New research from Eli Lilly ( LLY) demonstrates the difficulty that doctors, patients and, yes, investors face in balancing the benefits and risks of a drug. The drug is Evista, an osteoporosis medication that Lilly is testing as a treatment for invasive breast cancer. In a report from the The New England Journal of Medicine, a Lilly-sponsored study shows that Evista reduced the risk of breast cancer but also increased the chance of fatal strokes and the development of blood clots in the legs and lungs. "What level of breast-cancer risk would justify the use of Evista for the prevention of breast cancer for a given person, if one takes into account the competing risks and patient preferences?" asks Marcia L. Stefanick of the Stanford School of Medicine's prevention research center. "Complicating the answer is our inability to predict these risks with high accuracy on an individual basis." Stefanick's comments appeared Thursday in an editorial accompanying the study. Lilly hasn't commented on when it might seek Food and Drug Administration approval of Evista as a breast-cancer treatment, although the company has said it would file an application . Evista was approved by the FDA in late 1997 for osteoporosis. Evista is Lilly's fourth-largest drug, although its sales grew only 2% to $1.04 billion last year. Getting the FDA's approval for breast cancer would theoretically broaden the drug's sales prospects, but Albert Rauch, an analyst with A.G. Edwards, isn't so sure Lilly would make big gains. Even if the FDA grants approval, several other drugs on the market "have shown better efficacy, and we question Evista's ability to significantly" capture market share, he wrote in a research note. The other drugs include Arimidex from AstraZeneca ( AZN), Femara from Novartis ( NVS) and Aromasin from Pfizer ( PFE). Rauch, who has a buy rating on Lilly, doesn't own shares. His firm doesn't have an investment-banking relationship.