Expensive new antipsychotic drugs that are among the most widely prescribed pills in medicine are no more effective and no safer than an older, cheaper drug that has been largely discontinued, according to the most comprehensive comparative study ever conducted.

The surprising result of a federally funded study released yesterday challenges widespread assumptions among psychiatrists about the best way to treat serious mental illness and underscores the extent to which physicians, patients and policymakers can be blindsided by self-interested research by drugmakers.

The study also paints a sobering picture of the state of treatment of schizophrenia, a disabling illness that afflicts about 3.2 million Americans with symptoms such as delusions, hallucinations and disordered thinking: Every drug, old and new, caused serious side effects, and the vast majority of patients stopped taking each of them.

"The study has vital public health implications," said Thomas Insel, director of the National Institute of Mental Health, which funded the study. "It is the largest, longest and most comprehensive, independent trial ever done to examine existing therapies for this disease."

Heavily marketed on the grounds they caused fewer side effects, the newer drugs, known as atypical antipsychotics, cornered about 90 percent of the market.

All won Food and Drug Administration approval on the basis of short-term studies that showed they were better than sugar pills, and researchers emphasized yesterday that the medicines do work. But they have never before been systematically compared against each other in a long-term trial designed to guide doctors in deciding which to try first, and which would best suit particular patients.

The new study tracked patients for 18 months, which allowed researchers to compare the effectiveness of the medications as they actually are used. Patients had a variety of complicating factors, much as they do in the real world, and came from diverse backgrounds.

Columbia University psychiatrist Jeffrey Lieberman, who led the new study, said 90 percent of trials in the scientific literature -- which doctors rely on to guide treatment -- are sponsored by drug companies. Although the industry trials serve a useful purpose, he said, they are rarely designed to answer certain crucial questions.

The industry has recently come under fire for hiding unfavorable trial data, especially in studies of antidepressant medications for children. Companies conduct trials to win FDA approval or for marketing reasons, Lieberman said.

"These are not the same purposes that are consistent with the needs of clinicians, patients, family members, administrators and policymakers," he said. "How do currently marketed treatments compare? If there are differences in cost, are they justified? There has not been any consistent means to get that information."

Although the National Institutes of Health is increasingly funding such trials, Lieberman said they also ought to be funded by federal agencies that pay for drug treatment for huge numbers of patients, such as the Centers for Medicare and Medicaid Services (CMS).

"CMS pays tens of billions to support medication, and there is no way they know what the appropriate value of these medications is," he said.

The new study, which will be published in the New England Journal of Medicine, cost taxpayers $44 million. Last year, the United States spent $10 billion on the newer antipsychotic drugs, which include medications such as Zyprexa, Risperdal, Seroquel and Geodon.

Lieberman and the other researchers said they were surprised to find that an older generic drug called perphenazine, which is 10 times cheaper than the newer drugs, was about as effective -- and about as safe.

Older antipsychotics are known to cause involuntary muscle movements, and the newer drugs were heralded for not causing that problem. But Lieberman said comparisons with older drugs had mostly used a highly potent drug called Haldol, whereas the new study did not find the same degree of movement problems with perphenazine, a less potent drug.

In an editorial accompanying the study, Robert Freedman, a psychiatrist at the University of Colorado Health Sciences Center, said the uncontrolled movements caused by the older drugs were "less troubling than potentially fatal metabolic problems" associated with some of the newer drugs.

Yale psychiatrist Robert Rosenheck, who helped conduct the study, said it was "not a horse race" that produced a winner. Rather, he said, each drug had benefits and risks. Doctors will have to judge what works best for particular patients.

"It would be a tremendous mistake to assume from this study that the cheaper, older drugs are, quote, 'just as good,' " added Darrel Regier, director of the division of research for the American Psychiatric Association, who also said that doctors' judgment is crucial.

The study is likely to stoke one of the most contentious debates in psychiatry -- whether drug treatment ought to be forced on unwilling patients. The fact that three-quarters of patients discontinued treatment because of side effects or a lack of benefit showed that patients "trying to say no to forced neuroleptics [drugs] have had a better grip on reality than the medical community," said David Oaks, a patient advocate who has himself been given five antipsychotic drugs at various times for a range of diagnoses, including schizophrenia.

Pharmaceutical companies called attention to aspects of the trial that showed their products to advantage. Geodon kept patients' weight gain and cholesterol down, said Pfizer's Daniel J. Watts. Zyprexa had a lower discontinuation rate, manufacturer Eli Lilly said. The dose of Risperdal used in the study was too small, which was why the drug did not work as well as it should have, said Ramy A. Mahmoud, vice president for medical affairs at Janssen Pharmaceutica Inc.

Alan Goldhammer of the Pharmaceutical Research and Manufacturers of America, said, "We have always made it clear during drug development that it is only the first stage, that it . . . never tells the whole story about safety and efficacy."

Large studies, which compare different drugs including generics, are beyond the capabilities of individual companies, he said: "They are so costly and time-consuming that it would probably bring drug development to a halt."

Patient advocate Vera Hassner Sharav, who said her son died at 32 because of toxic side effects associated with the antipsychotic drug Clozaril, said the new study showed that the FDA is not doing its job. "We really only seriously test drug safety on the vast public," she said. "That is not right."