WASHINGTON -- Government-funded researchers tested AIDS drugs on hundreds of foster children over the past two decades, often without providing them a basic protection afforded in federal law and required by some states, an Associated Press review has found.

The National Institutes of Health-funded research spanned the country. It was most widespread in the 1990s, as foster care agencies sought treatments for ther HIV-infected children that weren't yet available in the marketplace.

The practice ensured that foster children, most poor or minorities, got care from world-class researchers at government expense, slowing their death rate and extending their lives. But it also exposed a vulnerable population to risks of medical research and drugs known to have serious side effects in adults and for which safety in children was unknown.

The research was conducted in at least seven states -- Illinois, Louisiana, Maryland, New York, North Carolina, Colorado and Texas -- and involved more than four dozen different studies. The foster children ranged from infants to late teens, according to interviews and government records.

Several studies that enlisted foster children reported that patients suffered side effects such as rashes, vomiting and sharp drops in infection-fighting blood cells. In one study, researchers reported a "disturbing" higher death rate among children who took higher doses of a drug.

The government provided special protections for young wards in 1983. They required researchers and their oversight boards to appoint independent advocates for any foster child enrolled in a narrow class of studies that involved greater than minimal risk and lacked the promise of direct benefit. Some foster agencies required the protection regardless of risks and benefits.

But researchers and foster agencies told AP that foster children in AIDS drug trials often weren't given such advocates, even though research institutions many times promised to do so to gain access to the children.

Illinois officials believe that none of their nearly 200 foster children in AIDS studies got independent monitors, though researchers signed a document guaranteeing "the appointment of an advocate for each individual ward participating in the respective medical research."

New York City could find records showing that 142 of 465 foster children in AIDS drug trials -- fewer than a third -- got monitors, even though city policy required them. The city has asked an outside firm to investigate.

Some states declined to participate in medical experiments.

Officials estimated that 5 percent to 10 percent of the 13,878 children enrolled in pediatric AIDS studies funded by NIH since the late 1980s were in foster care. More than two dozen Illinois foster children remain in studies today. Some foster children died during studies, but state or city agencies said they could find no records that any deaths were directly attributed to the treatments.

Researchers typically secured permission to enroll foster children through city or state agencies. And they frequently exempted themselves from appointing advocates by concluding that the research carried minimal risk and that the child would directly benefit because the drugs had already been tried in adults.

Those who made the decisions said the research gave foster kids access to drugs they otherwise couldn't get. They said they protected the children's interest by explaining risks and benefits to state guardians, foster parents and the children themselves.