 |
 |
Ann I have some experience in statistics but am far from being the answer man. But I can tell you the following: there were four groups in this study and the numbers represent the average for each group. Thus, the range of responses in each group would be higher and lower than these numbers. And statistics attempts to answer the question: are these groups really different from each other or are the differences so small that there really is no difference i.e. the drug or, in this case, enzyme had no effect. (by looking at averages and standard deviations) If I remember the results correctly, there was no diff between the control group and the others except for the 1200 mg group. This MIGHT mean that you need 1200 mg to achieve an effect. But I think there was a positve correlation found between the dosages and the results. This correlation, however, was not as significant(alpha was chosen to be .10 instead of .05, i.e. the probability of being right was only 90% not 95%) and wouldnt count if done in a more complete study(more participants). BUT remember, this study was very limited and no professional would recommend its usage until it was backed up with a better study. A patient with PD, however, might want to try something that showed promise even tho the study was only preliminary! For example, I now take 600 mg of CoQ10 each day--I hedged my bet--I didnt go with the 1200 mg due to cost and potential side effects. And, the % change in UPDRS may not be very large i.e. the numbers that we see are the CHANGES in the scale not the total rating. Ie, dont expect a 50-100% improvement even if the enzyme does work. You question the lack of logic in the data i.e. the 300mg group did better than the 600 mg group, etc. Welcome to the real world. Data arent always consistent. This might mean that the enzyme changes its effect based on dosage; it might mean that more data is needed; it might mean that the enzyme has no effect; or that the particpants were not truly randomized. Or all of the above. Sorry, if I've confused the issue even more. Bob ----- Original Message ----- From: "Ann Gibbons" <[log in to unmask]> To: <[log in to unmask]> Sent: Monday, November 04, 2002 1:50 PM Subject: CoQ10 Study Results > The following is a question I submitted to Ask the Doctor, and Dr. Lieberman's reply. > > question > > Dr. Lieberman: The final results of the study -- Placebo +11.99; 300mg +8.81; 600mg +10.82; 1200mg +6.69 are somewhat confusing to me. It would appear that the 600mg group progressed almost as much as the placebo group, and that the 300mg group did better than the 600mg group. Although the 1200mg group progressed the least, and therefore had the best outcome, the trend toward that result is statistically unusual. Can you comment? > > answer > > the comparison among the three groups involves a statistical method that compares the percent change from group to group using the number in each group and the probability the result could have been obtained by chance alone i am not a statistician and it is not a method with which I am well versed although the differences between 600 mg and 1200 mg do not seem great or profound, statistically they are the results with 1200 mg per day were not statistically significant to be so they would have to have a 95% probability they did not occur by chance alone they have a 92% probability they show a trend > abe lieberman > > > I don't think Dr. Lieberman answered my question, which is that the increases in dosage do not show a positive trend. It is true that the 1200mg group progressed the least. But it looks like the 300mg group did better than the 600mg group. It looks like the 600mg group did almost as badly as the placebo group. Something is wrong here. The study should typically show a positive trend from placebo to 1200mg. I am not a statistician, but my husband who has always worked with statistics also questions this. Can we talk about this? > > > Ann Gibbons > > > > ---------------------------------------------------------------------- > To sign-off Parkinsn send a message to: mailto:[log in to unmask] > In the body of the message put: signoff parkinsn > ---------------------------------------------------------------------- To sign-off Parkinsn send a message to: mailto:[log in to unmask] In the body of the message put: signoff parkinsn