rather than with some wacky fringe group that no-one is likely to pay any attention to. I don't mean this as a criticism, but more as a suggestion of future direction with greater yield of benefit in my sight anyway. If you are trying to encourage the restriction of medicines to some formulaic prescribing mechanism based on trial results, then I think that is irrational and detrimental rather than beneficial - vets in the UK need to be encouraged to prescribe drugs rationally rather than on limited, unseen data presented to a regulatory authority and the semantics of what is then written on the instructions in my view.

PS - I am no more clear on what EBM/EBVM is than what homopathy is - the definition of the former seems to range from "everything" (i.e. nothing then), to something akin to a game of top-trumps using studies published from "reputable" journals, I am most unclear as to what it is meant to replace or what the relevance to me is supposed to be.

RE: Can we know anything?

David Mills — Thu, 05 May 2016 19:46:51 GMT

[quote user="Eamon McAllister"]

The one thing that I do know is that I find it hard to follow this thread without losing focus and nodding off. Bunch of blokes in the playground of this forum demonstrating their erudition to one another methinks.

[/quote]

http://mobile.brainyquote.com/quotes/quotes/a/abrahamlin109276.html

RE: Can we know anything?

Eamon McAllister — Thu, 05 May 2016 16:18:21 GMT

RE: Can we know anything?

Francisco Gomez — Thu, 05 May 2016 10:58:10 GMT

[quote user="David Mills"]

it is not equipped to do so, epsitemically, and its grading of evidence according to study methodology is highly suspect (see Worral's papers on this). It is illogical to think a single methodology (whatever happened to its democratising credentials??) can capture the nuanced differences across evidence extending from large-scale clinical trials and anecdotal evidence or series reports. It's premises no not entail its conclusions.

[/quote]

I agree in the form, that a single benchmark can't ensure you best practice. But I'm afraid that we need to use 'some' benchmark, and the ones we're using now - for example, giving a double blinded prospective study more weight than a case report - are, historically, and just like democracy, the best of all bad forms of government, or in this case, governance.

Your suggestion risks something even worse that what we've got now, you're risking that goalposts would be moved based in what's is convenient to arrive to the conclusions that you want.

vd hqs.jpg

There is a tool in photoshop that it would put that tibia straight David, your own xray from another post. Following your own argument, there's no way you can stop me from believing that once I've photoshopped that leg, the resulting picture won't have a positive impact on your patient's gait.

Don't believe me? You must be applying the wrong epistemology.

RE: Can we know anything?

Hannah Wynne Richards — Thu, 05 May 2016 10:55:09 GMT

David

Official RCVS position that if you're not practicing EBVM then you're unfit to practice.............and yet they allow homeopathy!!!!!!!!!!!!!!!!!!!!!!!!!!!!

You couldn't make it up!

Wynne

RE: Can we know anything?

Alex Gough — Thu, 05 May 2016 09:44:21 GMT

[quote user="David Mills"](or, cynically put, their silence hijacked for political gain).[/quote]

Can we expect to see you running for RCVS council next year then? You would (probably) get my vote.

Alex

RE: Can we know anything?

David Mills — Thu, 05 May 2016 01:59:25 GMT

[quote user="alex gough"]On that basis, I have taken the liberty of correcting your post to more properly reflect this world view. [/quote]

The probablies are normally silent. That is not to say they do not exist! (or, cynically put, their silence hijacked for political gain).

[quote user="Malcolm Ness"]EBM is emphatically NOT what you are describing here. EBM is a methodology for practitioners to use to evaluate ALL of the relevant evidence (including experience, anecdote, consensus, eminent opinion as well as the published literature) in order to address a SINGLE clinical question framed around a patient centred problem. [/quote]

I call this the shapeshifter response, and it is commonplace amongst EBM advocates - see the response to my article in the Vet Record a while back from RCVS Knowledge. Lest we forget EBM has brought this fuzziness and grandeur on itself - it was Sackett et al that proclaimed it a 'paradigm shift'. EBM's defence that "what you're describing is not EBM" starts to wear a little thin - if critics truly don't know what it is, it is because EBM hasn't made a good enough job of defining itself, overreached itself (especially early on), or has allowed itself to be hijacked, transmuted, flattered to become something epistemically and practically bloated and unwieldy. And before it's said that none of the above is EBM's fault, well that take us back to the criticism of cookbook medicine, doctors and vets as the presser of the buttons to make the EBM computer whir: EBM cannot at once isolate itself from all the 'other stuff' whilst at the same time expect people to nod sagely and accept it can also encompass patient values and clincial expertise, both of which, epistemically, seem beyond its methodology.

