More than 70 people have crowded into reception rooms furnished with rows of attached blue plastic seats. Cabanerio is one of many regulars who gather at SFBC International Inc.'s test center, which, with 675 beds, is the largest for-profit drug trial site in North America.

Across the U.S., 3.7 million people have enrolled in drug tests sponsored by the world's largest pharmaceutical companies. The companies have outsourced 75 percent of experimental drug trials to centers like SFBC, a leader in a $14 billion industry.

At the same time, the U.S. Food and Drug Administration has farmed out much of the responsibility for overseeing safety in these tests to private companies known as institutional review boards. These boards are also financed by pharmaceutical companies.

So, the drug industry is paying the people who do the tests -- and most of the people who regulate those tests. And that combination can be dangerous, and sometimes deadly.

``The fundamental problem is a system in which investor- owned businesses have control over the evaluation of their own products,'' says Marcia Angell, editor in chief of the New England Journal of Medicine from 1999 to 2000. ``Oversight of clinical trials is too important to leave in the hands of drug companies and their agents.''

`I'm in a Bind'

Most of the people lining up at SFBC to rent their bodies to medical researchers are poor immigrants from Latin America, drawn to this five-story test center in a converted Holiday Inn motel.

Inside, the brown paint and linoleum are gouged and scuffed. A bathroom with chipped white tiles reeks of urine; its floor is covered with muddy footprints and used paper towels. The volunteers, who are supposed to be healthy, wait for the chance to get paid for ingesting chemicals that may make them sick.

They are testing the compounds Big Pharma, the name for the world's largest pharmaceutical companies, hopes to develop into best-selling medicines.

Cabanerio, who has a mechanical drafting degree from a technical school, says he left Venezuela because he lost his job as a union administrator. For him, the visit to SFBC is a last resort. ``I'm in a bind,'' Cabanerio says in Spanish. ``I need the money.''

Conflicts of Interest

Few doctors dispute that testing drugs on people is necessary. No amount of experimentation on laboratory rats will reliably show how a chemical will affect people. Helped by human testing, drugmakers have developed antibiotics capable of curing life-threatening infections as well as revolutionary treatments for diseases like cancer and AIDS.

These medical success stories mask a clinical drug trial industry that is poorly regulated and riddled with conflicts of interest. Every year, trial participants are injured or killed.

Rules requiring subjects to avoid alcohol and narcotics and to take part in only one study at a time are sometimes ignored by participants, putting them at risk and tainting the test data.

The consent forms that people in tests sign -- some of which say participants may die during the trial -- are written in complicated and obscure language. Many drug test participants interviewed say they barely read them.

Ken Goodman, director of the Bioethics Program at the University of Miami, says pharmaceutical companies are shirking their responsibility to safely develop medicines by using poor, desperate people to test experimental drugs.

`It's an Eye-Opener'

``The setting is jarring,'' says Goodman, 50, who has a doctorate in philosophy, after spending 90 minutes in the waiting rooms at SFBC's Miami center, which is also the company's headquarters. ``It's an eye-opener. Every one of these people should probably raise a red flag. If these human subject recruitment mills are the norm around the country, then our system is in deep trouble.''

Pharmaceutical companies distance themselves from the experiments on humans by outsourcing most of their trials to private test centers across the U.S. and around the world, says Daniel Federman, a doctor who is a senior dean of Harvard Medical School in Boston.

The chief executive officers of drug companies should be held accountable for any lack of ethics in these tests, he says.

``The CEOs of the companies have to be publicly, explicitly and financially responsible for the ethical approach,'' says Federman, 77, who still sees patients. ``It's not possible to insist on ethical standards unless the company providing the money does so.''

Pressure for New Drugs

CEOs of 15 pharmaceutical companies that outsource drug testing to firms including SFBC -- among them, Pfizer Inc., the world's largest drugmaker; Merck & Co.; and Johnson & Johnson -- declined to comment for this story.

SFBC Chief Executive Arnold Hantman says his center diligently meets all regulations. ``We take very seriously our responsibilities to regulatory authorities, trial participants, clients, employees and shareholders,'' Hantman, 56, says. ``We are committed to conducting research that fully complies with industry and regulatory standards.''

The pressure pharmaceutical companies face to develop new drugs has intensified in the past 15 years.

Faced with the expiration of patents on best-selling drugs like AstraZeneca Plc's Prilosec, which has helped tens of millions of people with heartburn and ulcers, Big Pharma has been in a frenzied race to find new sources of profit.

Oversight Secrecy

When the patent for a company's blockbuster drug expires, a lucrative monopoly vanishes. Such drugs typically lose 85 percent of their market share within a year of patent expiration, according to CenterWatch, a Boston-based compiler of clinical trial data.

The private independent review board companies that oversee drug trials operate in such secrecy that the names of their members often aren't disclosed to the public.

The oldest and largest review company is Western IRB, founded in 1977 by Angela Bowen, an endocrinologist. WIRB, an Olympia, Washington-based for-profit company, is responsible for protecting people in 17,000 clinical trials in the U.S.

The company oversaw tests in California and Georgia in the 1990s for which doctors were criminally charged and jailed for lying to the FDA and endangering the lives of trial participants. No action was taken against WIRB. Bowen says she didn't see human safety issues in those trials.

WIRB aims to visit test sites it monitors once every three years, Bowen says.

Unlicensed Employees

The FDA's own enforcement records portray a system of regulation so porous that it has allowed rogue clinicians -- some of whom have phony credentials -- to continue conducting human drug tests for years, sometimes for decades.

The Fabre Research Clinic in Houston, for example, conducted experimental drug tests for two decades even as FDA inspectors documented that the facility had used unlicensed employees and endangered people repeatedly since 1980. In 2002, the FDA linked the clinic's wrongdoing to the death of a test participant.

