|
matterunomama
Aug 8, 2010, 11:13 PM
Post #26 of 47
(5761 views)
Shortcut
Registered: Mar 11, 2005
Posts: 419
|
jt512 wrote: onceahardman wrote: Effectively treating chronic pain (for example) and proving your treatment effective, is much more difficult. There are various pain scales which claim to measure pain objectively, but it can be well argued that pain is at least partly a subjective experience. How does one measure a subjective experience objectively? With a pain scale. Jay
The pain scale is very subjective, it asks for personal opinion "1 being no pain.....10 being the worst pain you ever had" My rating of my broken ribs at 20 would have been higher than now, after childbirth, tooth abcesses...and I am told that cancer pain-which I hope I never have- is the worst. I know that toward the end we have to dope patients up so much that they can no longer talk. What CAN be measured and IS objective is the CHANGE in pain scale before and after an experimental intervention.
|
|
|
|
|
jt512
Aug 9, 2010, 12:19 AM
Post #27 of 47
(5744 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
matterunomama wrote: jt512 wrote: onceahardman wrote: Effectively treating chronic pain (for example) and proving your treatment effective, is much more difficult. There are various pain scales which claim to measure pain objectively, but it can be well argued that pain is at least partly a subjective experience. How does one measure a subjective experience objectively? With a pain scale. Jay The pain scale is very subjective, it asks for personal opinion "1 being no pain.....10 being the worst pain you ever had" My rating of my broken ribs at 20 would have been higher than now, after childbirth, tooth abcesses...and I am told that cancer pain-which I hope I never have- is the worst. I know that toward the end we have to dope patients up so much that they can no longer talk. What CAN be measured and IS objective is the CHANGE in pain scale before and after an experimental intervention.
It might interest you to know that, in principle, the effectiveness of a pain treatment can be measured using a pain scale objectively without pre-treatment measurements. Say you take 2000 patients and randomize half of them to an active treatment for pain and the other half of them to a placebo. After a prescribed period of time on either the active treatment or placebo, you have them rate their pain on a scale, say from 1 to 10. Then, if say the active treatment group's mean pain rating was sufficiently lower than that that of the placebo group, it would be correct to conclude that their actual level of pain was, in fact, lower, and hence that the experimental treatment was successful.
Jay
|
|
|
|
|
matterunomama
Aug 9, 2010, 12:49 AM
Post #28 of 47
(5733 views)
Shortcut
Registered: Mar 11, 2005
Posts: 419
|
jt512 wrote: matterunomama wrote: jt512 wrote: onceahardman wrote: Effectively treating chronic pain (for example) and proving your treatment effective, is much more difficult. There are various pain scales which claim to measure pain objectively, but it can be well argued that pain is at least partly a subjective experience. How does one measure a subjective experience objectively? With a pain scale. Jay The pain scale is very subjective, it asks for personal opinion "1 being no pain.....10 being the worst pain you ever had" My rating of my broken ribs at 20 would have been higher than now, after childbirth, tooth abcesses...and I am told that cancer pain-which I hope I never have- is the worst. I know that toward the end we have to dope patients up so much that they can no longer talk. What CAN be measured and IS objective is the CHANGE in pain scale before and after an experimental intervention. It might interest you to know that, in principle, the effectiveness of a pain treatment can be measured using a pain scale objectively without pre-treatment measurements. Say you take 2000 patients and randomize half of them to an active treatment for pain and the other half of them to a placebo. After a prescribed period of time on either the active treatment or placebo, you have them rate their pain on a scale, say from 1 to 10. Then, if say the active treatment group's mean pain rating was sufficiently lower than that that of the placebo group, it would be correct to conclude that their actual level of pain was, in fact, lower, and hence that the experimental treatment was successful. Jay
If the patients are randomized that's true, and well stated.
