Forbes.com

Medical Technology
Fixing Hospitals

Robert Langreth, 06.20.05

Medical errors kill 100,000 Americans every year. A new vanguard is out to fix the fatal flaws.

Josie king was a curious and precocious baby. At 18 months old she loved jumping on the trampoline and had recently learned to say "I love you." She toddled into the bathroom one day when no one was looking, slipped into the tub and turned scalding hot water on herself. After ten days in the pediatric intensive care unit at Johns HopkinsChildren's Center in Baltimore, her burns were healing so fast that doctors moved her to an intermediate care ward with hopes of releasing her soon. Her three older siblings started planning a coming-home party.

Then things went terribly wrong. Her mother noticed that Josie started screaming for liquid every time she saw a drink. At one point the girl desperately sucked water out of a washcloth. But nurses told Josie's mom, Sorrel King, to resist giving her daughter any drinks as she had been vomiting. That evening Josie's eyes started rolling back in her head, but when Sorrel pointed this out she was told Josie's vital signs were fine. The next morning Sorrel persuaded doctors to let Josie drink, but then one doc ordered heavy-duty narcotics against another's instructions. A few minutes later Josie's heart stopped. She was taken off life support two days later, dying of dehydration and bad medicine in one of the nation's premier hospitals.

"It was such simple stuff that went wrong; it wasn't one doctor or one nurse or one decimal point. It was a huge systems breakdown," Sorrel King says. Johns Hopkins resolutely admitted its mistakes and cooperated with a Baltimore Sun series on the 2001 incident. The hospital paid the family an undisclosed settlement and agreed to work with King to fix the flaws, revamping its systems for a sharper focus on safe care. But the broader problem remains, King says. "The general public still doesn't get it that thousands and thousands and thousands of people are dying every year because of medical errors."

Shoddy quality control plagues American medicine, killing at least a hundred thousand people every year and running up an estimated $500 billion a year in avoidable medical costs, or 30% of all health care spending. The quality crisis fuels the litigation crisis: Medical malpractice costs hit $27 billion in 2003.

"American health care operates with levels of unreliability, injury, waste and just plain poor service that long ago became unacceptable in many other industries," says Donald Berwick, a Harvard pediatrician and a crusader for fixing flawed care.

Many of the needed fixes are surprisingly simple, and new technology has arrived to modernize the antiquated systems that make the medical industry one of the least wired businesses in the U.S. economy.
61% of doctors wash their hands before examining a patient if they know someone is watching.
 
44% wash their hands if they think no one is watching.
University of Geneva study; Annals of Internal Medicine.
After Josie King died, Hopkins Children's added a multidisciplinary team to each floor--one unit is named after her--and started holding meetings twice a month to raise safety concerns.A new safety director visits nurses weekly to unearth impending problems. Some 145 issues have come up in two years, and more than 100 of them have been fixed. When it emerged recently that two asthma drugs came in vials so similar as to risk confusing them, the staff confiscated current stocks, and the hospital got a new contract for two drugs with dissimilar packaging.

Yet many hospitals are painfully slow to embrace a revamping. One of the biggest obstacles: American medicine's cult of the "doctor is God," says Harvard's Lucian Leape, an early researcher in doctor errors. Doctors, who typically are independent of a hospital's direct control and are free to prescribe drugs for uses that haven't been approved by regulators, resist any second-guessing and perceived threats to their autonomy. Trying to reform the system sparks turf wars and staunch resistance. Only 34% of doctors have participated in quality-improvement efforts, one study says; two-thirds of doctors in one recent survey oppose disclosing their track records to the public.
5% Portion of revenue hospitals spend on IT.
 
9% Portion of revenue the financial services industry spends on IT.
Forrester Research.
The doctor-as-God approach "doesn't work in the 21st century," says Dr. Leape, a professor at the Harvard School of Public Health, not with the proliferation of new drugs and tests and tools and surgical procedures. Hospital patients often take dozens of drugs simultaneously and undergo scores of procedures and tests every day. Doctors and nurses still try to keep track of it all using paper, pen and clipboard--just as they did 50 years ago. New tools and better vetting of doctor decisions are critical to fixing this flawed system. Leape calls it "a tectonic shift in the practice of medicine."

"There is a massive gap between where we are and where we could be," says Dr. Brent James, a vice president at Intermountain Health Care in Salt Lake City, Utah, a big hospital chain that was one of the first to focus on systemic quality improvements. Better quality control, says Dr. Kenneth Kizer, who heads the National Quality Forum, "has the potential to do more for health care than any foreseeable technology improvement--including a cure for cancer."

The good news:After years of an industry in denial, some pioneering doctors and hospital chains, including Intermountain, Ascension Health in St. Louis and OSF Healthcare System in Illinois, are finally taking action. A reform movement is gaining ground, aiming to pull off a radical overhaul of America's 5,764 hospitals. Some 2,000 hospitals have joined an initiative known as the "100,000 Lives" campaign led by Harvard's Berwick, chief of the nonprofit Institute for Healthcare Improvement in Cambridge, Massachusetts. He got religion after seeing his wife hospitalized for an autoimmune condition in 1999 and learning that "errors were not rare, they were the norm," as he put it later in a controversial speech; one urgently needed drug took 60 hours to get to her room. Now Berwick's 80 employees scour the world for simple, proven remedies for medicine's reliability woes. They farm out recommendations to the collaborating hospitals.

