Act Acting » Method Acting » DELTA

DELTA

Question:

: : I think every one is probably nervous about such a high dropout number.. : — : All these nervous assumptions that are made make me nervous. Isn’t this : called "faith, not Science" when I allegedly do it? : No, it is science because one can (and I am sure did) test the comparability of : the drop outs.  When this is done it represents faith in the method.  That is : different that assuming a fact on faith (Multiple infections cause aids). You are acting as if the latter cannot be tested also. Californ

No, I am acting as though one has beeen tested and analyzed while there is insufficent data to warrent investing much in testing the other.

Response:

- Hide quoted text — Show quoted text – : read in SCIENCE that a remote number (pun intended) of 54% of the ACTG 175 : cohort dropped out of the trial (i.e. discontinued study treatment). If more : than every second trial participant walks out of a trial, how can one : seriously assert that such a trial found anything? : Carefully and nervously.  One has to assume (or test in some way) that dropouts : were comparable in both groups. : I think every one is probably nervous about such a high dropout number.. — All these nervous assumptions that are made make me nervous. Isn’t this called "faith, not Science" when I allegedly do it?

No, it is science because one can (and I am sure did) test the comparability of the drop outs.  When this is done it represents faith in the method.  That is different that assuming a fact on faith (Multiple infections cause aids).

Response:

: : I think every one is probably nervous about such a high dropout number.. : — : All these nervous assumptions that are made make me nervous. Isn’t this : called "faith, not Science" when I allegedly do it? : No, it is science because one can (and I am sure did) test the comparability of : the drop outs.  When this is done it represents faith in the method.  That is : different that assuming a fact on faith (Multiple infections cause aids). You are acting as if the latter cannot be tested also. Californ

Response:

: read in SCIENCE that a remote number (pun intended) of 54% of the ACTG 175 : cohort dropped out of the trial (i.e. discontinued study treatment). If more : than every second trial participant walks out of a trial, how can one : seriously assert that such a trial found anything? : Carefully and nervously.  One has to assume (or test in some way) that dropouts : were comparable in both groups. : I think every one is probably nervous about such a high dropout number.. — All these nervous assumptions that are made make me nervous. Isn’t this called "faith, not Science" when I allegedly do it? Californ

Response:

– Hide quoted text — Show quoted text – First I should say that given what is known about slowing progression, many trials employing a placebo control would be unethical.  Survival is not the only meaningful endpoint.  Once a favorable effect is shown the effective therapby becomes standard therapy and this is the ethical reference point. This question of "ethics" was the real tragedy of the first study of AZT.  Even though 1)some of the original investigators have said that they were premature in stopping the trial,

The investigators did not stop the trial, an independent data safety monitoring board stopped the trial.  The investigators may or may not have been in agreement but it was not their decision.  Hindsight is always 20:20. 2)the TV show 60 Minutes did a piece that documented fraud in this trial, 3)John Lauritsen obtained papers from the Freedom of Information Act that also suggested fraud was rampant,

The impact of the "fraud" has been analyzed and found not to have affected the validity of the conclusions. THIS study was the one that made it "unethical" to *not* provide AZT.

No, it made it unethical for those who accepted its results to give placebos.  That is not the same thing at all.  In any event, If you had looked up the article I posted info about a few days ago you would know that not everyone has refrained from giving placebos.  You are incorrect in inferring (or believing someone who claims) that this is the only placebo controlled study. This, IMHO, was far worse than using this study to approve AZT to treat HIV/AIDS.  It made it unethical, as Dr Holzman says, to not give AZT and incidentally, to *not* use placebo controls for further studies.

At least this is labelled as opinion.

Response:

While I am at it … and conceding that I am *not* a statistician … I just read in SCIENCE that a remote number (pun intended) of 54% of the ACTG 175 cohort dropped out of the trial (i.e. discontinued study treatment). If more than every second trial participant walks out of a trial, how can one seriously assert that such a trial found anything?

