kv said: "Sometimes/somethings you understand instantly but othertimes/otherthings you may not--intentionally or un-intentially, I can't say. "

yes, kv, it's (very) funny you notice that, as i've had the same experience with hans. sometimes, i even think he speaks a language he made up all by himself!

__________________
"The need to perform adjustments for covariates...weakens the findings." BMJ Clinical Evidence: Mental Health, (No. 11), p. 95.... It's that simple, guys: bad numbers make bad science.

It is natural...as we can understand homeopathy,energetic/force side, beliefs etc. easily and instantly..others can understand conventional system, materialisitc side, skepticism etc. easily and instantly. It is upto what one have read/studied, understood and follow.

__________________
Homeopathic & Biochemic system existed because Drs.Hahnemann & Schuessler thought differently.
Successful people don't do different things, they do things differently..Shiv Khera

Reading comprehension, Neil. I did not call the errors trivial, I called your point trivial. You point, as I understood it, is that a tool wrongly used makes for a poor result. I agree, but that observation is hardly news.

The problems are not with the controlled trial, the problems are with the way people execute the controlled trial. You are trying to discredit the principle by pointing out that people don't follow it well, but that is a non-sequiteur.

Even if you were right, it would STILL not invalidate the principle.

Your toss of coin conclusion assumes that faults completely invalidate results, but that is not necessarily the case. Also, scientific validity of a theory rarely relies on a single trial result, there will usually be replications.

Quite. What is even cuter, and more pertinent to our discussions, is that whatever the details, this shows how difficult it is to get reliable results, due to biases, vested interests, human fallacy, etc. Yet, after emblazoning this, you proceed to suggest that YOUR trials should be unblinded, and that we should place trust in a "clinical record" that is at best biased and incomplete.

Post-publication review is not protocol-based. It nitpicks the protocols, points out weaknesses, confounders, etc. You can't be fradulent in that. You can, obviously, be wrong, but fraud is not into the picture.

We have simply never come to the point of seriously discussing poor fit of the method, partly due to your own diversions.

As for the gold standard, I am forced to repeat: If you want recognition for a drug, a double-blind trial IS the gold standard. It is what the authorities will demand of you, unless you can provide a VERY good explanation for not using it. No amount of discussion between you and me can change that. It doesn't matter one bit what you or I think of the method, as long as the authorities insist on it.

That is why I have asked you (and I have not noticed you answering):

Why are you suggesting a test of homeopathy?

1) To satisfy your own curiosity?

2) To confirm your belief in it?

3) To pave the way for wider recognition.

In the case of #1 and #2, you can design any protocol you will.
For #3, you will need a double blind protocol, unless you can come up with a very good explanation. And without a double blind protocol, you will need some very distinct results.

Hans

__________________
<i>You have a right to your own opinion, but not to your own facts.</i>

Yes, that is very interesting. Why is it that sometimes you are able to write perfectly understandable English, other times not? Is it just because you don't always bother, or is ambiguity sometimes useful for you?

However, for myself, I make a certain effort to understand you posts. If I don't succeed, I say so, I think that is the fair thing to do. Would you rather I misunderstood you?

As I said, you have obviously not understood the double blind principle.

Hans

__________________
<i>You have a right to your own opinion, but not to your own facts.</i>

quote=MRC_Hans: Reading comprehension, Neil. I did not call the errors trivial, I called your point trivial. You point, as I understood it, is that a tool wrongly used makes for a poor result. I agree, but that observation is hardly news.

The problems are not with the controlled trial, the problems are with the way people execute the controlled trial. You are trying to discredit the principle by pointing out that people don't follow it well, but that is a non-sequiteur. only partly correct. but a better way to look at it, or to understand my point more precisely, is that, apart from execution, there are limits to how the rct can be appropriately applied, as there are limits to anything. now, we disagree to whether homeopathy falls within the span of the rct, and if it does, how easily it can be applied there ... a minimal point i have tried to make is that it is very difficult to design an adequate protocol, certainly for a treatment trial, because of all of the confounders - some of which are simply in the nature of homeopathic process - e.g., the numbers of symptoms we look for, the length of time treatment takes, the generally mild and hard to identify appearance of symptoms in response to the "gentle" cure of homeopathy, etc.

in short, i am not and have specifically stated over and over that i am not questioning the principle of the rct - so please get that straight. what i am questioning is exactly the difficulties in the path of adequately adapting it to measure a difficult object.



