It seems investors need a dose of reality. If you own shares of Eli Lilly (LLY -1.00%), please repeat after me: "The trial failed. Everything else is just speculation."

Shares were up 5% on Monday after data on Lilly's Alzheimer's disease drug solanezumab was presented at a medical conference. Since the company originally disclosed that the trials failed, the stock is up 20%. Except for the general love affair with pharma lately, there isn't much to explain the increased valuation.

If the trials failed their primary endpoints, what's got everyone so excited? A subgroup analysis of data for just the patients with mild Alzheimer's disease showed a slowing of cognitive decline. Individually, only one of the two trials showed a statistically significant reduction in decline of cognitive function. If you pool the data, there was also a slowing of cognative decline.

There is no way the Food and Drug Administration will approve the drug based on these data. The agency has a very clear stance that primary endpoints must be met; if they aren't, any analysis after that is just exploratory. Alzheimer's disease is an unmet need, but I don't think that really matters. Idiopathic pulmonary fibrosis is, too, but the FDA wouldn't approve InterMune's (ITMN.DL) drug, and it even passed one of its two phase 3 trials. Trying to get the drug approved with secondary endpoints is just plain ridiculous.

Eli Lilly is in a tough spot. Running two new large trials on patients with mild Alzheimer's disease will take a lot of time and money to prove a weak hypothesis. And if it decides to shelve the project, I'm sure there will be an outcry from patient advocates that the company is putting profits first.

The pharma's best plan -- and maybe this has been the plan all along -- is to get those patient groups to pick up some of the costs. It's worked for other smaller diseases. Usually with a large indication, a pharma would want to retain all the upside, but given the increased risk, it could make sense.

I'd feel more comfortable with the data if another of the amyloid-attacking drugs -- such as Pfizer (PFE -3.85%), Johnson & Johnson (JNJ -1.15%), and Elan's (ELN) bapineuzumab -- had also shown an effect in early Alzheimer's patients. Even if the effect is real, it might be very hard to prove given the timeframe and variability of measuring cognitive function.

Roche might have the best shot at proving the hypothesis. The company's Genentech group is studying its amyloid beta antibody, crenezumab, in a group of patients that have a mutation that typically triggers Alzheimer's symptoms around age 45, which might make it easier to measure a clinically meaningful outcome.