Whilst I agree with your definition - to an extent - that EBM is ideally based around a single clinical question, I disagree with both the scope of evidence it aims to use (and, mundanely, the practicality of such, hence the new authoritarianism of systematic reviews and the like) - it is not equipped to do so, epsitemically, and its grading of evidence according to study methodology is highly suspect (see Worral's papers on this). It is illogical to think a single methodology (whatever happened to its democratising credentials??) can capture the nuanced differences across evidence extending from large-scale clinical trials and anecdotal evidence or series reports. It's premises no not entail its conclusions.

Interesting that you mention Greenhalgh who from inside the tent has produced of late several cutting critiques of EBM. Her work on narrative medicine in particular stands ill at ease with EBM advocates.

To an extent, this is being recognised in the medical world, where a paired-down version of EBM is being promoted, restricting itself essentially to evidence that can be statistically analysed. It is increasinlgy clear that to achieve patient-centred care, we must dispense with Feyeraband's tyranny of the method. Part of the problem has come from doctors and vets thinking their work is scientific - once we accept that we are a science-using profession, then things are more simple. Patient (ours included) values and clinician expertise will always be beyond the EBM 'methodology'. There has been much written recently on values-based medicine (using science, or a paired-down version of EBM) which has implications for our own profession. It would be a crying shame if EBVM tread the same mistaken path that EBM has over the last 20 years. Relegating it to what it does best - epidemiological analysis, and grading of results therein - would help clarify our future direction and temper the hubristic announcements from our own Sacketts. Did anyone know, for instance, that the official RCVS position is that if you're not practising EBVM then you're not fit to practise? Stunning.

RE: Can we know anything?

Alex Gough — Wed, 27 Apr 2016 23:11:05 GMT

So I believe your answer, David, is that nothing is truly knowable, only things are more or less probable. On that basis, I have taken the liberty of correcting your post to more properly reflect this world view. I look forward to your posts in the future being along the following lines:

Probably interesting. Do you think you were rational in arriving at your current conclusions about the world? Or was it simply what you were taught along the years and your experience roughly correlated with?

Francisco Gomez

One or way or the other, modern physics found a justification for it, and I feel safe to say that minor changes in how physics work over traditional physics makes no difference in how I treat animals today.

We're probably back in Kuhnian territory. I believe no-one throughout the ages, overall, has wilfully misinterpreted the world or deliberately constructed wrong theories of science. It is likely, if history teaches us anything that science as we know it now will be probably be unrecognisable and the source of much mirth 200 years hence. We probably can't see it because at the moment it works, though there are cracks - quantum theory, string theory, hell even gravity definition are probably challenging the bedrock of physics and they are probably not easily explainable using current models.

Regarding treating animals, thinking about the basis about what you think you know is probably enlightening and may lead to better medicine. It is probably the call to arms of EBVM.

Francisco Gomez

I don't think you can compare it to the level of rationale behind an homeopathic remedy.

Au contraire. There is probably nothing a priori that scieintific knowledge has that homeopathic knowledge doesn't. How do you respond to their charge that of course science doesn't show a benefit of our therapies because its the wrong epistemology? Don't forget that history probably teaches us that new or freshly packaged theories are often rejected initially before probably being accepted later.

Arlo Guthrie

and then that we draw our lines in the sand as to what constitutes sufficient evidence (still with you), but then criticise the black cab driver for where they've drawn that line (now you've lost me).

Sorry, the black cab driver logic was probably not a criticism of pragmatism in its truest philosophical sense, but merely that when confronted with difficult questions, scientists probably often counter well we're too busy working in the real world, mate, of course its x, y, z. This is probably a logical fallacy - the obviousness fallacy - and probably predicates itself on "common sense". We as a profession probably deserve better arguments than this.

Arlo Guthrie

It's not just a bit unlikely, it's off the scale improbable.

See this is probably more promising, though off the scale is probably unhelpful. I probably happen to agree, to an extent, in that it is probably highly improbable homeopathy works. What I probably detest is absolutes, and being told such. I probably don't believe there is enough evidence against homeopathy to say it definitely doesn't work, only it is very unlikely to work. Probably not the same thing.

I am probably not arguing for homeopathy. I am probably arguing against the assumptions and packaging of the call to effectively ostracise it.

