The Times found instances in which the recusals did not work as intended.
In the mid-to-late 1990s, Eastman, the diabetes researcher, participated in a series of decisions affecting the drug company employing him as a consultant, despite having signed a recusal. Separately, Katz, the director of the arthritis institute, signed a recusal involving his client, Schering AG, which nevertheless supplied the NIH with the drug involved in the kidney patient's death in 1999.
Katz said that he did not know at the time that Schering AG was the maker of the drug his institute was testing.
Compliance with the recusals can, itself, undercut the interests of the NIH and taxpayers, who support the agency. When heads of institutes and laboratories recuse themselves, they sometimes constrain their ability to carry out their government duties.
Kirschstein, who for the last eight years has personally reviewed requests from the institute directors to consult privately for pay, said she tended to approve the deals, unless she saw "real conflict."
"I've disapproved some -- and I've approved many," she said.
In her view, recusals have worked "extremely well" in avoiding conflicts of interest.
Other present and former officials say it is difficult, if not impossible, for researchers to keep separate their confidential government information when they consult for companies.
"You can't police the thing," Philip S. Chen Jr., a senior advisor in the NIH director's office who has served as an agency scientist or administrator since the 1950s, said in an interview last year. "The rules are there -- whether they follow the rules is another thing."
A former NIH director voiced surprise at the agency's loosened approach to conflicts of interest.
"There has been a lot of relaxation," said Dr. Bernadine P. Healy, who served as director from 1991 to 1993. Before, Healy said in an interview, "there were very strict ethics rules for NIH scientists. You couldn't have virtually any connection with a company if your institute was in any way doing research involving their products."
At least one vestige of the old days remains.
During last year's holiday season, workers were advised to refuse gifts from outsiders worth more than $20.
"Just a reminder," ethics coordinator John C. Condray wrote, introducing a five-page memo, "that sometimes gifts and events can create the appearance of a lack of impartiality."