Carefully and nervously.  One has to assume (or test in some way) that dropouts were comparable in both groups. Did they follow-up those  who discontinued?

I don’t know for sure but it is usual to follow people at least to document survival. I found this high number very disturbing and wonder now how reliable the claims in the end are which were made by the PIs? Any comments on this one?

I think every one is probably nervous about such a high dropout number, which even made the news pages of Science.  I am reluctant to opine more because all I know of the study is what I heard at the meetings.

Response:

I agree with you about the nature of the substitution but I think that, with the limitations on all such generalizations, the current knowledge of the effects of  AZT are appreciated in NYC by physicians and patients alike.  I don’t advcate physicians subtituting their judgement for that of their patients, although they should (IMO) be making their informed (?) judgements clear to the patients.

That’s great, if a majority of physicians and PWAs in NYC at least are aware of such things. In Australia, until a couple of weeks ago (literally) it was official state of the true believers in the research/activist/gay editor establishment that AZT confers survival benefits if given early. Largely due to the influence of David Cooper and his gang. I couldn’t even express my doubts in a little Letter to the Editor of the _Medical Journal of Australia_ about a year or two ago, because one of these ‘experts’ (any guess who sponsors the research of these oh so independent scientists?) rejected it :-) . While you talk about informed judgements. I remember my GP who once admitted that he has really not more time than to read the letters pages of the _Medical Journal of Australia_. I doubt that these judgements are in many cases informed. Then, ethicists (and legal ppl) have this funny ‘reasonable physician’ standard which probably means knowing what’s in the letters pages of the professional paper of your respective medical association :-) . Anyway, if they don’t have more time (and I believe that many really don’t have), there’s a problem with the level of information physicians have and with the level of information patients have who depend on their physicians as the only relevant source of treatments related information. It is important to reflect that there may not be a consensus on what is or is not established and that results in differences about what is and is not ethical.

Why not use the concept of professional community equipoise as an alternative to consensus. If a large number of professionals believes/thinks that there is an equipoise between active agent and placebo, then the placebo is ethically justifiable. Naturally, I don’t think this is the answer and in the end it doesn’t make placebo controlled trial designs more ethical per se, but at least there’s a reasonable response to this objection. While I am at it … and conceding that I am *not* a statistician … I just read in SCIENCE that a remote number (pun intended) of 54% of the ACTG 175 cohort dropped out of the trial (i.e. discontinued study treatment). If more than every second trial participant walks out of a trial, how can one seriously assert that such a trial found anything? Did they follow-up those who discontinued? I found this high number very disturbing and wonder now how reliable the claims in the end are which were made by the PIs? Any comments on this one? Udo

Response:

Thanks for this response, Robert. I am not sure whether we really talked about the same issue.

Sorry if I tried to cover too much and was diffuse as a result.  I think that we are agreed on most points covered, but see the comments below. I wrote.. First I should say that given what is known about slowing progression, many trials employing a placebo control would be unethical.  Survival is not the

You wrote.. Not really. Assumed you can slow progression but you don’t improve survival, then this means only that slowing progression up to a certain point results in better health & quality of life. This in turn you have to balance against the undesirable effects of the drug you give in order to achieve slowing progression w/out improving survival. Hence, in a way what you really do is to substitute survival thru quality of life considerations. I have no objections to this, but I have doubts that most PWAs are aware of the fact that this normative judgement has taken place. If they make a conscious decision to go for this solution that’s naturally acceptable.