Review boards can have blatant conflicts of interest. The one policing the Fabre test center was founded by Louis Fabre, the same doctor who ran the clinic. Miami-based Southern IRB has overseen testing at SFBC and is owned by Alison Shamblen, 48, wife of E. Cooper Shamblen, 67, SFBC's vice president of clinical operations. Both Shamblens declined to comment.

`Gives Me Hives'

SFBC's 2005 shareholder proxy, filed with the U.S. Securities and Exchange Commission, lists Lisa Krinsky as its chairman and a director of medical trials and refers to her 26 times as a doctor. Krinsky, 42, has a degree from Sparta Medical College in St. Lucia in the Caribbean; she is not licensed to practice medicine.

Arthur Caplan, director of the Center for Bioethics at the University of Pennsylvania in Philadelphia, says handing oversight of human drug experiments to private, for-profit companies is a mistake.

``This whole world gives me hives, this privatized review process,'' Caplan, 55, says. ``I've never seen an IRB advertise by saying, `Hire us. We're the most zealous enforcer of regulations you could have.' People say, `We'll turn it around faster. We're efficient. We know how to get you to your deadlines.'''

The Pharmaceutical Research and Manufacturers of America, a Washington-based trade association and lobbying group, says human drug tests in the U.S. are safe and well-monitored.

`Gold Standard'

``The vast majority of clinical trials conducted in the United States meet high ethical standards,'' PhRMA, as the group is known, said in a written response to questions. ``The U.S. regulatory system is the world's gold standard, and the Food and Drug Administration has the best product safety record.''

Joanne Rhoads, the physician who directs the FDA's Division of Scientific Investigations, says that view isn't realistic. ``What the FDA regulations require is not any gold standard for trials,'' Rhoads, 55, says.

The agency doesn't have enough staff to aggressively monitor trials, she says, adding that FDA regulations are a bare minimum and much more oversight is needed. ``You cannot rely on the inspection process to get quality into the system,'' Rhoads says. ``I know many people find this not OK, but that's just the truth.''

Michael Hensley, a pediatrician who was an FDA investigator from 1977 to 1982, says the agency has become less active in clinical trial oversight in recent years. Families of injured or dead trial participants seeking accountability for mistakes have to file lawsuits.

`Stopped Enforcing Rules'

``The FDA's backbone has been Jell-O,'' says Hensley, 60, who's now president of Chapel Hill, North Carolina-based Hensley & Pilc Inc., which advises pharmaceutical companies on FDA compliance. ``The folks at the FDA stopped enforcing the rules several years ago.''

By law, drug companies must first conduct tests to determine whether potential drugs produce dangerous side effects, such as organ damage, impaired vision or difficulty breathing. The FDA calls them phase I tests.

In 1991, 80 percent of industry-sponsored drug trials were conducted by medical faculty at universities, with protection for participants provided by the schools' own oversight boards, according to the New England Journal of Medicine.

Now, more than 75 percent of all clinical trials paid for by pharmaceutical companies are done in private test centers or doctors' offices, according to CenterWatch.

`No Qualifications'

Some test centers, FDA records show, have used poorly trained and unlicensed clinicians to give participants experimental drugs. The centers -- there are about 15,000 in the U.S. -- sometimes have incomplete or illegible records. In California and Texas, clinicians have used themselves, staff or family members as drug trial participants.

``Unfortunately, I don't think it's been recognized how important it is that people who actually conduct the trial be trained,'' Rhoads says. ``We oftentimes see people with no qualifications whatsoever, but they'll go to a one-day training course and they call themselves a certified study coordinator.''

These people often run 90 percent of the study with little involvement by physicians, she says.

Participants in Miami clinical trials talk openly about how they violate SFBC rules intended to protect the integrity of research findings. SFBC prohibits people from taking part in two clinical trials at the same time.

Multiple Trials

Roberto Alvarez, 36, an Argentine in the U.S. on a visa; Efrain Sosa, 35, a Cuban native; and Marlon Matos, a 27-year-old immigrant from Venezuela, say they've participated in more than one clinical trial in Miami at the same time or gone from one test to another, ignoring required waiting periods.

They say they do it for the money, without telling the test centers, and that no one has ever caught them violating the rule.

``We maintain many safeguards to help us ensure that the participants of our clinical trials are not participating simultaneously in multiple clinical trials,'' SFBC's Hantman says.

SFBC fingerprints participants to keep track of their tests at the company, he says. ``Unfortunately, there is no clearing house that we're aware of that would allow us to find if they were participating in another trial at the same time,'' he says.

In April, Alvarez signed up for a 36-day clinical trial at Miami testing company Elite Research Institute for a new sustained-release form of donepezil, an Alzheimer's drug that Tokyo-based Eisai Co. sells in the U.S. with New York-based Pfizer.

`I Hop Around'

At the time, Alvarez was in the middle of a 212-day test sponsored by Madison, New Jersey-based Wyeth at SFBC for an experimental muscular dystrophy drug, according to consent forms he signed.

``I hop around to get around that,'' says Alvarez, a part- time construction worker who's wearing a black T-shirt and jeans when he's interviewed in a bagel shop two doors down from SFBC. ``They ask, but I just don't tell them. Everybody does that.''

Steve Simon, a research biostatistician at Children's Mercy Hospital in Kansas City, Missouri, says that when people participate in more than one clinical trial at a time, it can be harmful to people and research.

``When neither researcher knows about the potential interactions with the other trial, that raises concerns about scientific validity,'' says Simon, who has a Ph.D. in statistical research. ``You don't know how these things might interact. It's asking for trouble.''

Carefully Planned

Ernesto Fuentes, Elite's clinical trial director, didn't return calls for comment. Eisai spokeswoman Judee Shuler says Elite did everything it could to ensure participants in the clinical trial weren't in other tests at the same time, including asking subjects verbally if they were.