Even if you were right, it would STILL not invalidate the principle. as i said, there is no disagreement here; this is a misrepresentation of my views. the clearest case, imo, in which the rct can be applied, is in the proving trial. treatment presents many difficult problems, and may be insurmountable, or nearly insurmountable, or easily surmountable once someone actually finds a way to delimit a particular trial without modifying/mongrelizing real treatment practices too far - you know, it is common to find that something is easy to do, once one has discovered how to do it ... just as, in hindsight, a scientific explanation may seem clear as day, when before it's discovery people might have not seen any way to resolve a particular problem.

Your toss of coin conclusion assumes that faults completely invalidate results, but that is not necessarily the case. Also, scientific validity of a theory rarely relies on a single trial result, there will usually be replications.

this is true. and yet, we still need to nitpick far more extensively than you would ever have suggested, in context of such wholesale error (for whatever reasons) as may be present. for example: the 20 or so trials of classical homeopathy reviewed in the lancet meta-analysis are all flawed by the same class of error and misunderstandings, so replication only replicates the systematic error, the equal of bias of course, making of the 20, essentially, 20 'anecdotes,' or 20 interations of the same error - in short, in such circumstances, quantitative methodology is no better than even informal observations, for correcting itself. then, in the case of the dsm-iv, the apparent widespread influences at work, affecting research, make it really problematic to sort out the wheat from the chaffe.

and, of course, beyond that is another limitation of the rct, in relation to the 'rare' side effect, and that is that the original resulsts are usually incomplete, in terms of risk, which also can affect its original findings regarding efficacy: the example at hand being the ssri's, which exacerbate depression rather than alleviate it. and hrt - all of the findings may be correct, but they are also incomplete, leading to many pitfalls in prescribing, if one prescribes this week, instead of next week when the next trial is published in the lancet - so those considerations also contribute to questioning the adequacy of the rct, to cover the waterfront, so to speak - in other words, another limitation, though still not undermining the principle.

Quite. What is even cuter, and more pertinent to our discussions, is that whatever the details, this shows how difficult it is to get reliable results, due to biases, vested interests, human fallacy, etc. Yet, after emblazoning this, you proceed to suggest that YOUR trials should be unblinded, and that we should place trust in a "clinical record" that is at best biased and incomplete.

my "trials?" i beg your pardon. my prospective trial is unblinded, yet structured around a comparator cohort receiving conventional treatment, and measured against laboratory findings - which, btw, can be blinded. my proving trial, otoh, is blinded. furthermore, i have never insisted on any single standard, blinded, unblinded, quantitative, experimental, or empirical. you are the one who has nailed his dreams to the single standard of the rct, you are the one who imagines that his method is 'good as gold.'

Post-publication review is not protocol-based. It nitpicks the protocols, points out weaknesses, confounders, etc. You can't be fradulent in that. You can, obviously, be wrong, but fraud is not into the picture.

of course, that is my point. the reviewers can make mistakes in evaluating methodology, too; they can make mistakes in calculation; they can misunderstand the demands of adapting rod to object: you see this all the time: dr x writes to the editor concerning dr y's mistakes in his published article; and dr y responds, pointing out how dr x has misunderstood the situation; and dr z writes in next month, demonstrating (sic?) how both x and y are mistaken - and then we find out that all 3 are paid consultants to competing pharmacological enterprises!

i am making it sound worse than it has to be - but it is, nevertheless, no pretty. if you'd stop apologizing for all the inadequacies for a few minutes, and stop imagining that i was trying to cast the rct into the dump, you might make more of a contribution to figuring out how to make the rct work better ... you know, as you have admitted, demand better products, better blinding, better controls over the controlled trial.

We have simply never come to the point of seriously discussing poor fit of the method, partly due to your own diversions.diversions? like my 2 articles that explored the problem in detail? the reason we haven't discussed it more than we have, is that you can't conceive that measuring a year long treatment with a 3 month trial doesn't so badly mongrelize the treatment as to make the trial outcomes meaningless ... or that any of the numerous other problems can conveniently be ignored for purposes of making the rct 'practical.'

As for the gold standard, I am forced to repeat: If you want recognition for a drug, a double-blind trial IS the gold standard. It is what the authorities will demand of you, unless you can provide a VERY good explanation for not using it. No amount of discussion between you and me can change that. It doesn't matter one bit what you or I think of the method, as long as the authorities insist on it.

but we are in position to advise the authorities on the limits of the rct. you know, gold really is gold, and deserves the name 'gold standard.' the rct may be gold for conventional medical efficacy, but it may not apply as clearly to homeopathy. that's the debate. well, that and the problem of skill and honesty of the researcher.

That is why I have asked you (and I have not noticed you answering):

Why are you suggesting a test of homeopathy?

