As pharmaceutical makers find it increasingly difficult to bring new drugs to market, they are turning to genetic tools to seek uses for medicines that failed to make it out of the development pipeline.

The discovery of new links between genes and diseases can help not only to design new treatments, but to salvage drugs that are shelved when they come up short in clinical trials.

The idea is "to take some of these compounds, capitalize on past investments sitting idle, and now selectively accelerate them in the development process," says Terry Hisey, a pharmaceutical-industry strategist at the consultancy Deloitte LLP.

The ultimate goal is to develop medicines tailored to patients with a particular genetic makeup or who have diseases with a particular gene variation that are more responsive to a specific treatment. While few have been fully realized so far, there are a small but growing number, such as cancer drug Gleevec from Novartis AG and NitroMed's BiDil, a heart-disease medicine.

"There's a big buzz about it," says Edward Abrahams, executive director of the Personalized Medicine Coalition, a nonprofit education and advocacy group comprising industry, academia and government organizations.

While the markets for these therapies are smaller than for those that treat the general population, pharmaceutical companies are realizing there are hefty profits to be made, because patients are more likely to use — and stick with — a tailored medicine that works better than a one-size-fits-all drug. For instance, biotechnology giant Genentech Inc.'s Herceptin, which is taken by women with breast cancer who have a particular variant of the HER-2 gene, reaped $1.3 billion in sales last year.

The broader shift toward so-called personalized medicine dovetails with pharmaceutical makers' recognition of how hard it is to maintain a business model that relies on producing a few drugs that bring in billions of dollars a year.

"The old sort of pharmaceutical model is under pressure right now," says Wayne Rosenkrans, chairman and president of the Personalized Medicine Coalition. "Personalized health care is one of those potential solutions."

Drug development typically can take 10 years or more. Failures cost companies both time and money, from $15 million per compound at the first stage of human testing to $86 million in late-stage trials, according to a 2003 study on the cost of drug innovation published in the Journal of Health Economics.

Historically, these cost have increased over time above the rate of inflation, but there aren't more recent data available for pharmaceuticals, says Joseph DiMasi, a study author who is director of economic analysis at the Tufts Center for the Study of Drug Development.