@SJ - I should interject that NNT statistics are themselves bogus, because they are derived from randomized trials which are almost always misleading/ useless as a guide to clinical practice.
Also, in real life good medical practice, a drug is tried out and if it doesn't work it is stopped, and pretty quickly.
Trials design ignores this good practice, indeed discourages it (because of analysis being based on the random allocation and 'intention to treat').
For example, an antibiotic may be given for a treatment whose exact diagnosis and sensitivity may not be established at the time the course begins. If the bug is sensitive to the drug, there will probably be a cure; but if either the diagnoses turns out to be wrong, or the bug is insensitive then it won't work, or the drug is swiftly stopped due to side effects then it won't work - yet trails average these in with the results of treating the correct disease with the correct drug.
A drug may therefore be close to 100 percent effective (NNT = 1) when given to 'the right patient' - yet come out as only modestly effective (with a high NNT) when the matching of drug and treatment is suboptimal (and this matching may be very suboptimal in trials where triallists are striving to include the largest possible number of subjects, perhaps being paid per subject).
I should interject that NNT statistics are themselves bogus [etc.]
Well, I have no trouble believing this...
Also, in real life good medical practice, a drug is tried out and if it doesn't work it is stopped, and pretty quickly.
But not for primary prevention!
24 January 2013 at 15:55
*
Note added: False alarm - didn't continue to work - presumed freakish interaction...
News only two days old is that 75 mg aspirin (the tiny tablets used to prevent heart attacks and strokes) produces about 75 percent relief from the pain in my arthriticky knees!
Getting up from a chair or climbing stairs is no longer a torment/ comic book parody of decrepitude; and I am not distracted by continuous burning sensations.
(Of course the joints will objectively be just as bad as before, but it was discomfort rather than functionality that was the major problem.)
Such striking pain relief within a couple of hours, in the context of months of increasing pain and restriction and side effects despite a variety of supposedly much stronger pain killers... well, this is like penicillin for pneumonia all over again - a therapeutic miracle!
The rationale for this treatment? I have none. (75 mg is a long way sub-analgesic, and sub-sub-anti-inflammatory - the effect of this small dose is supposedly on preventing platelet aggregation.) It was just an idea that popped into my head (and thanks to whoever popped it there).
I award myself a spiritual Nobel Prize for this n=1 triumph.
Will it last more than a coupla days? Will side effects prove intolerable? - We shall see what we shall see...
Untitled
6 Comments -
http://cardiobrief.org/2012/06/06/real-world-bleeding-risk-of-aspirin-in-primary-prevention-examined/#respond
24 January 2013 at 10:45
Krill oil or fish oil might help also, provided the side effects of blood thinning and immune dampening aren't a problem for you.
24 January 2013 at 11:12
@D - Thanks. My usage is as treatment - not primary prevention. Very very few drugs are beneficial overall when used as primary prevention.
@MR - I've tried them. They don't help at all.
24 January 2013 at 11:26
Very very few drugs are beneficial overall when used as primary prevention.
Roffle. It's a horrifying day in medical school when you learn about "Number Needed to Treat", and realize that the NNT for most drugs is like 50.
24 January 2013 at 13:55
@SJ - I should interject that NNT statistics are themselves bogus, because they are derived from randomized trials which are almost always misleading/ useless as a guide to clinical practice.
http://www.biomedcentral.com/content/pdf/cvm-2-1-002.pdf
Also, in real life good medical practice, a drug is tried out and if it doesn't work it is stopped, and pretty quickly.
Trials design ignores this good practice, indeed discourages it (because of analysis being based on the random allocation and 'intention to treat').
For example, an antibiotic may be given for a treatment whose exact diagnosis and sensitivity may not be established at the time the course begins. If the bug is sensitive to the drug, there will probably be a cure; but if either the diagnoses turns out to be wrong, or the bug is insensitive then it won't work, or the drug is swiftly stopped due to side effects then it won't work - yet trails average these in with the results of treating the correct disease with the correct drug.
A drug may therefore be close to 100 percent effective (NNT = 1) when given to 'the right patient' - yet come out as only modestly effective (with a high NNT) when the matching of drug and treatment is suboptimal (and this matching may be very suboptimal in trials where triallists are striving to include the largest possible number of subjects, perhaps being paid per subject).
24 January 2013 at 14:23
I should interject that NNT statistics are themselves bogus [etc.]
Well, I have no trouble believing this...
Also, in real life good medical practice, a drug is tried out and if it doesn't work it is stopped, and pretty quickly.
But not for primary prevention!
24 January 2013 at 15:55