News 
Market Data 
Discover
Replies to post #219946 on Biotech Values
07/05/18 9:23 PM
A complicated new clinical trial could give some hope to Alzheimer's patients and investors in drug companies. But there are lots of devilish details.
…The design of the clinical trial in question in complicated. It originally was designed to compare five different doses of BAN2401 compared to placebo. The trial utilized Bayesian statistics, a type of statistical methodology that allows researchers to make changes to a study in the middle. It also utilized what's called an adaptive trial design, meaning that if a dose appeared ineffective, patients would be switched to another dose that was more likely to work.
…[As reported in 2017: #msg-137105873]: …after 856 patients had been studied for 12 months, there was no finding of success. This was the main goal of the study. "The hurdle was extremely high because we did not want to identify a compound that may not follow through at its promise," says Lynn Kramer, the chief medical officer at Eisai.
But he says it was always the plan to do a more conventional statistical analysis at 18 months, the time normally required by regulators. At that analysis, he says, patients declined less on the very highest dose than on placebo – and the result was statistically significant not only at 18 months, but at six months and 12 months. In other words, while the trial had failed by Bayesian statistics, it succeeded when a more traditional statistical analysis was used.
AI