Arlo Guthrie

And isn't it somewhat ironic that the most commonly used argument for homeopathy is 'I've seen it with me own eyes, guvnor'

And probably in veterinary medicine, alas. Plenty of examples on this forum. Probably.

;)

Arlo Guthrie

Surely, therefore, it demands the very highest standards of evidence to be recommended by a doctor or vet?

You see, you would hope so. But glass houses and all that - veterinary medicine's evidence base is minute, biased, and skewed towards referral populations. Arguably because animals cannot consent, we should be aiming beyond human medicine. But we're not, because its implausible. Much of what we do is anecdotal - taking us back into the homeopaths redoubt. This should make us uncomfortable and more evidentially humble, and aware of our limitations and fallibility.

Arlo Guthrie

the Campaign For Rational Veterinary Medicine isn't calling for a ban, just that it should be mandatory that patients are able to give informed consent. In other words, that it becomes mandatory for practitioners to inform owners/patients that the weight of evidence - the crushing weight of evidence (sorry, lapsing into PR-speak but it is 'crushing') - is that homeopathy does not work.

See above. This is not something veterinary medicine would be advised to get into. Further scrutiny of our thinking and processes shows them as not too pretty.

As a final point, I think it's a bad name, and a leading one. After all, who would say they're going to practice Irrational Medicine? Its a nonsense. To think we are rational beings is also a fallacy. Kahneman in Thinking, Fast and Slow masterfully expounds this: most of us are tremendously irrational, most of the time.

RE: Can we know anything?

Arlo Guthrie — Tue, 26 Apr 2016 10:39:06 GMT

[quote user="David Mills"]

[quote user="Arlo Guthrie"] and then that we draw our lines in the sand as to what constitutes sufficient evidence (still with you), but then criticise the black cab driver for where they've drawn that line (now you've lost me).[/quote]

Sorry, the black cab driver logic was not a criticism of pragmatism in its truest philosophical sense[/quote]

Sorry, my mistake, it sort of read that way to me

[quote user="David Mills"]

[quote user="Arlo Guthrie"]It's not just a bit unlikely, it's off the scale improbable.[/quote]

See this is more promising, though off the scale is unhelpful.[/quote]

Well, I'm an ex-PR man, and you're a scientist :)

All the same, I do have an innate love of science and I agree with you. I would much rather express myself in absolute terms. Problem is, I can't think of a way of putting a number to the improbability of something which would require (as I said earlier) the observations - scientific and otherwise, not just of the proposition itself, but also of the many other things that would be required in order for the proposition to be true - of millions of people.

I mean there would be so many 0000's involved that I think 'off the scale', whilst not a term that you as a scientist are comfortable with, is a fair reflection of the reality.

[quote user="David Mills"]What I detest is absolutes, and being told such.[/quote] OK.

[quote user="David Mills"]I don't believe there is enough evidence against homeopathy to say it definitely doesn't work, only it is very unlikely to work. Not the same thing.[/quote]

And therein lies the problem. Your job is scientific. Mine is to communicate. As a scientist, you detest absolutes and PR language. I sympathise (you'll never know how much hyperbole I have to strip from press releases).

BUT, when communicating a scientific concept to non-scientific people, it is necessary to speak their language in a way that accurately reflects reality.

Take your description above. You say there is not enough evidence against homeopathy to say it definitely doesn't work. Given the number of fundamental scientific principles that would need to be broken in order for it to work, I presume you mean this in the same way as you might say that there is not enough evidence to say the moon isn't made out of cheese (after all, only six moon landings so far).

As a scientist, you are only comfortable with 'very unlikely that homeopathy works'.

As a communicator, I am very uncomfortable with that, because I think it gives patients and owners an unrealistic hope. Something that is 'very unlikely' to work still might. It's still worth a go. Still worth spending money on.

In your world, I think the current evidence should allow you to say more than 'it's very unlikely that homeopathy works'.

More accurately, you might say: It's very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very²⁰⁰⁰ unlikely that homeopathy works

But it's not a terribly economical use of words (and how many verys is right?). The more succinct way of expressing it in a way that communicates the practical day-to-day implication for patients and owners is simply:

It doesn't.

RE: Can we know anything?

Martin Atkinson — Tue, 26 Apr 2016 09:16:59 GMT

RE: Can we know anything?

Francisco Gomez — Tue, 26 Apr 2016 08:42:06 GMT

[quote user="David Mills"]

[quote user="Francisco Gomez"]Or else I would have to find a rational argument to justify the change...[/quote]

Interesting. Do you think you were rational in arriving at your current conclusions about the world? Or was it simply what you were taught along the years and your experience roughly correlated with?

