Replies to post #36546 on Biotech Values

[DewDiligence]

> EMIS – On the surface, the results seem promising, in that the high dose insulin group did show a signif decrease in HbA1c.<

It did not—the metric with the 0.037 p-value was a non-standard binary-responder metric that EMIS pulled out of thin air. The mean reduction in HbA1c had a p-value >0.05 (probably a lot greater) for all three insulin doses.

You should not blame yourself for being confused, however. On first inspection, I read the PR the same way you did.

It’s fairly obvious that the PR was deliberately confusing and was released a few minutes before the market open to cause even greater bewilderment.

[dewophile]

i couldn't agree more. to project on hypoglycemia using hgb A1c shows a level of ignorance that any medical student could remedy. monthly monitoring was also a silly design. More frequent monitoring would screen for subclinical hypoglycemia, and in addition there should be careful reporting of clinical signs of hypoglycemia (which should then reflexively require a blood glucose). based on more intense (perhaps even daily) monitoring then one could conclude hypoglycemia not a risk and blood monitoring of little or no value

I too bleieve there is a potential gem here, but the data is incomplete and from what I saw today speaks for a level of competence in management that makes me wary of an investment here

[DewDiligence]

> the claim that oral insulin does not cause hypoglycemia is preposterous based on the limited info presented.<

According to the CC, there were ~10 "hypoglycemia events” about equally distributed among the four trial arms.