Posted by LlurpsieBlossom on January 6, 2007, at 19:00:50
In reply to Generics » linkadge, posted by ed_uk on January 6, 2007, at 16:45:45
> Hi Link,
>
> >So, generics are allowed to be as much off as +/- 20%?
>
> >If you were taking 375mg of effexor, you could have as much as much as 150mg difference between two doses.
>
> No, the amount of drug in a generic tablet/cap will be almost identical to the amount of drug in a branded tablet/cap. When comparing a new generic product to the orginal branded product, bioequivalence studies are performed. In the UK, the following criteria are assessed:
>
> Product A = the test product ie. the new generic
> Product B = the reference product ie. the original brand
>
> A group of volunteers are given the medication. Pharmacokinetic parameters (AUC, Cmax and Tmax) are determined for each product (A and B). The ratio of A/B for AUC and Cmax and 90% geometric confidence intervals for AUC and Cmax respectively are determined. These values must fall within the accepted regulatory range of 80-125% (0.80-1.25 range) to indicate that A is bioequivalent to B.
>
> EdOh goody! statistics!
how come the confidence interval is so large? I thought that the alpha for most peer-reviewed studies examining significant differences was .05?In my field we use a 95% confidence interval when we report such things.
Hmm false positives occur more often as alpha increases. Thus, a significant difference between brand A and brand B would be more likely to be detected when alpha is larger. I guess having a larger alpha is in the best interests of detecting differences between brand A and brand B.
Now the crucial statistic here is a FAILURE to find a difference, when one actually exists. In order to minimize this danger, one would want to have a large sample size, thus having enough ´´power´´ to detect a difference.
see, if me and eduk and linkadge and phillipa are the only volunteers to test a set of drugs, we may all have very different patterns of response to brand A.
me .1
eduk 1
linkadge 10
phillipa 100and our average is 111.1/4 or about 27, but our standard deviation is HUGE!!!
but, if we test a whole BuncH!! of people the outliers are much less likely to affect the statistics.
then, when we find out that the average for drug A is 25 with a standard deviation of plus or minus 3 and another drug B has average of 43 with a standard deviation of .01, then we know that theres something kooky. someone spiked the punch, you know ?
okay, I have statistics withdrawal. oh, how I miss my silver computer with all the SPSS and SAS and all that good stuff. heck, Id just settle for having an American keyboard. I can{t get apostrophes and these ñññfunny letters keep popping up all over the place!
`CIAO MY LOVELIES,
LLUIR´STJHPIE
poster:LlurpsieBlossom
thread:719885
URL: http://www.dr-bob.org/babble/20070101/msgs/719958.html