[/quote]

Are those mutually exclusive? If they're not, yes to both.

[quote]

[quote user="Francisco Gomez"]One or way or the other, modern physics found a justification for it, and I feel safe to say that minor changes in how physics work over traditional physics makes no difference in how I treat animals today.[/quote]

We're back in Kuhnian territory. No-one throughout the ages, overall, has wilfully misinterpreted the world or deliberately constructed wrong theories of science. It is likely, if history teaches us anything that science as we know it now will be unrecogniable and the source of much mirth 200 years hence. We can't see it because at the moment it works, though there are cracks - quantum theory, string theory, hell even gravity defintion are challenging the bedrock of physics and they are not easily explaiinble using current models.

Regarding treating animals, thinking about the basis about what you think you know is enlightening and leads to better medicine. It is the call to arms of EBVM.

[/quote]

Again what you're saying is tomorrow's medicine might be today's reiki. I disagree, purely because there is a strong drive towards unbiasedness, current methods and models purposely try to take this into account, and they (those who investigate) agree that they can't be 100% unbiased.

I guess you're next question would be about being unbiased being necessarily the ultimate benchmark about science and medicine. And the answer is 'yes it is', cause otherwise I would be calling a mechanic each morning before I drive to work, you know, to make sure the car works and not that the last thousand times the Micra took me to work was by chance.

[quote]

[quote user="Francisco Gomez"]I don't think you can compare it to the level of rationale behind an homeopathic remedy.[/quote]

Au contraire. There is nothing a priori that scieintific knowledge has that homeopathic knowledge doesn't.

[/quote]

I'd guess the clue is in 'a priori' right?

[quote]

How do you respond to their charge that of course science doesn't show a benefit of our therapies because its the wrong epistemology? [/quote]

I answered this in my previous post. We build over something that is hard enough, had proven to be cemented and is going to take loads to move it, if we can even move it. Accepting that argument puts in question that the Earth is round, which in turn questions GPS functionality, which in turn puts in question Pythagoras. You're pretty much asking me to discredit Pythagoras theorem because we can't know for sure if the same rules applied to geometry two thousand years ago.

RE: Can we know anything?

Malcolm Ness — Tue, 26 Apr 2016 08:24:41 GMT

[quote user="David Mills"]We use induction in EBM most strikingly in interpreting study results - p-values etc allow some probabilification of whether the null hypothesis is likely to be true.[/quote]

[quote user="David Mills"]Moreover, the EBM movement (rational veterinary medicine if you like) thinks mechanism of action is decidedly unimportant.[/quote]

EBM is emphatically NOT what you are describing here. EBM is a methodology for practitioners to use to evaluate ALL of the relevant evidence (including experience, anecdote, consensus, eminent opinion as well as the published literature) in order to address a SINGLE clinical question framed around a patient centred problem.

The selection of a paper from the literature and it pseudo-philosophical analysis in order to answer or refute wide-ranging, and often ill-defined general medical questions is a hobby of feckless academics and is, in my opinion, quite worthless.

If EBM is to be criticised, it must first be understood. Authors involved in the Cochrane Collaboration have published widely on the subject. 'How to Read a Paper' by Greenhalgh, Pub BMJ. is a good starting point that will lay to rest many of the misconceptions that are apparent in this and other veterinary discussions.

RE: Can we know anything?

Niall Taylor — Tue, 26 Apr 2016 06:41:10 GMT

[quote user="David Mills"]We can hardly begin to call upon mechanism of action now, EBMers, in our hour of need.[/quote] I don't see why not. Evidence-based medicine has demonstrated itself ill equipped and too naive to cope with the Byzantine convolutions and excuses of alternative medicine.

[quote user="David Mills"]... we owe it to ourselves and our patients not to be so complacent and lazy with our epistemology.[/quote]

I would disagree, we owe it to our patients to leave abstract philosophising such as this behind at the consulting-room door and give our clients some real world answers.

I genuinely enjoy these debates, and you clearly know a lot about the philosophy of science which makes them better still, but in the end such arguments are circular and self-defeating and no basis for coping in the real world, including practicing veterinary medicine.

[quote user="David Mills"]Most or all good scientists would disagree with you... Fallacious logic[/quote]

And presumably you know these assertions to be true by virtue of the same criteria which tell you how true it is that nothing is true?

Niall

RE: Can we know anything?