1) To satisfy your own curiosity?

2) To confirm your belief in it?

3) To pave the way for wider recognition.

and i have answered, neither of which responses find their way onto your list (which is pretty typical): a) to establish more realistic (accurate) estimation of the limitations of the controlled trial - to challenge the basis for its trendy popularity these days; b) to contribute to our understanding ... you know, science is about discovering the 'truth.'

In the case of #1 and #2, you can design any protocol you will.
For #3, you will need a double blind protocol, unless you can come up with a very good explanation. And without a double blind protocol, you will need some very distinct results.

Hans

__________________
"The need to perform adjustments for covariates...weakens the findings." BMJ Clinical Evidence: Mental Health, (No. 11), p. 95.... It's that simple, guys: bad numbers make bad science.

Mr hans,

I mean that it may not be always proper or conclusive to base DB studies or other clinical trials unless these are in field, prolonged, life-long and even for many genarations. Otherwise, these can just be "still experimantals".

__________________
Homeopathic & Biochemic system existed because Drs.Hahnemann & Schuessler thought differently.
Successful people don't do different things, they do things differently..Shiv Khera

Of course. I can only agree with that. However, the limits imposed by lack of human honesty or due dilligence are not limiting factors for applicability of rct. If anything, then the contrary, because the rct system not only makes it more difficult to cheat, it also makes it easier to detect cheating, during review.

Exactly, and this is what we should spend our energy discussing.

Well, I think you are mixing things up quite throughly, here. But I think we need to discuss a specific protocol to clear that up.

Well, I'm trying to get your position in rct straight, but I'm afraid you are being quite ambiguous about it.

Well, here we agree, at least. A proving trial ought to be a breeze.

Well, I see less problems than you do.

Adequacy is another matter. I'm sure everybody would be delighted if somebody could come up with an even better solution, but till then we will have to make the best of what we have. As I think I have said before: The present testing methods (including the peer review system) is somewhat like democracy; it is far from ideal, but it is the best we have.

Trial, then, sorry. Yes, exactly: It is unblinded. Sure, the lab assessments can be blinded, but I seem to remember that you consider the complexity of assessing symptoms to be so complicated that it requires the evaluation of a clinician, which can hardly be blinded. Anyway, my question was how you could assume that, considering the problems with blinded tests, an unblinded test could be better?

No, now it is your turn to misrepresent. I am not saying that it is as good as gold. The term "gold standard" is applicable to something that is considered the current authoritative method. Just like elections, with all their problems, are the "gold standard" for democracy. For all I care, you can make your own choice of metal, but the double-blind, randomized, placebo controlled trial is the standard you will need to meet if you go for scientific recognition of drugs.

....Out of time. More later.

Hans

__________________
<i>You have a right to your own opinion, but not to your own facts.</i>

I'm sorry, but I think you will find that the FDA will be quite unwilling to take your advice. Have you ever been dealing with the FDA? You will find that they are very unlikely to take advice from the people they are about to audit.

It will certainly take more than your articles to convince any authorities that conventional testing does not apply to homeopathy.

The honesty and skill of the researcher will be that of the homeopathic researcher. You have agreed with me that the skills appear to be less that ideal, and personally, I will have my doubts about honesty.

I think we both know that part of the purpose of such lists is to make a point, so rewriting them is fine.

Mmmm, while a worthwhile quest in itself, that can not be the purpose of testing homeopathy.

It certainly is, and I don't even see any need for the quotation marks. The purpose of science is even to find The Truth, although we must realize that that is an endless quest.

Now, in the case of homeopathy, IMHO, the first truth we should set out to discover is whether it works.


Hans

__________________
<i>You have a right to your own opinion, but not to your own facts.</i>

Originally Posted by bwv11
only partly correct. but a better way to look at it, or to understand my point more precisely, is that, apart from execution, there are limits to how the rct can be appropriately applied, as there are limits to anything.


Of course. I can only agree with that. However, the limits imposed by lack of human honesty or due dilligence are not limiting factors for applicability of rct. If anything, then the contrary, because the rct system not only makes it more difficult to cheat, it also makes it easier to detect cheating, during review.

as i said, "apart from the execution," that is, not considering cheating, there are limits to anything, including the rct.true or not true? we are obviously agreed that blinded research is not possible in the case of surgery - a limit on the applicability of the rct, regardless of considerations regarding skill or honesty of the researcher. we disagree whether - or how far or how easily - the rct can be applied to homeopathy.

