STUDIES MAY BE OVERSTATED, NEW JAMA REPORT SAYS

Nearly one out of every three published medical studies may be exaggerated
or simply untrue, a new study says.

The study's results, which were published in the Journal of the American
Medical Association, concluded that 16 percent of the studies examined were
later contradicted and another 16 percent were weakened by further testing.

These numbers led the authors to conclude that almost one-third of
published medical research findings may not hold up.

Donald Heistad, a UI professor of internal medicine, was glad to see the
study because it brings to light important facts about medical research the
public and media don't realize, he said.

"There is an inherent probability that some findings will be wrong," he
said. "When you have to rely on statistics [for research], there's a
built-in chance for error."

He added that the media are to blame for some of the misinterpretation.

"The media latch onto surprising findings; if a study happens to find that
cholesterol isn't harmful, they'll focus on it," he said. "I think it's an
error by the press to overstate such findings and also an error by the
public to accept them."

Robert Wallace, a UI professor of epidemiology, agreed, but he attributed a
large part of the exaggeration to lack of technical understanding by writers.

"Most of the media are pretty responsible, and they don't want to go beyond
what the results say - they just don't have the technical expertise [to
interpret the results]," he said.

He has been on both sides of the issue. He and his colleagues found results
during an observational study 25 years ago indicating that estrogen
protected the heart from coronary disease. Later, randomized trials
contradicted their conclusion, and further research was halted.

"There's no great sin to have scientific findings that turn out to be
wrong," he said. "[Scientists] put their findings out to be refuted."

The study also emphasized that journals tend to include results that are
new and exciting. Doing this may place additional expectations on researchers.

"Scientists know it's easier to publish a positive finding where a therapy
has been proven effective rather than a negative one where a therapy's
effectiveness is questioned," Heistad said. "This doesn't mean they will
exaggerate, it just means they'll have an easier time getting [a positive
finding] published."

Wallace also acknowledged the problem, adding that major medical journals
are taking steps to reduce the problem.

"It's very difficult for the public to find truth [and] there is only one
solution; talk to an expert," Heistad said. "People should talk to their
doctors; they're expected to be knowledgeable about these recent findings."