Braden Collins — Tue, 26 Apr 2016 01:58:01 GMT

When I worked around the western tip of Cornwall, one of the old farmers told me that his dad believed urinating in a cows ear was a treatment for milk fever. (True story. MAybe it just annoyed a cow so much once that it struggled and got up?) I was still quite happy to group that theory with homeopathy into the "so unlikely to work that we can pretty safely write it off" group.

I have always liked the expression "Keep an open mind, but not so open that your brain falls out".

RE: Can we know anything?

David Mills — Tue, 26 Apr 2016 00:49:48 GMT

[quote user="Francisco Gomez"]Or else I would have to find a rational argument to justify the change...[/quote]

Interesting. Do you think you were rational in arriving at your current conclusions about the world? Or was it simply what you were taught along the years and your experience roughly correlated with?

Regarding treating animals, thinking about the basis about what you think you know is enlightening and leads to better medicine. It is the call to arms of EBVM.

[quote user="Francisco Gomez"]I don't think you can compare it to the level of rationale behind an homeopathic remedy.[/quote]

Au contraire. There is nothing a priori that scieintific knowledge has that homeopathic knowledge doesn't. How do you respond to their charge that of course science doesn't show a benefit of our therapies because its the wrong epistemology? Don't forget that history teaches us that new or freshly packaged theories are always rejected initially before being accepted later.

Sorry, the black cab driver logic was not a criticism of pragmatism in its truest philosophical sense, but merely that when confronted with difficult questions, scientists often counter well we're too busy working in the real world, mate, of course its x, y, z. This is a logical fallacy - the obviousness fallacy - and predicates itself on "common sense". We as a profession deserve better arguments than this.

[quote user="Arlo Guthrie"]It's not just a bit unlikely, it's off the scale improbable.[/quote]

See this is more promising, though off the scale is unhelpful. I happen to agree, to an extent, in that it is highly improbable homeopathy works. What I detest is absolutes, and being told such. I don't believe there is enough evidence against homeopathy to say it definitely doesn't work, only it is very unlikely to work. Not the same thing.

I am not arguing for homeopathy. I am arguing against the assumptions and packaging of the call to effectively ostracise it.

[quote user="Arlo Guthrie"]And isn't it somewhat ironic that the most commonly used argument for homeopathy is 'I've seen it with me own eyes, guvnor' [/quote]

And in veterinary medicine, alas. Plenty of examples on this forum.

[quote user="Arlo Guthrie"]Surely, therefore, it demands the very highest standards of evidence to be recommended by a doctor or vet?[/quote]

[quote user="Arlo Guthrie"]the Campaign For Rational Veterinary Medicine isn't calling for a ban, just that it should be mandatory that patients are able to give informed consent. In other words, that it becomes mandatory for practitioners to inform owners/patients that the weight of evidence - the crushing weight of evidence (sorry, lapsing into PR-speak but it is 'crushing') - is that homeopathy does not work. [/quote]

See above. This is not something veterinary medicine would be advised to get into. Further scrutiny of our thinking and processes shows them as not too pretty.

RE: Can we know anything?

Francisco Gomez — Mon, 25 Apr 2016 19:57:45 GMT

I can't even sell a metronidazole tablet without asking for a signature, is it really that much harder to ask for one if you intent to supply a product that doesn't even meet most basic POM-V standards?

RE: Can we know anything?

Arlo Guthrie — Mon, 25 Apr 2016 19:48:17 GMT

David Mills, I'm a bit confused how you can say this:

[quote user="David Mills"]I don't think anything can be proven or disproven. It comes down to drawing our lines in the sand as to what constitutes sufficient evidence. [/quote]

Followed by this:

[quote user="David Mills"]You see, it is this pragmatism that sticks in the craw: It's the London black cab driver logic. Oh it MUST be true, guv, I live in the real world, say how things are, spades a spade.[/quote]

What I mean is that you argue nothing can be proven (fine, with you so far), and then that we draw our lines in the sand as to what constitutes sufficient evidence (still with you), but then criticise the black cab driver for where they've drawn that line (now you've lost me).

Given we can prove nothing, surely what constitutes sufficient evidence of a proposition is defined by three things:

It's impact on a person or people
The subjectivity of the proposition
The degree to which the proposition contradicts the observations of a person or people (in which I include scientific and other observations documented over history). In other words the probability of the proposition being correct.