Quote:
now, we disagree to whether homeopathy falls within the span of the rct, and if it does, how easily it can be applied there

Exactly, and this is what we should spend our energy discussing. we have done so - without resolution. i have made the case at some length in my articles.


Quote:
... a minimal point i have tried to make is that it is very difficult to design an adequate protocol, certainly for a treatment trial, because of all of the confounders - some of which are simply in the nature of homeopathic process - e.g., the numbers of symptoms we look for, the length of time treatment takes, the generally mild and hard to identify appearance of symptoms in response to the "gentle" cure of homeopathy, etc.

Well, I think you are mixing things up quite throughly, here. But I think we need to discuss a specific protocol to clear that up. been there done that - the bell trial in particular, in my articles again. feel free to nitpick my articles point by point.


Quote:
in short, i am not and have specifically stated over and over that i am not questioning the principle of the rct - so please get that straight. what i am questioning is exactly the difficulties in the path of adequately adapting it to measure a difficult object.

Well, I'm trying to get your position in rct straight, but I'm afraid you are being quite ambiguous about it. i really don't see any ambiguity: i am stating clearly that rct is best applied to conventional efficacy trials; it is applicable to the h. proving trial, and may be adaptable to a treatment trial though more difficult there. lab experiments (as with rat guts and tumors) are also a good target for the rct, i would add to my list at this time.i assume this clears up any perceived ambiguity?


Quote:
as i said, there is no disagreement here; this is a misrepresentation of my views. the clearest case, imo, in which the rct can be applied, is in the proving trial.

Well, here we agree, at least. A proving trial ought to be a breeze. as i said, this is the clearest case - but the bell trial i reviewed adequately demonstrates how many pitfalls await the researcher who underestimates the problems adapting the rct to real world medical practice in homeopathy. i.e., not a 'breeze,' hans.


Quote:
treatment presents many difficult problems, and may be insurmountable, or nearly insurmountable, or easily surmountable once someone actually finds a way to delimit a particular trial without modifying/mongrelizing real treatment practices too far - you know, it is common to find that something is easy to do, once one has discovered how to do it ... just as, in hindsight, a scientific explanation may seem clear as day, when before it's discovery people might have not seen any way to resolve a particular problem.

Well, I see less problems than you do. right. or, as i would say it, you see less problem modifying (mongrelizing, in effect) treatment for the sake of "practicalities" in research.


Quote:
and, of course, beyond that is another limitation of the rct, in relation to the 'rare' side effect, and that is that the original resulsts are usually incomplete, in terms of risk, which also can affect its original findings regarding efficacy: the example at hand being the ssri's, which exacerbate depression rather than alleviate it. and hrt - all of the findings may be correct, but they are also incomplete, leading to many pitfalls in prescribing, if one prescribes this week, instead of next week when the next trial is published in the lancet - so those considerations also contribute to questioning the adequacy of the rct, to cover the waterfront, so to speak - in other words, another limitation, though still not undermining the principle.

Adequacy is another matter. I'm sure everybody would be delighted if somebody could come up with an even better solution, but till then we will have to make the best of what we have. As I think I have said before: The present testing methods (including the peer review system) is somewhat like democracy; it is far from ideal, but it is the best we have.

this type of statement continues to astonish me, coming from you. in science (and in laws and regulations), if the object is pursuit of 'truth,' with or without quotes , then the 'best we have' must be 'right,' or it is useless, or at least well down the road to useless, and very possibly misguided and destructive, if it legislates on the basis of error (whether due to incompetence, fraud, conflict of interest, bias, or intrinsic limitations of the method).would you concede, at least in principle, that if systematic error in fact underlies the negative findings in controlled trials of homeopathy to date, and if the nature of homoepathic practice makes it very difficult to avoid such errors, then "the best we have" is inadequate - if the 'best' leads to erroneous results in a large percentage of cases?

Quote:
my "trials?" i beg your pardon. my prospective trial is unblinded, yet structured around a comparator cohort receiving conventional treatment, and measured against laboratory findings - which, btw, can be blinded.

Trial, then, sorry. Yes, exactly: It is unblinded. Sure, the lab assessments can be blinded, but I seem to remember that you consider the complexity of assessing symptoms to be so complicated that it requires the evaluation of a clinician, but here i have not talked about assessing symptoms, but recording lab outcomes - you know, Factor X = 2.8, normal range 2.1-3.4. movement of outcomes in the h. cohort into the normal range, across a broad selection of standards, as compared to the conventionally treated cohort, would be quite convincing, i'm sure you'd agree, and is easy to blind. yes? which can hardly be blinded.but i was talking about lab outcomes, not symptoms. Anyway, my question was how you could assume that, considering the problems with blinded tests, an unblinded test could be better?not my assumption. my assumption, or argument, is that the prospective test, complete with comparator conventionally treated cohort and pre, intermittent, and post lab tests, would provide a wealth of useful information. remember, i'm not out there stumping for my favorite candidate, as you are: gold schmold, hans. we don't only need replication, we need independent confirmation: and that means bringing differing methodologies into agreement - the rct and clinical observation and lab tests and basic research all, in the end, must agree.yes? in short, not only replication, but independent verification.