Pfizer spokesman Stephen Lederer says his company had no role in the donepezil tests. Gerald Burr, a Wyeth spokesman, says the company carefully planned and monitored the clinical trial.

The FDA requires pharmaceutical companies to hire monitors to audit clinical trials to ensure patient safety and scientific validity. ``Our sponsors visit our facilities frequently to monitor our trials and also routinely audit our work,'' SFBC's Hantman says.

Pharmaceutical company monitors spend more time scrutinizing data being gathered than watching out for people's safety, Harvard's Federman says. ``There are no monitors of monitors,'' he says. ``It's like looking at a dark cloud. There's minimum training. They're relying on people running the trials.''

Vioxx Verdict

The shortcomings of human drug testing may come to light in the welter of litigation surrounding Vioxx, the blockbuster pain reliever that Whitehouse Station, New Jersey-based Merck pulled off the market last year after its own studies found long-term use posed twice the normal risk of a heart attack.

A 2004 study by David Graham, the FDA's associate director for science and medicine, estimated that Vioxx caused as many as 140,000 heart attacks and strokes, killing as many as 55,000 people.

On Aug. 19, a Texas jury ordered Merck to pay $253 million to the widow of a Vioxx user, an amount that will be reduced to $26 million under state law. The company has been sued by more than 5,000 people who say they were hurt by the drug.

Before Vioxx was approved by the FDA, Merck tested it on thousands of people in early phase I clinical trials across the U.S., including at SFBC's Miami center.

Pharmaceutical companies sponsored 36,839 new clinical trials from 2001 to 2004, six times more than in the period from 1981 to 1985.

$50 Billion at Stake

The search for the next money-spinning drug is fueling the surge in human testing. Pharmaceutical companies that make 28 top-selling drugs will lose a total of $50 billion in revenue as their patents expire from 2003 to 2008, according to Norwalk, Connecticut-based market research firm BCC Inc.

Schering-Plough Corp., for example, suffered a drop in revenue after losing U.S. exclusivity for Claritin, an allergy treatment, in December 2002.

The Kenilworth, New Jersey-based company's sales fell 18 percent to $8.3 billion in 2003 from $10.2 billion the year before, and the company reported a net loss of $92 million in 2003 compared with a profit of $1.97 billion in 2002. Schering- Plough shares averaged $17.42 in 2003, down from an average price of $25.99 in 2002.

Schering-Plough has used SFBC for clinical tests, including trials in the past year comparing different forms of Claritin. ``We believe that they are at the industry standard, and the appropriate checks and balances are in place,'' Schering spokeswoman Rosemarie Yancosek says.

`Ethical Violations'

As drug companies try to get new drugs to market, time is literally money. They lose as much as $5 million a day waiting to get approval of new medications. Eighty percent of all experimental drugs tested in humans are never approved by the FDA.

Big Pharma has an insatiable demand for people to be in clinical trials, says Angell, a doctor and former editor of the New England Journal of Medicine.

``Human subjects are in very short supply, so it's not surprising that under the growing pressure to find them, there are sometimes terrible ethical violations,'' says Angell, 66, a Harvard Medical School senior lecturer. ``Drug companies may claim innocence, but they need to take responsibility.''

In 1978, the National Commission for the Protection of Human Research Subjects, an advisory committee appointed by President Richard Nixon, recommended, in what became known as the Belmont Report, that clinical trial participants be fully informed of risks and sign a consent form.

So-called informed consent wasn't required by the FDA until 1981.

`Genuine Social Problem'

Interviews with people in clinical trials and relatives of participants who died in medical experiments across the U.S. suggest that researchers often don't fully explain risks and potential side effects.

Bowen, whose Western IRB has overseen trials at SFBC sites, says phase I centers often don't conduct the informed consent process properly. ``I'd say it's fairly widespread,'' she says. ``It's a genuine social problem that needs to be dealt with.''

Alvarez, the clinical trial participant from Argentina, says he skimmed over the 12-page consent form for a test SFBC managed for KW-6002, an experimental Parkinson's disease drug made by Tokyo-based Kyowa Hakko Kogyo Co., before signing the form on Aug. 30.

``The thing I pay most attention to when filling this thing out is this,'' says Alvarez, flipping through the form, written in Spanish, to a page that describes payment terms. ``How much it pays and how long it takes. I don't read them too carefully.''

8-Night Confinements

Page 8 of the consent form explains that the 57-day test Alvarez has signed up for pays $4,300, spread out in payments tied to completion of three 8-night ``confinements.'' During confinements, participants aren't allowed to leave the SFBC building unless they decide to drop out of the trial.

They live in 12-foot (3.66-meter) by 24-foot rooms outfitted with three double-decker beds. The center has recreation rooms with televisions, pool tables and video games.

The payment schedule provides an incentive for participants to stay the course: About half of the money, or $2,355, isn't paid until the last week of the eight-week test.

The trial, which was scheduled to end today, also includes 12 outpatient visits to test for levels of KW-6002 in people's bloodstreams. The consent form says KW-6002 can produce side effects that include heart palpitations, sleep disorders and breathing difficulties.

`A Whole Team'

An SFBC employee asked if Alvarez had read the consent form and understood what the test entailed when he signed up, Alvarez says. He told the clinician he had read the form, and the clinician didn't say anything more about risks, he says.

SFBC Executive Medical Director Kenneth Lasseter says the center always explains risks to participants. ``We have a whole team of people,'' he says. ``They go over the risks and discomforts and explain them to the subject.''

Lasseter says he's never before heard that participants said they weren't fully informed of risks in tests. ``Everyone who is screened has a one-on-one interview with one of the screening team that goes over the informed consent,'' he says. ``If they are denying that, that's simply a fabrication. They simply are not being truthful.''