I agree with you about the nature of the substitution but I think that, with the limitations on all such generalizations, the current knowledge of the effects of  AZT are appreciated in NYC by physicians and patients alike.  I don’t advcate physicians subtituting their judgement for that of their patients, although they should (IMO) be making their informed (?) judgements clear to the patients. only meaningful endpoint.  Once a favorable effect is shown the effective therapy becomes standard therapy and this is the ethical reference point. The favorable effect, at least in the case of AZT means *may be* slowing progression (definitely not in all patient groups,

Agreed, but see the meta-analysis in the Annals of Int. Med. that I referencedc a day or so ago.   On the statistical issue.  The point I was trying to make was that, given a single endpoint — survival, showing AZT better than placebo (and also given that there is enough corroberation that this is not a type I error), your observation that an agent that was measurably better than AZT should also be better than placebo is correct.  However there are other considerations of experimental design and comparability accross studies carried out at different times and therefore with populations getting different types of ancillary treatment which could affect survival and might lead one to question if the estimates of effect magnitude were comparable or if the differences would continue to achieve statistical significance.  In essence this is the problem of randomized versus historical controls which was so nicely summarized recently. The experimental dilemma is that a placebo study, though unethical, is the efficient way to answer the first question.   It is only unethical when you have clear evidence that the experimental agent is better than placebo.

It is important to reflect that there may not be a consensus on what is or is not established and that results in differences about what is and is not ethical. Placebo studies are not per se unethical.

I never said so and am a strong advocate for them. Best wishes.

Response:

First I should say that given what is known about slowing progression, many trials employing a placebo control would be unethical.  Survival is not the only meaningful endpoint.  Once a favorable effect is shown the effective therapby becomes standard therapy and this is the ethical reference point.

This question of "ethics" was the real tragedy of the first study of AZT.  Even though 1)some of the original investigators have said that they were premature in stopping the trial, 2)the TV show 60 Minutes did a piece that documented fraud in this trial, 3)John Lauritsen obtained papers from the Freedom of Information Act that also suggested fraud was rampant, THIS study was the one that made it "unethical" to *not* provide AZT. This, IMHO, was far worse than using this study to approve AZT to treat HIV/AIDS.  It made it unethical, as Dr Holzman says, to not give AZT and incidentally, to *not* use placebo controls for further studies. Todd

Response:

– Hide quoted text — Show quoted text – [...] First I should say that given what is known about slowing progression, many trials employing a placebo control would be unethical.  Survival is not the only meaningful endpoint.  Once a favorable effect is shown the effective therapby becomes standard therapy and this is the ethical reference point. [...] What a bizarre idea. I wonder if we here in the UK had bought into the idea that everyone with an "HIV+" diagnosis was "entitled" on ethical grounds to receive AZT, would we not be seeing a similar incidence of AIDS to the US? [...] I never said what you say I did.  You claim I said that every one with an "HIV+" diagnosis was "entitled" (YOUR quotation marks, not my quote) to AZT.  

I did not. I quoted exactly what you said at the beginning of the post. You see it right there in front of your nose. What is the point of such mendacity? My use of quotes around "HIV+" and "entitled" is intended to indicate that there is doubt over what "HIV+" actually means and whether subjects in tests could be "entitled" to receive AZT because it would be unethical to refuse it (thereby absolving tests without placebo controls), which is the clear implication of your post. I don’t need to quote it back to you. It is there at the top of the page. It is a bizarre idea. And a very odd concept of what is meant by ethics. It may be possible to draw an exquisite distinction between what is ethical for a test subject and what is ethical for a patient but I am not sure this is your intention. Your explanation of why there are no placebo controls in AZT trials (that it would be unethical to include them) remains totally unconvincing. I am surprised you even offer it. Leaving out these controls has the effect of turning what should be scientific studies into advertisements. This is very shabby science. John

Response:

Sorry to try it here again, but I didn’t get a reply to this when I posted this question to s.m.a. awhile ago. We know that AZT does not increase survival compared to placebo.

Not quite right. We know that AZT reduces survival time of people in only a very small group (sez Concorde) in early vs deferred treatment groups.