The cab driver sees a spade. He's seen and used a spade all his life. He's seen textbooks with spades. He's been to the garden centre. If that spade turns out to be a banana, it's unlikely to present him with any immediate threat (unless he was trying to dig his way out of a hole at the time). Does he need more evidence that a spade is a spade? Is he to be frowned upon? Surely not.

Clearly the line is a movable one (BTW, just explaining my thinking here, not trying to sound like I'm capable of teaching you anything!), and the burden of proof needed by a general practitioner advising patients is obviously higher than that used by the cab driver to assure himself that the spade is a spade, still higher by the specialist, and thence on to the philosopher.

So we have the proposition that ultra-dilutions of random substances can cure disease. Something that could have a profound impact on humanity. Something which would contradict the observations - scientific and otherwise, not just of the proposition itself, but also of the many other things that would be required in order for the proposition to be true - of millions of people. It's not just a bit unlikely, it's off the scale improbable.

Surely, therefore, it demands the very highest standards of evidence to be recommended by a doctor or vet? Surely your default position should be that it is untrue until the standard of evidence you require proves otherwise? Not the other way around, as you seem to be arguing.

And isn't it somewhat ironic that the most commonly used argument for homeopathy is 'I've seen it with me own eyes, guvnor' (more or less the exact words used to me by a homeopathic vet, not so long ago).

The now numerous studies that have now been conducted and analysed by far brighter minds than my own have also shown at best a placebo effect, and that the better the studies are controlled against the flaws inherent in personal observation, the more likely they are to show no effect.

Does this all weigh in favour of a ban? In fact, the Campaign For Rational Veterinary Medicine isn't calling for a ban, just that it should be mandatory that patients are able to give informed consent. In other words, that it becomes mandatory for practitioners to inform owners/patients that the weight of evidence - the crushing weight of evidence (sorry, lapsing into PR-speak but it is 'crushing') - is that homeopathy does not work.

RE: Can we know anything?

Francisco Gomez — Mon, 25 Apr 2016 16:42:20 GMT

I'm sorry David Mills, I (think I) understand the phylosophical argument, I remain unconvinced however that your point is of value in this particular discussion.

I don't care who ate the cheese, however I base my assumptions on the Earth rotating around the Sun. Do I have proof of it? Well, just like you say, nothing can be proven or disproven. I 'kinda' have proof it, most of my scientific knowledge in general and my veterinary knowledge in particular would fall apart it it wasn't that way. Or else I would have to find a rational argument to justify the change...

One or way or the other, modern physics found a justification for it, and I feel safe to say that minor changes in how physics work over traditional physics makes no difference in how I treat animals today. Mostly because none of them are expected to travel anywhere close to light speed, at least in relation to me or their owners.

Either we accept that there's enough knowledge already, backed and tested by people before us in the most unbiased way that they could find, and we are happy to build over it or we will have to question what is the best time to enter the bathroom every morning.

I can accept that Cushing's disease relies heavily in confidence of diagnosis when I don't have the funds to visualise a tumour. And even then, I can question if the ultrasound machine is showing an artefact. But are you suggesting I call the technician each time I make a diagnosis?

I don't think you can compare it to the level of rationale behind an homeopathic remedy.

RE: Can we know anything?

Martin Atkinson — Mon, 25 Apr 2016 15:20:08 GMT

David which of these are you?:

RE: Can we know anything?

David Mills — Mon, 25 Apr 2016 13:09:54 GMT

[quote user="alex gough"]Could you expand on your assertion that inference or induction is fundamentally illogical?[/quote]

Induction always requires a leap of faith to a conclusion, which in logic theory, makes it illogical.

Consider

the cheese in the larder had bite marks in it

I heard scratching in the larder

Conclusion: a mouse ate the cheese.

We make these kind of inductions/inferences all the time, but the thing to recognise is that the conclusion is not truth, it is a guess. It may be true, and justified by the premises evidence, but the truth, or correctness is not entailed or guaranteed by the premises. The premises underdetermine the conclusion: it could have been the maid, a child, a rat, a ghost, a ghost-rat, aliens, a Welshperson. Hence, it is illogical.

Inference and induction also hide an underlying assumption that Hume called Universality of Nature (UN). To make inductions, we have to assume that nature behaves itself as it has done before. So, when we walk through our door into the parlour, we assume we will reach the parlour and not fall 18 storeys. However we cannot use the fact the door always leads to the parlour as evidence for UN because it is circular - UN 'guarantees' the door leads to the parlour, so we cannot say that the door leading to the parlour is evidence for UN.

