Data that could save money and help doctors make smarter treatment decisions are often unavailable at the time new medicines hit the market, according to U.S. researchers.

In a study out Tuesday, they found nearly a third of new drug approvals from the Food and Drug Administration included no data on how well the medications compare with existing alternatives.

"Even when these things are accessible, it's hugely time-consuming to go through it," said Joshua Gagne, a pharmacist at Harvard Medical School in Boston.

That leaves insurers and healthcare providers at a loss when trying to find the best and cheapest drugs for their patients, Gagne, who led the new work, told Reuters Health.

In 2009, Congress earmarked $1.1 billion to support drug comparison research, which would help doctors choose between various drugs.

But that information is unlikely to appear until years after new medicines reach the market, because drugmakers only need to show their products work better than a sugar pill — not how they compare to existing treatments.

"There is a gap between the time the drug hits the market and the time this information is generated," Gagne said.

With his colleagues, he went through public FDA data for 197 drugs approved between 2000 and 2010.

After excluding medications for diseases with no alternative treatments, 70 percent of the FDA approval packages contained some data on how the new products compared to existing ones.

Experts who spoke with Reuters Health were split in their reaction to the new study, published in the Journal of the American Medical Association.

Dr. Robert Temple, who directs FDA's Office of Medical Policy at the Center for Drug Evaluation and Research, said he was surprised by how many drugmakers had provided some comparative data, despite not being required to do so.

"Considering that there is absolutely no requirement, it is pretty impressive," he said in a phone interview.

Another expert took the opposite view.

"A third of new drugs that are approved have no comparative data at all," said Dr. Alec B. O'Connor of the University of Rochester School of Medicine and Dentistry in New York. "I think we'd all say that is sad."

He said he believes new drugs should always be tested against those already approved to treat a given disease.

In an earlier study, O'Connor had tried to compare the cost-effectiveness of different painkillers. But that turned out to be next to impossible, because almost no researchers had tested the drugs head-to-head.

"When a new drug becomes available, if it's not compared with a drug we already use, we don't know if it is better or worse -- we're just guessing. If you're a patient, I don't think you would want your doctor to be guessing," he told Reuters Health.

"You could save a lot of money if you had better comparative data at the time of approval," he added. "It's a win-win for patients, for doctors and for societal costs."

Meanwhile, Gagne and his team say the comparative FDA data, when it exists, should be made more accessible.

But according to an e-mail from Temple, of the FDA, that is not without problems.

"Comparative data are often too limited to allow a firm conclusion about comparative effectiveness (the size of studies that would be needed to rule out a small difference is usually daunting)," he told Reuters Health. "FDA would not want to suggest that a small study showing no difference between two treatments 'proves' equivalence, when the data fall well short of such proof."