Informed consent documents routinely fail to satisfactorily explain risks to potential participants, says Laura Dunn, a professor of psychiatry at the University of California, San Diego, who wrote an article on informed consent that appeared this year in the Journal of the American Medical Association.

`Poor Understanding'

``Decades of research show that poor understanding of informed consent documents is widespread,'' she says.

The title of the KW-6002 consent form says the test is a phase I clinical trial. The document doesn't explain what phase I means, that the purpose is to determine the side effects and safety of an experimental drug.

The test, the consent form says, aims to determine how the active ingredient in KW-6002, istradefylline, is ``absorbed, distributed, decomposed and eliminated from the body.'' Joseph Brindisi, a spokesman for Kyowa's U.S. unit, declined to comment.

It's inevitable that tests that often make healthy people sick rely on the poor, says Greg Koski, who from 2000 to 2002 was head of the federal Office for Human Research Protections. A division of the Department of Health and Human Services, the office oversees all federally funded clinical trials; it doesn't review pharmaceutical company-sponsored tests in private centers.

`Disproportionate Burden'

``I have little doubt that there is a disproportionate burden of risk that falls on the disadvantaged members of our society,'' says Koski, 55, who's now an anesthesiologist in Boston.

SFBC Executive Vice President Greg Holmes says money is the main reason people sign up for phase I tests. ``Look at the benefits,'' he says. ``There is little benefit other than getting paid. There's no secret there.''

SFBC conducted a test in June of a drug that may treat overactive bladders. The test was sponsored by Theravance Inc., a South San Francisco-based company that's 21 percent owned by GlaxoSmithKline Plc.

The London-based company, the largest drugmaker in Europe, has marketing rights for new Theravance drugs, according to filings with the SEC.

``The goal of this study is to determine the highest daily dose of TD-6301 that will not cause an undesired increase in heart rate,'' the consent form says.

University of Miami bioethicist Goodman says the wording is misleading and confusing. ``They're saying it backwards to a population that may not be of the highest education level,'' he says.

`Don't Want to Say That'

The only way to accomplish the intent of the study is to raise the dosage of the experimental drug until heart rates increase, Goodman says. ``The real purpose of the study is, `We're going to make you sick in order to find out at what level you get sick when given this drug,''' Goodman says. ``Obviously, they don't want to say that.''

SFBC's Lasseter says the wording in that consent could be better. ``It's clear to me,'' he says. ``Perhaps it needs to be explained more.''

GlaxoSmithKline spokesman Rick Koenig says his company wasn't involved in the clinical trials. He adds that GlaxoSmithKline has the right, but not the obligation, to develop the Theravance drugs. Theravance spokesman David Brinkley says his company policy is not to comment on specific clinical trials.

Cabanerio, the Venezuelan immigrant, says he reads consent forms and questions doctors and clerks at SFBC closely to weigh the risks against his need for cash for his family. In July, he says, he needed the money so badly he was willing to enroll in a test that could have had fatal results.

Potential Risk: Death

Cabanerio signed up for a trial that mixed alcohol with an experimental opiate pain reliever called Oros Hydromorphone, made by Alza Corp., a unit of New Brunswick, New Jersey-based Johnson & Johnson. The test paid $1,800.

Participants who chew Oros tablets, as opposed to swallowing them whole as directed, can overdose, which can cause a heart attack or death, a June 21 consent form in Spanish for the test says. People also can have allergic reactions to Oros, which, if severe, can be fatal, the form says.

``It's not the job I would choose, but financial circumstances require you to do it sometimes,'' Cabanerio says.

The doctors who examined Cabanerio during the screening process for tests at SFBC asked him to recite a couple of side effects listed on the test's consent form to see if he understood the risks, he says.

One Woman Fainted

While being screened for the Oros test, Cabanerio says, a doctor told him there were few risks involved. ``He said the strongest reaction would be like a shot of whiskey,'' Cabanerio says. ``He said it would be fun.''

The test included four 3-night stays in which some patients were given Oros and up to 40 percent alcohol mixed with orange juice on an empty stomach, according to the 14-page consent form.

After Cabanerio and 18 other people began the test on the fourth floor of SFBC's center, one woman fainted, Cabanerio says. Another woman in the test got so drunk after drinking the brew that she began imitating a striptease dancer.

Cabanerio says he didn't feel bad because he was in a different group of participants that received lower doses of alcohol and were allowed to eat beforehand.

Cabanerio participated in the test in July. That's the same month the FDA asked Purdue Pharma LP, a Stamford, Connecticut- based drug company, to withdraw another opiate tested with alcohol at SFBC's Miami center.

Backloading Payments

Purdue withdrew the drug, Palladone, because its time- release mechanism is dissolved by alcohol, which could cause a deadly release of all the opiates at once, according to the FDA. Participants were given naltrexone to block the opiates.

Alza ensures tests of its drugs are safe for participants by following FDA rules and guidelines approved by IRBs, company spokesman Ernie Knewitz says. ``Patient safety is the most important element in each clinical study conducted by Alza,'' Knewitz says.

The Purdue experiment paid volunteers $2.78 an hour, or $66.72 per 24-hour day, for the first nine days of confinement. For those who remained, payment jumped to $333.33 a day for the final three days, with a bonus of $800 paid following a single follow-up visit.

Such payment backloading is coercive and thus unethical, says Peter Lurie, a physician who is deputy medical director of Public Citizen, a Washington-based group that monitors patient safety issues.

`Powerful Incentive'

``It provides a very powerful incentive for somebody to continue in a study even if they're being made uncomfortable by it,'' he says.

Purdue's payment schedule complies with guidelines set by the FDA and international regulators, company spokesman James Heins says. He says any experiment dropouts willing to return for the follow-up visit were paid $800.