Of course, we use induction all the time, else we wouldn't do anything all day. But that is not the point. We can, and do, use induction successfully: but the key is in recognising the limits of what induction can tell us, and maintaining a skepticism that prevents us from over-egging our knowledge as the ultimate truth.

We use induction in EBM most strikingly in interpreting study results - p-values etc allow some probabilification of whether the null hypothesis is likely to be true. Putting aside the problems with statistical philosophy of p-values (why 0.05? Nobody knows) and the like, what studies do not tell us is which hypothesis is true, only whether there was a difference noted in the study and whether the null hypothesis is statitistically likely to be true given the data examined. To find the true hypothesis, we need to test every possible hypothesis - we cannot.

Moreover, we use inference to generalise results from the obersved to the unobserved, adding another illogical layer. The strict epsitemic reading of studies is that the difference is only applicable to those studies - anything else is on shaky ground. So even in our vaunted scientific model nothing can be said to be true, or even nearly true I contest - we can only probabilify the truth. Evidence helps with this justification/calculation, but and in veterinary medicine to talk of overwhelming evidence is laughable - we simply ain't got the evidence to say this about most of what we do.

Popper's response was that to distinguish itself from other sources of knowledge it needs to be logical, hence the hypothetico-deductive scheme he proposed in which the conclusions are entailed by the premises. He also held that a theory can never be proven, only falsified. We may have evidence that fits the theory, but no amount of evidence can prove it conclusively: good scientists recognise this, and factor it into their work. In a way, veterinary science does work like this in the construction of RCTs in EBM - whilst our conclusions may be inductive, the construction of RCTs is Popperian.

[quote user="alex gough"]With this in mind, what level of evidence would you require to consider a treatment disproven? Or would you assert that nothing is ever proven or disproven? And if so how does that affect our clinical practice?[/quote]

I don't think anything can be proven or disproven. It comes down to drawing our lines in the sand as to what constitutes sufficient evidence. However this is subjective, and illogical, and therefore suspect to the vagaries of campaigning, politicalisation, etc etc. So one can becry 'disproven' when what one means is that 'I have evidence to believe that most homeopathy does not work' - this is very different, less dramatic, and a less strong, though correct, message. Banning sits uncomfortably because of these reasons - there is not the epistemic weigth to do so. Beware the man who knows not what he knows, especially when he's selling an idea.

[quote user="Martin Whitehead"]Aside from the fact that "scientific" and "rational" hardly seem to be emotive words (at least, by the definitions of those words I know)[/quote]

Come, come, let us at least debate as learned adults. Scientism (and to a lesser degree rationality) is a well recognised phenomenon. Adding scientific or rational to any claim lends it epistemic weight - just add chronic dieters - by conjuring up notions of objectivity and reliabality. Any reading of philosophy of science shows science isn't objective, from individual studies to scientific trends, it is the subject of much influence, cultural, personal etc; it may be more objective but it isn't purely objective.

[quote user="Martin Whitehead"]would only be true if we have reason to think that different homeopathic remedies have different mechanisms of action[/quote]

This counterfactual would only hold, and even then probably not, if we knew the true, complete mechanisms of all our scientific drugs. Moreover, the EBM movement (rational veterinary medicine if you like) thinks mechanism of action is decidedly unimportant. It' very core, the RCT, does away with it altogether: it was designed as the height of pragmatism - study something to see its 'true' effect. Indeed it was possibly the very reason for EBM - to move away from the (philosophical) rationalist model of basic biology into diease and replace it with empiricism: "what really works". And see how far mechanistic reasoning comes down the list of strength of evidence! We can hardly begin to call upon mechanism of action now, EBMers, in our hour of need.

Besides, and I'm no homeopath, there are lots of mundane arguments about remedey production and the mechanism being different that barely need to be said.

[quote user="Niall Taylor"]n real-world terms, to certainties.[/quote]

Most or all good scientists would disagree with you. See above. Kuhn describes it well - scientisits work within a fallible theory, using models of reality to explain the world. We guess, educatedly, and don't take anything for certain.

You see, it is this pragmatism that sticks in the craw: It's the London black cab driver logic. Oh it MUST be true, guv, I live in the real world, say how things are, spades a spade. However we owe it to ourselves and our patients not to be so complacent and lazy with our epistemology. And then run with this to start talking about absolutes when our evidence ain't got the legs or lifting power to say such.

[quote user="Niall Taylor"]In the same way it is perfectly acceptible to state that homeopathy does not work. [/quote] Fallacious logic.

RE: Can we know anything?