|
|
|
|
|
onceahardman
Aug 9, 2010, 9:35 PM
Post #29 of 47
(5678 views)
Shortcut
Registered: Aug 3, 2007
Posts: 2493
|
jt512 wrote: matterunomama wrote: jt512 wrote: onceahardman wrote: Effectively treating chronic pain (for example) and proving your treatment effective, is much more difficult. There are various pain scales which claim to measure pain objectively, but it can be well argued that pain is at least partly a subjective experience. How does one measure a subjective experience objectively? With a pain scale. Jay The pain scale is very subjective, it asks for personal opinion "1 being no pain.....10 being the worst pain you ever had" My rating of my broken ribs at 20 would have been higher than now, after childbirth, tooth abcesses...and I am told that cancer pain-which I hope I never have- is the worst. I know that toward the end we have to dope patients up so much that they can no longer talk. What CAN be measured and IS objective is the CHANGE in pain scale before and after an experimental intervention. It might interest you to know that, in principle, the effectiveness of a pain treatment can be measured using a pain scale objectively without pre-treatment measurements. Say you take 2000 patients and randomize half of them to an active treatment for pain and the other half of them to a placebo. After a prescribed period of time on either the active treatment or placebo, you have them rate their pain on a scale, say from 1 to 10. Then, if say the active treatment group's mean pain rating was sufficiently lower than that that of the placebo group, it would be correct to conclude that their actual level of pain was, in fact, lower, and hence that the experimental treatment was successful. Jay
That does interest me, and it is entirely true. Thank you.
Perhaps it is because of the population of patients I see. I'm pretty sure PTs tend to "get" patients who have been resistant to more traditional types of treatment for pain.
For example (expanding upon your good example above), lets say that 65% of 1000 randomized patients saw a significant improvement in their symptoms, compared to 30% in the placebo group.
That still leaves 35% of 1000 randomized patients who did not benefit from a clinically "proven effective" modality. These are the patients who often get referred to PT. Still a pretty large group of patients. How to treat them? This is where the subjective nature of pain becomes important, and the "art and science" of healing becomes valuable.
Often, I think, the statisticians (as VERY valuable and rightfully respected as they are) tend to forget that even at a 99% confidence interval, still leaves (statistically speaking) 1%, or 3,000,000 people (roughly) in the USA who are outside that interval.
Put another way, if you freeze your legs in a block of ice, and dump boiling water over your head, on average, you are quite comfortable.
Or, the average earthling has one testicle and one breast.
(those last 2 are just jokes, no need to respond, unless you know some other funny examples).
|
|
|
|
|
davidnn5
Aug 10, 2010, 12:58 AM
Post #30 of 47
(5657 views)
Shortcut
Registered: Dec 8, 2009
Posts: 348
|
jt512 wrote: It might interest you to know that, in principle, the effectiveness of a pain treatment can be measured using a pain scale objectively without pre-treatment measurements. Say you take 2000 patients and randomize half of them to an active treatment for pain and the other half of them to a placebo. After a prescribed period of time on either the active treatment or placebo, you have them rate their pain on a scale, say from 1 to 10. Then, if say the active treatment group's mean pain rating was sufficiently lower than that that of the placebo group, it would be correct to conclude that their actual level of pain was, in fact, lower, and hence that the experimental treatment was successful. Jay
That's not exactly rocket science, and it surprises me that it took you so long to provide this response. The million dollar question is:
What size cohort is required to obviate the skew introduced by patient subjectivity in their response to questions about pain?
And if we have an answer to the above, what is the likelihood of obtaining funding for, or finding an abundance of double-blinded randomised controlled trials that meet this threshold?
|
|
|
|
|
reno
Aug 10, 2010, 2:39 AM
Post #31 of 47
(5642 views)
Shortcut
Registered: Oct 30, 2001
Posts: 18283
|
davidnn5 wrote: That's not exactly rocket science, and it surprises me that it took you so long to provide this response. The million dollar question is: What size cohort is required to obviate the skew introduced by patient subjectivity in their response to questions about pain? And if we have an answer to the above, what is the likelihood of obtaining funding for, or finding an abundance of double-blinded randomised controlled trials that meet this threshold?
And the cynical question: How do you determine that the patient is telling the truth? It's well established that some analgesic drugs provide the side effect of a "high", while others do not (see: Demerol vs. Toradol.) I've had hundreds... HUNDREDS... of patients who looked at me calmly, without a worry on their face, and said "My pain is still a 10/10... the only thing that works for me is Demerol", when I'd been dosing them with enough Morphine to snow a horse.
I *knew* the patient was lying/full of shit/looking for a high.
How does that figure into your "objective" measurement of a subjective symptom?
|
|
|
|
|
jt512
Aug 10, 2010, 3:03 AM
Post #32 of 47
(5632 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
davidnn5 wrote: The million dollar question is: What size cohort is required to obviate the skew introduced by patient subjectivity in their response to questions about pain?