Quote:
my proving trial, otoh, is blinded. furthermore, i have never insisted on any single standard, blinded, unblinded, quantitative, experimental, or empirical. you are the one who has nailed his dreams to the single standard of the rct, you are the one who imagines that his method is 'good as gold.'

No, now it is your turn to misrepresent. I am not saying that it is as good as gold. The term "gold standard" is applicable to something that is considered the current authoritative method. Just like elections, with all their problems, are the "gold standard" for democracy. For all I care, you can make your own choice of metal, but the double-blind, randomized, placebo controlled trial is the standard you will need to meet if you go for scientific recognition of drugs. yes, precisely right - it is the political standard. but 'gold standard' implies, hans, that the standard is immutable, that it is sound, and can be applied with confidence across a well-defined range. the rct may be the best we have, quantitatively, but compared to lab results or basic research outcomes, it plays second fiddle, let alone the argument about empirical evidence.

....Out of time. More later.

Hans
Quote:
Originally Posted by bwv11
we are in position to advise the authorities on the limits of the rct. you know, gold really is gold, and deserves the name 'gold standard.' the rct may be gold for conventional medical efficacy, but it may not apply as clearly to homeopathy. that's the debate. well, that and the problem of skill and honesty of the researcher.


I'm sorry, but I think you will find that the FDA will be quite unwilling to take your advice. Have you ever been dealing with the FDA? You will find that they are very unlikely to take advice from the people they are about to audit. we are informing them right now, if they care to read, or informing people who might inform them ... you know, public debate. neither you nor i will likely get an invitation to the dance, but there is, politically, no reason to discuss this at all if it does not have some impact on some small corner of the universe, you know, like the medical or skeptical communities.

It will certainly take more than your articles to convince any authorities that conventional testing does not apply to homeopathy.

The honesty and skill of the researcher will be that of the homeopathic researcher. You have agreed with me that the skills appear to be less that ideal, and personally, I will have my doubts about honesty.i am quite willing to concede that the homeopathic community could turn out to be no better than the conventional system, and await the day when lancet, jama, n eng j med, bmj, etc uncover as gross a record of deceit and incompetence as they have documented in the case of "your" system. !!


Quote:
and i have answered, neither of which responses find their way onto your list (which is pretty typical):

I think we both know that part of the purpose of such lists is to make a point, so rewriting them is fine. really? well, yes, it does seem you like to rewrite what i (and others) say, and then (at least sometimes) refuse to admit openly that you have misrepresented the matter - dontcha?

hmmmmm....



Quote:
a) to establish more realistic (accurate) estimation of the limitations of the controlled trial - to challenge the basis for its trendy popularity these days;

Mmmm, while a worthwhile quest in itself, that can not be the purpose of testing homeopathy. of course this is a purpose for testing homeopathy: if a more adequate test demonstrates efficacy, then ipso facto it demonstrates the gross failure of the rct in many iterations to date to take an accurate or reliable measure of homeopathy - leading one to examine, i presume, where you and your friends have gone wrong. right? granting the significant "if."


Quote:
b) to contribute to our understanding ... you know, science is about discovering the 'truth.'

It certainly is, and I don't even see any need for the quotation marks. The purpose of science is even to find The Truth, although we must realize that that is an endless quest.

Now, in the case of homeopathy, IMHO, the first truth we should set out to discover is whether it works.


Hans

__________________
"The need to perform adjustments for covariates...weakens the findings." BMJ Clinical Evidence: Mental Health, (No. 11), p. 95.... It's that simple, guys: bad numbers make bad science.

hans said: "Have you ever been dealing with the FDA? You will find that they are very unlikely to take advice from the people they are about to audit."

good one, hans! i almost missed that.

but, certainly, you meant to say, "except in the case of conventional medicine."

you know, it was pogo, wasn't it? who said, "we have met our friend, and he is us." or something like that.

love that dsm-iv!!

__________________
"The need to perform adjustments for covariates...weakens the findings." BMJ Clinical Evidence: Mental Health, (No. 11), p. 95.... It's that simple, guys: bad numbers make bad science.