Heins says anyone who dropped out in the middle of a confinement period without a health reason was considered ``noncompliant'' and was paid $25 a day.

Under federal regulations, anyone can drop out of a clinical trial at any time. University of Pennsylvania bioethicist Caplan says it's often not easy to voluntarily leave a test.

He says he enrolled himself in a trial in which a clinician inserted a tube down his throat. Caplan says after the procedure started, he told a nurse: ``You know, I don't like this. I don't want to do it anymore.''

He says the nurse told him: ``You can't do that. You can't stop!'' He completed the procedure.

Like a Burning Shock

Wyeth sponsored trials at SFBC this year to find out what dosages of an experimental drug to treat muscular dystrophy caused side effects, according to the consent form for the trial.

Possible side effects included severe allergic reactions that can cause breathing difficulty, abdominal pain, increased heart rate and death, according to the consent form.

Healthy people were paid $5,500 for staying in the center for 15 nights during a 26-week test. Another version of the test with a 29-night stay in the center paid $6,900.

John Juarez, who was born in Miami, says the injections of Wyeth's experimental drug felt like a burning electric shock searing his body from within. ``It made me feel really weird,'' says Juarez, 22.

In the last few weeks of testing, Juarez developed red hives up and down one arm that wouldn't go away for days, he says. And he started growing hair all over his body, including thick sideburns that he still wears.

Handled Properly

Wyeth has documented that an IRB approved the trial, consent was handled properly and the test followed all FDA rules, Wyeth spokesman Burr says. ``Wyeth is committed to sponsoring and supporting carefully conducted clinical trials as the fastest and safest way to find treatments that work in people and ways to improve health,'' he says.

The FDA depends on IRBs to approve and review trials. For drug tests conducted at SFBC in Miami, London-based AstraZeneca, Merck and Purdue have used Southern IRB, the review board owned by Alison Shamblen, the wife of SFBC Vice President Cooper Shamblen.

Purdue, whose Palladone tests were monitored by Southern IRB, didn't know Southern was owned by a relative of an SFBC executive, Heins says.

``If Purdue had been aware of the relationship you allege, the company would have looked into the issue before conducting trials at the site,'' Heins says. ``Purdue will address this issue should we decide to work with SFBC in the future.''

Not Merck's Choice

Merck, which has relied on Southern IRB to monitor tests at SFBC, including an April experiment for a drug to prevent nausea and vomiting, says the company wasn't responsible for using an IRB owned by a relative of an SFBC executive.

Merck chose SFBC because for years it had worked with Clinical Pharmacology Associates, which SFBC bought in 2003, Merck spokeswoman Janet Skidmore says. ``SFBC selects which institutional review board is most appropriate,'' she says. ``Merck did not choose Southern IRB -- SFBC did.''

SFBC's Hantman says Alison Shamblen hasn't been affiliated with Southern IRB since early 2005. Rosa Fraga, Southern IRB's chairwoman, says Shamblen still owned the IRB as of Oct. 10. Fraga says Alison Shamblen decided in October to shut Southern IRB after 16 years. Fraga herself will soon open a new company called Southern IRB Services, she says.

The FDA has found ``significant objectionable conditions'' during three inspections of SFBC since 2000. In 2002, the FDA found SFBC conducted invasive procedures on people without getting proper consent from the participants. In March 2005, the FDA wrote up a significant objectionable conditions finding that it hasn't yet made public.

`Positive Feedback'

SFBC's Hantman declined to release the report. ``We have consistently received positive feedback from the FDA's reviews,'' he says. SFBC's Lasseter describes the FDA reports as being ``like a traffic ticket.''

SFBC Chairwoman Krinsky says the company hasn't received a warning letter, which is more serious than a significant objectionable conditions citation, from the FDA in more than 20 years. She says the company has addressed all observations by the agency.

The phrase ``institutional review board'' dates back to the time when most boards -- like the clinicians they monitored -- were part of universities or hospitals. Today, the review industry is dominated by a handful of large, for-profit companies with enormous power.

No Training Needed

IRBs have the duty to reject or stop a clinical trial if the risks are found to outweigh the benefits. Nobody knows for sure how often trials are stopped since there is no central database that tracks IRB actions.

The exact number of IRBs is also a mystery. There are an estimated 3,000 to 5,000 of them, according to the Government Accountability Office, the investigative arm of Congress. The number is unknown because the companies don't have to register with the FDA.

IRB members don't have to be trained or certified.

FDA oversight of IRBs is scarce -- and becoming scarcer. The agency conducted 175 inspections of IRBs in the year ended on Sept. 30, down from 327 in the year ended on Sept. 30, 2002, according to FDA records.

When the FDA conducts an inspection, it reviews informed consent documents and checks that an IRB has at least one person with a scientific background, one layperson and one community member.

The agency reviews the IRB's record-keeping to see whether it has maintained proper minutes of meetings. ``The regulations for IRBs are fairly loose,'' the FDA's Rhoads says.

Researcher Is Also Monitor

The inadequacy of the IRB system is illustrated by the case of Louis Fabre, the Houston psychiatrist who ran at least 400 clinical trials with 20,000 people for more than 50 drug companies at his Fabre Research Clinic from 1973 to 2005.

To monitor those trials, Fabre, 64, used an IRB that he had founded himself. He called it the Human Investigation Committee, and its members included his business partner, psychiatrist Stephen Kramer, 64; and his lawyer, Bruce Steffler, 60.

The Human Investigation Committee allowed Fabre Research Clinic to run tests even as FDA inspectors found conduct that put people at unnecessary risk during six inspections from 1980 to 2005.

