Quote:
Originally Posted by johnt
My feeling is that we would benefit more from increasing the frequency of testing. To see why, consider two measurement strategies:
A, test at the beginning of the trial and at the end of a year, say. Suppose, by way of an example, A shows a 12% improvement
B, test every month. B shows a non-compounded 1% improvement each month.
Although both methods have the same final outcome, method B has more statistical power. With modern technology we can test every second of every day.
Increasing the frequency of testing also prepares the ground for the focus of the trial to be on the individual, rather than the cohort.
|
No doubt increasing the frequency of testing would help improve the reliability (although not necessarily the validity) of the tests. As would increasing the sample sizes and sample subgroup cohorts. However, the trial investigators all live in the real world where it's almost impossible to get funding for the studies as currently designed, as well as finding volunteers. Adding on more frequent testing, while great from a statistical viewpoint, just becomes that much more difficult in terms of funding and recruiting. I know we all complain about "big-pharma", and they should be able to fund any costs. However, most of the research that effects our area of science directly, is being conducted by small biotech companies and universities. Most don't have the required $ to conduct studies with the optimum trial design.