There is no "skew" due to such subjectivity, and if there were, a larger sample size would not diminish it.
In reply to: And if we have an answer to the above, what is the likelihood of obtaining funding for, or finding an abundance of double-blinded randomised controlled trials that meet this threshold?
An abundance of double-blind randomized controlled trials of glucosamine for arthritis have already been conducted, so I guess the likelihood is 100%.
Jay
(This post was edited by jt512 on Aug 10, 2010, 3:15 AM)
|
|
|
|
|
jt512
Aug 10, 2010, 3:10 AM
Post #33 of 47
(5630 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
reno wrote: And the cynical question: How do you determine that the patient is telling the truth?
The statistical answer to your cynical question is that in a double-blind randomized controlled trial is that lying, or any other type of measurement error, is not fatal to the validity of the study. Due to patient blinding and randomization, lying would be approximately equal between the active treatment and the control group, and hence the effect of lying would wash out in the computation of the effect of the treatment compared with the control.
Jay
(This post was edited by jt512 on Aug 10, 2010, 3:13 AM)
|
|
|
|
|
davidnn5
Aug 10, 2010, 4:45 AM
Post #34 of 47
(5614 views)
Shortcut
Registered: Dec 8, 2009
Posts: 348
|
jt512 wrote: davidnn5 wrote: The million dollar question is: What size cohort is required to obviate the skew introduced by patient subjectivity in their response to questions about pain? There is no "skew" due to such subjectivity, and if there were, a larger sample size would not diminish it. In reply to: And if we have an answer to the above, what is the likelihood of obtaining funding for, or finding an abundance of double-blinded randomised controlled trials that meet this threshold? An abundance of double-blind randomized controlled trials of glucosamine for arthritis have already been conducted, so I guess the likelihood is 100%. Jay
Let me put this a little differently. In an experiment where you can see objective results (for example, flipping a coin), there is only one factor that needs to be considered: the outcome - in this case heads or tails.
In a trial considering the relative level of pain experienced by the subject, there are several significant factors to be considered: the outcome reported (1-10), the patient's previous experience of and perception of pain, and the possibility of the patient lying.
The only way I can think of to adjust for these is to scale the trial to the point where the law of averages prevails, in line with your suggestion. But again, at what cohort size does this occur?
There is also the potential for population skew. Let's say your cohort was all from Detroit, where everyone gets shot daily.... 
I'd be interested to know further details about the glucosamine trials, and how they adjusted for the above factors. My understanding was that the jury's still out on the demonstrable efficacy of glucosamine (although undoubtedly safe, and most likely cost-effective).
Edit for atrocious typos.
(This post was edited by davidnn5 on Aug 10, 2010, 4:49 AM)
|
|
|
|
|
onceahardman
Aug 10, 2010, 8:02 PM
Post #35 of 47
(5572 views)
Shortcut
Registered: Aug 3, 2007
Posts: 2493
|
reno wrote: davidnn5 wrote: That's not exactly rocket science, and it surprises me that it took you so long to provide this response. The million dollar question is: What size cohort is required to obviate the skew introduced by patient subjectivity in their response to questions about pain? And if we have an answer to the above, what is the likelihood of obtaining funding for, or finding an abundance of double-blinded randomised controlled trials that meet this threshold? And the cynical question: How do you determine that the patient is telling the truth? It's well established that some analgesic drugs provide the side effect of a "high", while others do not (see: Demerol vs. Toradol.) I've had hundreds... HUNDREDS... of patients who looked at me calmly, without a worry on their face, and said "My pain is still a 10/10... the only thing that works for me is Demerol", when I'd been dosing them with enough Morphine to snow a horse. I *knew* the patient was lying/full of shit/looking for a high. How does that figure into your "objective" measurement of a subjective symptom?
I kind of agree, but with a bit of different perspective.
I don't think a chronic pain patient is "lying" necessarily, when he complains that "only demerol" works for me.
As I said earlier, I think chronic pain is primarily a brain problem, and I think that brains become confused when withdrawing from pain meds. This confusion results in complaints of *onset of pain* when the real problem is not onset of pain, but *withdrawal from meds*. I've had some very close friends addicted to pain meds, and suspect we all know someone who is, even if you aren't aware of it. The problem is very common.
