When Dr. Peter Pronovost assembled a checklist of proven safety procedures and required critical care doctors to use it, the results were indisputable: It prevented infections and saved lives.
But the veteran Johns Hopkins safety researcher also drew critical scrutiny from government regulators. In a decision last month that upset his colleagues nationwide, the federal Office of Human Research Protections said using a safety checklist - and studying its effects - amounted to conducting an experiment without a patient's consent.
So the agency directed the 70 Michigan hospitals that participated in Pronovost's study to stop sending him data on future infection rates.
Both Pronovost and federal officials play down the dispute. But elsewhere, researchers say the government's heavy-handed treatment of a simple, commonsense project that has saved more than 1,500 lives - at almost no cost - could delay similar safety initiatives nationwide.
"I think it's ludicrous. It probably will kill people if they don't fix this soon," said Dr. Robert Wachter, a national expert on patient safety at the University of California, San Francisco. "This is regulation run amok."
Richard J. Umbdenstock, president of the American Hospital Association, called the decision "wholly inappropriate and detrimental to the patients and communities we serve."
Experts considered Provonost's findings, reported in the New England Journal of Medicine in 2006, a major step in efforts to reduce the estimated 90,000 deaths from hospital infections each year.
Both federal officials and Pronovost, a Hopkins professor and and medical director of the Center for Innovation in Quality Patient Care, emphasized that hospitals in Michigan, Maryland and elsewhere are continuing to use intensive-care unit checklists and that no sanctions would be imposed.
"Punishment was not on the table. We told them to take corrective action, and they have," said Ivor Pritchard, acting director of the OHRP.
But before Pronovost can continue to study infection rates in Michigan's intensive-care units, hospitals there must have the work reviewed by independent oversight groups called institutional review boards.
In some cases, that could take months, researchers said. They suggested that federal officials misapplied regulations originally designed to protect patients in clinical trials of experimental drugs and procedures.
"It's a bit of a bump in the road," said Christine Goeschel, manager of operations for Pronovost's patient safety group. "Any time a group pushes the envelope, there are going to be cases where existing rules and guidelines are not always a good fit."
Federal officials say they're impressed with Pronovost's results and encourage hospitals to use checklists. All sides agree that the procedures involved - rigorous hand-washing, use of sterile masks, gowns and gloves, and proper handling of catheters - are accepted safety measures.
But the government said writing them down in a checklist, making doctors follow it and methodically analyzing the results involved a change of practice. That qualified it as a research project that requires formal patient consent and institutional reviews.
"If you're deliberately manipulating the way you deliver care and you study the effects for the purpose of gaining new knowledge from it, that's research," Pritchard said.
But some patient safety experts fear that the decision will delay or discourage a wide range of safety measures nationwide. For example, it could require patient consent before hospitals make such basic improvements as installing gel dispensers in hospital rooms to increase hand washing.
"We'll all be looking to see if we violated any rule by measuring any type of quality of care initiative," said UCSF's Wachter.
"It's just absurd on its surface," added Dr. Lucian Leape, an expert on preventing medical errors at the Harvard School of Public Health.
Leape formally asked the OHRP to rescind the decision in a letter that he also sent to his congressional delegation. "This ruling has the potential to cripple the truly impressive progress we are finally making in patient safety," his letter says.
The federal ruling could be "a big step backward," added Dr. William Minogue, executive director of the Maryland Patient Safety Center. Minogue said the state's hospital ICUs made checklists standard practice in 2005 with help from a expert panel that Pronovost chaired. He called the OHRP decision "shocking."
This week, the American Hospital Association asked the government to "immediately retract" statements implying that advance, independent reviews and patient consent are required for improvements such as checklists, computerized medical staff reminders and team approaches to patient care.
For their part, Pritchard and other officials say there's nothing to prevent a hospital from implementing programs to improve care, such as hand sanitizers or a computer upgrade that tracks a patient's progress. "If you're implementing what established as a best practice or something you think will help, that's not research," Pritchard said.
In Michigan, Pronovost focused on checklists designed to prevent ventilator-assisted pneumonias and infections from central venous catheters, which are tubes inserted in most intensive care patients.
The study was part of an effort to improve overall Michigan ICU safety known as the Keystone Initiative. "The number one savings was in lives," said Sam R. Watson, executive director of what is now the Keystone Center for Patient Safety and Quality.
Within 18 months after Michigan hospitals introduced the checklists in 2004, central line infections dropped by two-thirds. But according to Watson, the biggest achievement was a permanent change in attitudes among medical personnel.
"You need to make a significant effort to change a culture, so the tool, in this case the checklist, doesn't become just a passing fad," he said.
In New Jersey, a similar project reduced central line infections by 73 percent. Ventilator-associated pnuemonias there declined by 55 percent, according to Alene Holmes, a vice president of the New Jersey Hospital Association.
Pronovost, working with experts in other fields, has developed checklists for patient care in surgery, emergency medicine, pediatrics and other specialties. He said he hopes to begin checklist projects with medical groups in Tennessee, California and Washington state.
"Doctors and nurses are so busy, we spend a lot of our time just putting out fires," said Pronovost, whose efforts were profiled in a recent New Yorker article.
Pronovost has been in Switzerland this week to help the World Health Organization set up safety programs for hospitals around the world. "It's inefficient for every hospital to invent his own wheel," he said.
Checklists are hardly new ideas: They have been standard tools for years in the automotive and airline industries.
But patient safety experts began to consider them about eight years ago. That's when the first of two reports from the Institute of Medicine showed that hospital deaths were often caused by the failure of doctors and nurses to take routine precautions, such as washing their hands.
The 1999 IOM report, entitled To Err Is Human, estimated that up to 98,000 patients died annually from preventable medical errors.
"As physicians taking care of patients, there was this concept of our own infallibility, and that's a dumb way of organizing things when there are a thousand moving pieces, when there are all these patients to take care of, I have people all around me and my beeper's going off every two minutes," Wachter said.
Here's the five-item safety checklist that medical personnel in Michigan intensive-care units employed to reduce infections and save patients' lives:
Rigorously wash hands with a sanitizing chlorhexidine soap
Wear sterile masks, gowns and gloves
Avoid inserting catheters in the groin area
Thoroughly clean and use barrier protections in areas around catheters
Remove catheters as soon as possible
[ Dr. Peter Pronovost, New England Journal of Medicine, 2006]