Not quite right. Therefore, whenever a combination of drugs increases survival over AZT mono, is it unreasonable to assume that this is also better than placebo even when no placebo control did exist? Or is this unreasonable to argue because in the end, even though much later deferred means they got AZT, too at one

point? Udo

This topic was dealt with on sma quite extensively (look for actg 175 threads) and one post from Greg Folkers (?) of NIH put it well. I wish I had it archived to re-post for you, but I don’t so I paraphrase and interpret: The days of placebo controls are over, wrt to AIDS drugs. Patients won’t tolerate being treated with nothing. Exhaustive placebo trials have already taken place for the old front line treatments and their data are well known, and a baseline established. Now trials are conducted with the old front line drugs as the controls. This is progress and precludes continually starting from square one. Also, from my own point of view: please be aware as you hear trial data, of the difference between longevity, or time from infection to death; and period of good health, or time from infection to AIDS; and the period from AIDS to death. These periods are now becoming statistically distinctive. Cameron

Response:

Personally, I find it unusual practice to cater to the whims of subjects in a study at the expense of scientific methodology.  

How many experiments involving volunteer human subjects have you done?

Response:

Sorry to try it here again, but I didn’t get a reply to this when I posted this question to s.m.a. awhile ago. We know that AZT does not increase survival compared to placebo. We know that AZT reduces survival time of people in only a very small group (sez Concorde) in early vs deferred treatment groups. Therefore, whenever a combination of drugs increases survival over AZT mono, is it unreasonable to assume that this is also better than placebo even when no placebo control did exist? Or is this unreasonable to argue because in the end, even though much later deferred means they got AZT, too at one point? Udo

Response:

- Hide quoted text — Show quoted text – Sorry to try it here again, but I didn’t get a reply to this when I posted this question to s.m.a. awhile ago. We know that AZT does not increase survival compared to placebo. We know that AZT reduces survival time of people in only a very small group (sez Concorde) in early vs deferred treatment groups. Therefore, whenever a combination of drugs increases survival over AZT mono, is it unreasonable to assume that this is also better than placebo even when no placebo control did exist? Or is this unreasonable to argue because in the end, even though much later deferred means they got AZT, too at one point? Udo

First I should say that given what is known about slowing progression, many trials employing a placebo control would be unethical.  Survival is not the only meaningful endpoint.  Once a favorable effect is shown the effective therapby becomes standard therapy and this is the ethical reference point. The question you ask is deceptively simple.  Different experimental designs ask different questions. For a single endpoint, if AZT is equivalent to placebo then you are correct that finding an improvement over AZT is the same as finding an improvement over placebo.  The problem is, if AZT is worse than placebo then it is possible to find diferrences  of the sort seen below. While most of us "establisment" types think AZT has some benefits other than survival, the failure to show a significant improvement over placebo still leaves open the possibility that AZT is worse. (Assume an ordered list of effects and that any underlined effects do not differ significantly after a three arm clinical trial)   Treatment arm:            AZT  Placebo  BetterCombo                                           but not placebo) Thus, whether AZT can substitute for placebo depends (ethics aside) on the precision of the measurement of the difference between AZT and placebo. If, for example, the 95% confidence interval for the difference includes a large range of values which favor placebo then using AZT may cause you to falsely declare the new treament to be effective. Using AZT as the reference group means that in a trial you are no longer asking "does this agent have an effect on AIDS" but rather "is this agent more powerful than AZT".  These are two different questions and both are worth answering.  The experimental dilemma is that a placebo study, though unethical, is the efficient way to answer the first question.  An AZT reference group, the ethical study, is the proper way to answer the second, but an experimental design to show equivalence of AZT and the alternate therapy may be very inefficient (which means, expensive, taking a long time to complete, requiring many participants). A final dilemma is that using an AZT comparison group rather than placebo group complicates evaluation of toxicity of the new agents.  Again, the placebo control asks the question "What are the properties of the new drug" and the AZT control asks "are the toxicities worse, better, or different than AZT". Remember that all these measurements are being made against a very active background of medical care (conventional and alternative) which has all manner of toxicities of its own.  The placebo group provides an estimate of these factors.

Response:

Related Posts

Leave a Reply