In 1980, the FDA reported that a woman enrolled in one of Fabre's experimental psychiatric drug tests had killed herself during the study. The FDA wrote that the woman was supposed to be in an inpatient study, and Fabre managed the study instead as an outpatient trial. Fabre was never censured for that incident.

FDA's Letter

In January 2005, the FDA wrote a letter to Fabre detailing his wrongdoing in connection with the death of Garry Polsgrove, an unemployed and homeless Vietnam veteran, in his clinic in May of that year.

Polsgrove, 55, died during a trial for generic schizophrenia drug clozapine that was sponsored by Miami-based Ivax Corp., the largest U.S. maker of generic drugs.

Fabre left Polsgrove in the care of John Rodriguez, who had no medical credentials, according to the FDA. Just six days before Polsgrove enrolled in the experiment, an FDA inspector visited Fabre's clinic and found that Rodriguez had screened subjects, performed physicals and conducted electrocardiograms.

The inspector believed Fabre's false claim that Rodriguez was a licensed physician's assistant, Rhoads says. A call to the Texas Board of Medical Examiners would have revealed that Rodriguez was unlicensed.

Rhoads says FDA inspectors don't normally verify medical licenses. ``On a routine inspection, it's not likely that they're going to dig because it takes a lot of work to do that,'' Rhoads says.

Two Purple Hearts

The agency waited almost three years after Polsgrove's death before it moved to ban Fabre from running trials.

Polsgrove was an ex-Marine who had won two Purple Hearts for his service in Vietnam in 1967. Polsgrove's sister, Nancy Gatlin, who says her brother was healthy before starting the drug trial, says Fabre killed her brother.

``He should have been stopped a long time ago,'' she says. Fabre, who now runs a drug development company in Houston, declined to comment. He denied wrongdoing in a response to the FDA.

The FDA can investigate a trial site at any time. Rhoads says when inspectors review a test center, they follow a checklist.

``The bottom line is, the inspections by the FDA field investigators are done by people who are trained in investigation, but they don't always have a tremendous scientific or medical background,'' Rhoads says. ``They're basically doing an audit process.''

`You Get a Whitewash'

Unable to oversee human drug testing by itself, the FDA has left much of the job to IRBs. Bowen's Western IRB had $20 million in revenue in 2004. It has grown at about 20 percent a year for the past decade, she says.

Bowen, 73, who used to be president of a drug company called William P. Poythress Inc. in Richmond, Virginia, says Western is the IRB for more than half of all new drug submissions to the FDA.

Bowen says WIRB is the best in the industry because of the professionalism of her members, their training and expertise and their willingness to turn down drug company tests they don't approve of.

Harvard's Federman sees WIRB differently. ``If you listen to themselves talk about themselves, you get a whitewash,'' he says.

`He Fooled Everybody'

In the 1990s, WIRB oversaw 23 clinical trials conducted by Robert Fiddes, a Los Angeles doctor who was charged with lying to the FDA. The FDA's investigation found that Fiddes repeatedly fabricated data and improperly included employees and family members in trials.

He pleaded guilty in 1997 and was sentenced to 15 months in federal prison.

A 1999 FDA inspection report criticized WIRB for its role in the doctor's experiments. ``There is a failure to have complete documentation of the board's knowledge, discussion and decisions regarding research activities,'' FDA investigators wrote of WIRB.

Bowen says WIRB didn't know about Fiddes's fraud. ``He fooled everybody,'' she says.

Pharmaceutical companies would be amazed at how poorly some clinical tests are run, she says. ``Some of the companies would be embarrassed if they saw the quality of the people doing research,'' she says. ``I call them clueless.''

WIRB's headquarters has 44,000 square feet (4,088 square meters) of office space on an 18-acre (7.28-hectare) campus studded with towering Douglas fir trees. It has 250 employees, who refer to themselves as ``Wirbies.''

Four Minutes for Discussion

Review board members attend about 40 four-hour meetings each month to approve new experiments and trial recruiting materials, review ongoing tests and examine reports of serious side effects, Bowen says.

About 60 items are considered at each meeting, giving members an average of four minutes to discuss each issue. The meetings and their minutes are closed to the public, as are the names of the board's members.

``If you were a plaintiff's lawyer, wouldn't you like to have the identities of all the membership?'' Bowen asks.

The FDA most recently inspected WIRB in August 2002. The agency found that WIRB's computer system lacked an audit function, meaning data entered could be altered without a record of the changes. The FDA called that a ``significant objectionable condition.''

In a 1999 inspection, the FDA criticized WIRB's role in the case of Richard Borison, a Georgia doctor convicted in 1998 of stealing more than $10 million of drug research money in experimental tests and sentenced to 15 years in state prison.

Psychiatric Drug Experiments

In 1990, Borison, the chairman of the psychiatry department at the Medical College of Georgia in Augusta, hired WIRB to oversee his experiments with psychiatric drugs. During the time WIRB was monitoring him, Borison stole the money provided for clinical trials by Pfizer, Wyeth and Basel, Switzerland-based Novartis AG.

As a department chairman, Borison was required by college rules to use the school's IRB. Instead, the doctor used WIRB, located 2,300 miles away, to help conceal his fraud from the school, says George Schuster, chairman of the college's IRB.

``Borison bypassed us and went to WIRB,'' he says. ``We didn't know until the whole thing blew up that they were using WIRB. If WIRB had followed its own rules, we'd have notified them it wasn't acceptable. We wouldn't have allowed the fraud to continue.''

Forged Signatures

WIRB's rules required it to notify a school when it was hired to oversee research. Bowen says WIRB didn't inform the Medical College of Georgia because Borison had told WIRB he was a part-time professor. Letters from Borison to WIRB were on the school's letterhead, listing Borison as chairman of the psychiatry department.

In his indictment, Borison was also accused of endangering the lives of participants by using inadequately trained employees and permitting his signature to be forged on prescriptions.

