New anti-psychotic medicines such as Seraquel, Risperdal and Zyprexa are some of the biggest sellers in the drug industry. But Chicago researchers say most of the time, those drugs are prescribed for uses not approved by regulators, and not supported by science.

This new generation of antipsychotics, called atypicals, were thought to be a big improvement over older drugs. But research is increasingly showing that those expectations haven’t panned out.

“There was promise that atypical agents were substantially safer and possibly more effective,” said Caleb Alexander, assistant professor of medicine at the University of Chicago. “And time and time again, we’ve seen we’re wrong in that case.”

Still, the drugs are widely prescribed for so-called off-label use, like dementia and anxiety, for which there’s scant evidence they’re effective. Alexander found unsupported use of the drugs has more than doubled since the mid-1990s. In 2008, some 54 percent of treatment visits involving atypicals were for off-label use for which the supporting science is sketchy at best. That adds up to about $6 billion in annual costs to patients, insurers and the government.

Alexander’s paper is out in the Journal Pharmacoepidemiology and Drug Safety.