This, I think, is largely the result of too much reliance upon statistical models for pain meds, and then pain meds getting overused, because they are "proven" to be more effective than placebo, leaving (in my example) 35% of people treated without good effect.
Clearly, statistical models which measure pain on a visual analog scale (VAS) are reliably measuring something, but what they are measuring is not always "pain", because pain so much varies between individuals.
The same injury to the same individual will be scored differently based upon the circumstances present at the time of the injury. The statistical treatment of the data may well be mathematically without error, but the data might not be measuring what you intend it to be.
Subscribing "proven" pain meds is not an unreasonable first try for an MD, but when they don't work well (this is a frequent occurence), then the "proven" meds should be discontinued, IMO.
|
|
|
|
|
jt512
Aug 10, 2010, 8:14 PM
Post #36 of 47
(5564 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
onceahardman wrote: Clearly, statistical models which measure pain on a visual analog scale (VAS) are reliably measuring something, but what they are measuring is not always "pain", because pain so much varies between individuals. The same injury to the same individual will be scored differently based upon the circumstances present at the time of the injury.
It doesn't matter very much. In the typical pre-post design with a control group, differences among individuals that are consistent between pre and post measurements will be completely controlled for, while individual-level inconsistencies between pre and post measurements will just add random noise to the data, and hence will bias the treatment effect toward the null hypothesis. Therefore, it is a conservative error. If a treatment is found to be statistically significantly better than placebo in the presence of such noise, then the study actually underestimates the true difference between active treatment and placebo.
Jay
|
|
|
|
|
dingus
Aug 10, 2010, 8:20 PM
Post #37 of 47
(5559 views)
Shortcut
Registered: Dec 16, 2002
Posts: 17398
|
|
|
|
|
|
welle
Aug 10, 2010, 8:27 PM
Post #38 of 47
(5553 views)
Shortcut
Registered: Aug 8, 2008
Posts: 212
|
dingus wrote: welle wrote: WTF is science-based medicine? what on oxymoron as in medicine is not science based? that's called voodoo. Herbal-based medicine? Folk-lore based medicine. Chinese medicine? I think there are many medicine's that are not based on science of any kind whatsoever. AM band talk radio on a Saturday morning will prove hugely instructive in this matter, a strange universe where vitamins cure cancer, baldness and erections galore! DMT
Dingus, I think some herbal and Chinese medicine are based on empirical evidence, which I consider scientific method as well...
|
|
|
|
|
dingus
Aug 10, 2010, 8:30 PM
Post #39 of 47
(5552 views)
Shortcut
Registered: Dec 16, 2002
Posts: 17398
|
welle wrote: dingus wrote: welle wrote: WTF is science-based medicine? what on oxymoron as in medicine is not science based? that's called voodoo. Herbal-based medicine? Folk-lore based medicine. Chinese medicine? I think there are many medicine's that are not based on science of any kind whatsoever. AM band talk radio on a Saturday morning will prove hugely instructive in this matter, a strange universe where vitamins cure cancer, baldness and erections galore! DMT Dingus, I think some herbal and Chinese medicine are based on empirical evidence, which I consider scientific method as well...
I deleted my post as I read I realized it was late to the party, welle.
So I won't argue the science behind tiger ball soup.
Are you the welle I know from SP?
Cheers
DMT
|
|
|
|
|
welle
Aug 10, 2010, 8:31 PM
Post #40 of 47
(5550 views)
Shortcut
Registered: Aug 8, 2008
Posts: 212
|
yep :)
|
|
|
|
|
onceahardman
Aug 10, 2010, 8:44 PM
Post #41 of 47
(5543 views)
Shortcut
Registered: Aug 3, 2007
Posts: 2493
|
Jay, Thanks, but I understand this.
What I'm trying to get at is, there remain large groups for whom "proven effective" pain meds are actually not effective.
The stats can be done perfectly, but the patient still has symptoms. He should be better, statistically speaking, but he isn't.
Like the old surgeon's joke, "The surgery was a success, but the patient died."
|
|
|
|
|
jt512
Aug 10, 2010, 9:11 PM
Post #42 of 47
(5531 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
onceahardman wrote: What I'm trying to get at is, there remain large groups for whom "proven effective" pain meds are actually not effective. The stats can be done perfectly, but the patient still has symptoms. He should be better, statistically speaking, but he isn't.