An FDA inspection report on Borison in 1997 also detailed patient protection violations, finding that untrained employees administered experimental drugs, evaluated side effects and decided when to increase dosages. The FDA sent its findings to WIRB, which had allowed Borison's tests to proceed for six years.

Today, seven years after Borison's conviction, Bowen says WIRB did nothing wrong in its oversight of the Georgia tests. ``I didn't see that there were patient safety issues,'' says Bowen, who sat on the panel that oversaw Borison's experiments.

`That's Just Bogus'

WIRB told its staff to send its research approvals directly to Borison's home and not to the school, according to WIRB documents obtained by state prosecutors. An undated WIRB memo says, ``Arrangement with Dr. Borison is to have all correspondence sent to his home address.''

Bowen says WIRB clients are free to use any address. ``We send it to where they ask us to,'' she says. ``We didn't know it was his residence.''

Prosecutor David McLaughlin of the Georgia Attorney General's Office in Atlanta says he was astonished by Bowen's attitude about Borison.

``I'm a prosecutor, sitting in her office, telling her they did this and that, and she was saying, `It's not a problem for us,''' he says. ``That's just bogus. I had such a bad taste in my mouth when I left.''

The state brought no charges against WIRB.

Pfizer spokesman Lederer says the results of Borison's research were removed from Pfizer's database and weren't sent to the FDA. Wyeth spokesman Burr declined to comment.

Genentech Test

In addition to monitoring phase I trials, WIRB plays a leading role in supervising phase II and III trials. In a phase II test, clinicians experiment with various doses of a medicine to test effectiveness.

In phase III, they aim to collect enough data on larger groups of patients to demonstrate that the substance works well enough to be approved by the FDA. In all phases, clinicians monitor for side effects.

In 2000, Bill Hamlet, a 58-year-old artist and woodcarver in Pittsboro, North Carolina, entered a phase III clinical trial for a proposed psoriasis treatment made by Genentech Inc. Hamlet enrolled on the recommendation of his physician, Mark Fradin, 45, a doctor running the test.

Hamlet says the medication he was taking before the test, methotrexate, successfully controlled his psoriatic arthritis, a condition causing inflammation of the skin and joints.

When Hamlet began the drug experiment, his doctor instructed him to stop taking methotrexate. He became sick after going off the medication. During the trial, he spent weeks in bed because he was barely able to walk. Hamlet was left with permanent knee damage, his medical records show.

`A Train Wreck'

``It was like a train wreck,'' Hamlet says, recalling the pain and discomfort that became part of his life for six months. ``My whole persona was taken away in one fell swoop by a medical trial.''

When the test began, Hamlet wasn't told by his doctor that he might be given a placebo, a substance with no active medicine, he says. Nor was he told three months later that he had been switched to the experimental drug.

By design, many drug trials don't allow participants or clinicians to know who is getting placebos at the time of the tests.

Hamlet sued Genentech, Fradin and WIRB, which was overseeing protection for participants in the clinical trial. All three settled the lawsuit this year without disclosing terms.

Genentech spokeswoman Tara Cooper says the company can't comment because the settlement has a secrecy agreement. Bowen says that clinicians and WIRB did nothing wrong in the Hamlet trial. Fradin's lawyer, William Daniell, says the doctor did nothing improper.

Five Deaths

In 2001, WIRB was hired by Johns Hopkins School of Medicine in Baltimore to help review research at the school after a clinical trial participant died in the same year.

Minutes of WIRB meetings from the first quarter of 2004, which are available at the medical school because a Maryland law requires such minutes to be public, show shortcomings in WIRB's own review of research.

A recruiting script for participants was approved by a WIRB panel even after a doctor on that panel said she didn't understand it. She abstained from the vote.

The board also complained that it took seven weeks for WIRB's staff to inform the board of the death of five people in a clinical trial. WIRB deleted some details and all names from the minutes provided for review by Bloomberg News.

No-Show Panel Members

For the Johns Hopkins review, WIRB's nine-member panels often met with just five members present, the minutes show. Alternate members made up the majority of WIRB boards 20 times from Jan. 1, 2004, to March 31, 2004. Twice in three months, all of the members were alternates.

In 2003, the Hopkins minutes show, WIRB required a clinical trial sponsor to make changes in the recruiting materials for a trial in order to better protect participants in the experiments. The sponsor, whose name was deleted from the minutes by WIRB, asked WIRB to reconsider its decision.

On Feb. 26, 2004, the same WIRB panel, acting with four alternates and one of its regular members present, reversed its decision and allowed the company to keep its original proposed language.

``That was worrisome,'' Bowen says, after being informed of what had happened. ``I wish somebody had caught it sooner.''

`Manipulation'

Daniel Ford, vice dean for research at Johns Hopkins School of Medicine, says the reversal by WIRB concerns him. ``It's possible you could have manipulation,'' he says, ``One of the big things WIRB sells is speedy review.''

Ford says WIRB provides high-quality service.

The University of Pennsylvania's Caplan disagrees. He says WIRB has failed to protect participants in clinical trials.

``It appears they have basically reneged on their obligation toward subject protection and have become complicit in protecting the interests of their sponsors because it serves an important business interest,'' he says. ``That's just what you fear from commercial IRBs. They've had conflicts of interest since the beginning.''

In a 2002 Seton Hall Law Review article, WIRB's director of regulatory affairs wrote that there's an inherent conflict within independent IRBs because their fees come from the same pharmaceutical companies whose trials they're asked to monitor. ``The conflict of interest faced by independent IRBs is real and substantial,'' David Forster wrote.

No Federal Rules

``Independent IRBs are paid by sponsors and investigators to protect subjects who are participating in research conducted by those sponsors and investigators,'' Forster wrote.

