When data are clearly good or bad, go/no go decisions on product development are easy. But when data are equivocal - not quite good enough on primary end points, but very interesting on secondary end points or for subgroups - it's awfully hard not to keep on going.

Amylin Pharmaceuticals Inc. has found itself in that position as it continues to analyze data from its completed Phase III trials of pramlintide synthetic amylin in Types I and II diabetes (two other trials are ongoing). And while