Only if you misunderstand the statistics. Studies report the percentage of responders in each treatment arm of the study. I think that most competent doctors understand that the percentage of positive responders to many treatments is less than 100%. Even in the glucosamine study I linked on the first page, only 40% of gluocsamine subjects responded to the treatment. Glucsoamine was "successful" (in part) because only 21% of placebo subjects "responded" to the placebo. So despite what you would call the "proven effective[ness]" of this treatment, no one should be even remotely surprised if an individual patent doesn't get a benefit.
Jay
|
|
|
|
|
jomagam
Aug 10, 2010, 9:42 PM
Post #43 of 47
(5524 views)
Shortcut
Registered: Nov 10, 2008
Posts: 364
|
Even if the Glucosamine works only 21% of the time like placebo, that's still more than 0. Shouldn't pharma companies claim to have a cure for everything and worst case you only get the placebo benefits with no side effects ? Only half joking.
|
|
|
|
|
jt512
Aug 10, 2010, 9:53 PM
Post #44 of 47
(5522 views)
Shortcut
Registered: Apr 12, 2001
Posts: 21904
|
jomagam wrote: Even if the Glucosamine works only 21% of the time like placebo, that's still more than 0. Shouldn't pharma companies claim to have a cure for everything and worst case you only get the placebo benefits with no side effects ?
No, they shouldn't
In reply to: Only half joking.
If you were half joking, then shouldn't your post have been at least a little funny?
Jay
|
|
|
|
|
hafilax
Aug 10, 2010, 10:05 PM
Post #45 of 47
(5517 views)
Shortcut
Registered: Dec 12, 2007
Posts: 3025
|
This thread got me wondering if there are any research groups out there looking at pain from a physiological standpoint. Found one group that claims they can quantify pain using fMRI. A little expensive for practical use. I wonder if any other methods could work (eeg, nerve activity, squids etc.)?
|
|
|
|
|
onceahardman
Aug 11, 2010, 10:31 PM
Post #46 of 47
(5477 views)
Shortcut
Registered: Aug 3, 2007
Posts: 2493
|
jt512 wrote: onceahardman wrote: What I'm trying to get at is, there remain large groups for whom "proven effective" pain meds are actually not effective. The stats can be done perfectly, but the patient still has symptoms. He should be better, statistically speaking, but he isn't. Only if you misunderstand the statistics. Studies report the percentage of responders in each treatment arm of the study. I think that most competent doctors understand that the percentage of positive responders to many treatments is less than 100%. Even in the glucosamine study I linked on the first page, only 40% of gluocsamine subjects responded to the treatment. Glucsoamine was "successful" (in part) because only 21% of placebo subjects "responded" to the placebo. So despite what you would call the "proven effective[ness]" of this treatment, no one should be even remotely surprised if an individual patent doesn't get a benefit. Jay
I agree that most competent physicians have a fair knowledge of statistics, in regards to reading current literature.
But I also think that many competent MDs also tend to look at the studies, and breeze through the conclusions, without really considering the meaning of the stats.
The meat of any study lies in the "Methods" section, not the "Conclusions" section, and especially not in the abstract.
For example, in the study on glucosamine, even though the effect is beyond placebo, statistically speaking, the treatment will be ineffective for >50% of a random population, according to the study itself (even though it is not expressed that way).
Yet, it is touted as proof of effectiveness.
That is not an indictment of statisticians. The stats are what they are. But they need to be understood better by docs, so they can be explained to patients, who depend on the doc to actually treat their condition effectively.
|
|
|
|
|
onceahardman
Aug 11, 2010, 10:49 PM
Post #47 of 47
(5469 views)
Shortcut
Registered: Aug 3, 2007
Posts: 2493
|
hafilax wrote: This thread got me wondering if there are any research groups out there looking at pain from a physiological standpoint. Found one group that claims they can quantify pain using fMRI. A little expensive for practical use. I wonder if any other methods could work (eeg, nerve activity, squids etc.)?
Real quantification of pain is really still only at the animal research phase. Because of the invasiveness required, it will probably be a long time before human studies are truly practical.
Here's a quick example:
http://www.molecularpain.com/content/2/1/31
|
|
|
|
|
|