There aren't any federal rules requiring for-profit IRBs, which are often located thousands of miles away from trial sites, to visit or inspect the test center at any time.

Nobody has ever studied the effectiveness of IRBs or tracked how many people are injured or killed each year while participating in clinical trials, says Harvard's Federman, who chaired a national committee on clinical trial safety in 2003.

``An intelligent person would assume we know this,'' Federman says. ``We don't know the number of persons harmed in clinical trials each year and are missing a registry of all subjects that participate in trials.''

Government agencies have repeatedly warned about inadequate protections for people in trials. ``Pressures to recruit subjects can lead researchers and IRBs to overlook deficiencies in efforts to inform subjects of potential risks,'' the Government Accountability Office cautioned 10 years ago.

`Race to the Bottom'

In 2000, the inspector general of the Department of Health and Human Services wrote, ``In a highly competitive marketplace, with few rules or guidelines governing recruitment, there is a very real danger of a race to the bottom.''

In 2002, after three people died in clinical trials at medical schools, bills were introduced in both houses of Congress to strengthen protections for people in drug tests.

The bills, sponsored by Democratic Senator Ted Kennedy of Massachusetts and Democratic Representative Diana DeGette of Colorado, stalled in committee and never made it to the floor for a vote.

``I hope Congress will act,'' Kennedy says. ``Recent failures of the current system have given new urgency to the need to guarantee the safety of clinical research and prevent similar tragedies in the future. We need to protect research participants.''

Panel Folded

Testing companies must fully inform people of risks in clinical trials, says Senator Charles Grassley, a Republican from Iowa. ``The burden is on the research companies to go out of their way to make sure study participants are fully informed when consent is given,'' Grassley says. ``Patient safety should never be sacrificed for short-term profit by a corporation.''

The National Bioethics Advisory Commission, a presidential panel created in 1995 by executive order of President Bill Clinton, issued human protection recommendations in 2001.

That panel folded in 2001, and President George W. Bush replaced it with the President's Council on Bioethics, which has issued reports on ethical issues of human cloning and stem cell research.

``Business has taken a much higher profile at the FDA because of the current administration,'' says Mary Faith Marshall, associate dean for social medicine and medical humanities at the University of Minnesota Medical School in Minneapolis. ``There's a much friendlier attitude toward Big Pharma and less emphasis on human subject protection.''

`An Independent Agency'

``The FDA is an independent agency,'' White House spokesman Trent Duffy says. ``It has maintained its independence. President Bush supports a strong FDA that protects American consumers.''

Harvard's Federman says politics is at issue. ``This type of inquiry is not a high profile for the current administration,'' he says. ``This is not a government that particularly looks at big business. Pharmaceutical companies have a huge lobbying operation.''

PhRMA, which represents more than 40 drug companies, spent more than $16 million last year on lobbying, a 12.5 percent increase from the year before. PhRMA hired 136 lobbyists in 2004, according to Public Citizen. PhRMA declined to comment about its lobbying activities.

``PhRMA and its member companies are certainly willing to review proposals that could make a good safety record even better,'' the group says.

One way pharmaceutical companies could improve safeguards for clinical trial participants is by checking to see whether the people running the tests are actually licensed as doctors.

Drug Study Institute

In Jupiter, Florida, a drug testing center called the Drug Study Institute lists its director of clinical research as Melody Sanger, who's identified as a primary care physician. Florida state records show Sanger, 50, isn't a licensed doctor. She's licensed only as a registered nurse, according to the Florida Department of Health.

The company Web site says she has run trials for AstraZeneca, Merck, Novartis and Pfizer.

Sanger never misrepresented her credentials to Merck, company spokeswoman Skidmore says. AstraZeneca spokeswoman Carla Burigatto says the Drug Study Institute did good-quality work, adding that Sanger didn't serve as a doctor on trials for the company. Sanger declined to comment.

SFBC describes Chairwoman Krinsky as a medical doctor in SEC filings and company literature. She's never been licensed to practice medicine in the U.S., SFBC's Hantman says. Krinsky's laboratory technician license in Florida expired in 1998. Krinsky is in charge of SFBC's phase I clinical trials.

`It's Misleading'

Hantman says Krinsky is a company executive who doesn't run any clinical trials. ``She is not required to be licensed in Florida,'' he says. Hantman says the SFBC center has five physicians, as well as nurses and emergency personnel.

Harvard's Federman is concerned that SFBC refers to Krinsky as a doctor without disclosing she's not licensed. ``It's misleading in that most, perhaps almost all, readers would assume she is a licensed and fully trained physician,'' he says.

Hantman is SFBC's treasurer as well as its CEO. Company SEC filings say he's a certified public accountant. Hantman's Florida CPA license expired in 1989, public records show. Hantman says he's been a lifetime member of the American Institute of Certified Public Accountants, a trade organization.

In Houston, the Fabre clinic used Rodriguez to give experimental drugs to people and make medical decisions during tests. The FDA found that Rodriguez had neither a medical license nor any clinical credentials in the U.S.

Better Ways

There are better ways to do research, says Koski, the physician who headed the federal agency for human protection for two years. Koski says a single U.S. agency should oversee all experimental tests.

The National Bioethics Advisory Commission suggested that informed consent discussions between researchers and participants be audio- or videotaped to ensure they're done right.

The commission also recommended a system to compensate people for research-related injuries and said all IRBs should have to register with the federal government. In addition, it said all IRB members should be trained in research ethics.

Mark Yessian, who oversaw investigative reports on IRBs over the past decade as Boston's regional inspector general for the Department of Health and Human Services, says changes are needed.

``The drug industry is trying to bring products to market,'' says Yessian, who retired in October. ``We don't want to suffocate that, but we need to do it in a more balanced way to give subjects confidence that there are people looking out for their interests.''