Niall Taylor — Sun, 24 Apr 2016 08:15:01 GMT

[quote user="David Mills"]All the campaign can say is that as far as we know, homeopathy doesn't work and we believe is unlikely to work in most cases.[/quote]

While science doesn't work in absolutes there are many firmly established areas of science which, although still referred to as 'theories' (gravity, evolution etc...) are tantamount, in real-world terms, to certainties. True, if someone came along with another, more plausible theory which fitted all the evidence supporting evolution and provided ways of predicting future findings then this new theory might end up replacing the current theory of evolution. But that is almost certainly not going to happen. In the meantime philosophical 'purists' who insist evolution is 'only a theory' simply provide ammunition for non-scientific people who argue that God created everything in 6 days.

It is perfectly acceptible, in colloquial speech, to state that evolution is a fact.

In the same way it is perfectly acceptible to state that homeopathy does not work. True, in theory, we may one day discover all the notions of physics, biology, quantum mechanics, chemistry etc... which it apparently violates, have gaping holes in them which noone apart from Samuel Hahnemann was able to see; we may discover the vast numbers of scientific papers and clinical trials which found homeopathy to be no more effective than placebo were all somehow flawed. But we know, in the real world that's not going to happen. And that's not scientism, it's pragmatism.

In the meantime, well meaning philosophical types who point out we can never be absolutely certain that homeopathy doesn't work (while at the same time not quantifying the degree of uncertainty involved) simply provide extra ammunition to homeopaths by giving them leave to persist in their fallacy that it's possible to cure cancer with sugar tablets.

Homeopathy is indistinguishable from placebo, there has never been good quality evidence showing it is effective and plenty which shows it to be ineffective. All the successes claimed by homeopaths are easily explicable by other means (coincidence, regression to the mean, mis-diagnosis etc).

It is therefore reasonable to work on the 'assumption' that homeopathy is ineffective and call for its use to be restricted.

Niall

RE: Can we know anything?

Martin Whitehead — Sat, 23 Apr 2016 17:02:04 GMT

[quote user="David Mills"]wallpapered with emotive words such as 'scientific' and 'rational' [/quote]

Aside from the fact that "scientific" and "rational" hardly seem to be emotive words (at least, by the definitions of those words I know), how are we to discuss a subject for which science and rational thought are fundamental to that discussion, without using words such as "scientific" and "rational"?

David, you appear to be arguing that, although (i) quite a few homeopathic remedies have been shown to be ineffective, and (ii) there is no definitive evidence of effectiveness (beyond placebo) for any homeopathic remedy, we should not argue on the basis of that evidence that homeopathy as a whole is ineffective because there might be other homeopathic remedies that haven't yet been tested (or even produced) that may be effective.

However, leaving aside the fact there is no known 'scientific' or 'rational' (sorry to be so emotive!) mechanism of action of homeopathic remedies, your argument would only be true if we have reason to think that different homeopathic remedies have different mechanisms of action. If we test one type of conventional antibiotic drug for a particular disease and find it to be effective or ineffective, we know that we cannot generalise that result to other types of antibiotic, because we know that other antibiotics have different mechanisms of action (i.e., they disable/kill bacteria in different ways). If all homeopathic remedies have the same mechanism of action then, presumably, clinical trial testing need only show that relatively few remedies are ineffective for us to be able to conclude with reasonable certainty that all of them are ineffective - and I would argue we are in that position now. As I understand it, homeopaths themselves believe that - while they have to take care to select the right remedy to fit the 'symptom picture' of each patient - all their remedies 'work' by the same mechanism (often described in terms of 'balancing' an out-of balance flow of energies in the body). Do you have reason to believe that different homeopathic remedies have different mechanisms of action?

Robert Mathie (of the British Homeopathic Association) one told me that there are about 3,000 remedies in the homeopathic Materia Medica, each with its own specific indications. As far as I am aware, homeopathic remedies can be made from absolutely anything - any species of animal or vegetable or microorganism, any mineral, any substance artificial or natural, organic or inorganic, and even from electromagnetic energy (there are remedies made from light from the planets Venus and Saturn, and from emanations from computers and TVs) - so, presumably there are many millions, at least, of possible remedies. If the remedies each have different mechanisms of action - each of which can be transmitted via water, often even in the absence of any of the original substance - how many remedies would need to be shown to be ineffective in clinical trials (bearing in mind that 5% of them would come up 'positive' by chance at a p<0.05 significance level) before we could confidently conclude that homeopathy as a whole is ineffective?

Martin