In this new focus on error-free medicine, hospitals are dropping age-old assumptions that their doctors are doing a good job and instead are measuring performance and issuing report cards. They are standardizing basic procedures to help reduce errors and condensing long-winded practice guidelines that harried doctors don't read into simple checklists integrated into daily work flow. They are instituting teamwork training, encouraging physicians to swallow their egos and go on rounds with nurses and empowering nurses to challenge doctors when they spot potential goofs--an especially heretical act. They are pushing for the digitization of medical records, at long last, so that crucial test results and other vital records can be tracked and communicated. In one controversial move, reformers are pushing hospitals to release their performance results to the public.

It may be too much change too fast for some. "Doctors are wary of changes that reduce their ability to act autonomously," says Berwick. Many in medicine, however, say they have no other choice but to adopt big changes. "There are three parts to medicine: understanding disease, identifying effective therapies and then making sure the therapies are used safely and consistently. We have totally neglected the third part," says Dr. Peter Pronovost, an intensive care specialist and safety researcher at Johns Hopkins University. His father died in 1990 in excruciating pain from misdiagnosed cancer. "I was told that I had to accept it. I thought patients deserved more," he says.

The reform movement doesn't yet have definitive clinical results, but early results are promising. Innovative hospitals are showing they can drastically cut medication errors and all but eliminate some deadly hospital infections by consistently applying known steps. "We don't have any idea of how low we can go, because the benchmarks are all based on a broken system. But we are seeing performances that are unheard of," says Pronovost.

He works on one project aimed at reducing the death rate from sepsis, a leading killer, and says thousands of lives could be saved each year if every hospital implemented the measures used in the study. It involves a network of 20 intensive care units at Johns Hopkins and elsewhere; the overall rate of sepsis deaths has fallen from 26% of sepsis patients in late 2003 to 20% now. The rate fell thanks to the addition of systems to ensure the simple stuff--that every patient with the ailment is quickly identified and immediately gets all recommended treatments.

For decades doctors figured occasional mishaps were inevitable and that the only solution was to work harder. One of the first to investigate the issue more deeply was Harvard's Leape. In 1991 he was part of a group of researchers seeking to catalog medical errors; they reviewed 30,000 in-patient records from 51 New York hospitals and found that roughly 4% of patients had been seriously harmed by their care. Two-thirds of the injuries were caused by errors. This later was extrapolated to estimates of almost 100,000 deaths from errors nationwide, and since then studies around the world have echoed the New York results.

Curious, Leape went to Harvard's vast medical library to see what had been written about preventing medical error. To his astonishment he was unable to find a single journal article on the subject. No one had even considered the matter, so alien was the concept that bad systems, not just bad doctors, might be responsible for patient injuries. Leape did turn up extensive safety literature from cognitive psychology, aviation and nuclear power, gleaning lessons--reducing reliance on memory, standardizing processes, promoting teamwork--that might apply well to medicine. He wrote his ideas up for publication, likening medicine's injury problem to "three jumbo jet crashes every two days." The article, published in late 1994, didn't go over well with most of his colleagues. But some saw the wisdom, and the patient safety movement began.

A decade later some hospitals are achieving miraculous results from rethinking the way they deliver and monitor care.

Getting Back to Basics

In Utah, intermountain's LDS Hospital was an early innovator in studying medical reliability. LDS researchers benefited from one of the U.S.' first computerized medical record systems, in place by the early 1980s, which provided a trove of data to track real-time results and spot problems. One early focus was slashing postsurgical infections. In the mid-1980s LDS infectious disease specialist John Burke and colleagues found only 40% of surgery patients at LDS were getting antibiotics during the optimal time window, zero to two hours before an operation. The remainder had vastly higher infection rates. Later other studies confirmed the problem was prevalent. Today 99% of surgical patients at LDS get antibiotics within the correct window. The biggest changes were cultural, Burke says. "We only made progress when we got the team together and got them to commit to using a checklist, like in a 747 before takeoff."

Burke and colleague David Classen then turned their attention to drug reactions. They designed software that scans drug orders and lab tests in the hospital database and flags problems. It showed only 1 in 60 toxic drug reactions were being reported through normal channels; in most cases doctors hadn't even noticed the problem. Today the software is used at LDS to spot mismatches in real time.

By the Numbers
 
A Casualty Count
 
3% or more of hospital patients are hurt by medical error.
 
1 in 300 patients die from such mistakes.In U.S. aviation only 1 in 5 million flights ends in a deadly accident.
 
24% of people say they or a family member have been harmed by a medical error.
 
90,000 people die of hospital-acquired infections annually. More than half of these deaths may be preventable.
 
180,000 elderly outpatients die or are seriously injured by drug toxicity. Half of these incidents may be preventable.
 
7,000 patients die from drug errors each year.
 
554 errors in four months were found at one six-bed intensive care unit; 147 were potentially serious or life-threatening.
 
55% of recommended care actually gets administered.
 
$2,000 Annual cost to employers per insured worker, due to poor-quality care.
 
Sources: Lucian Leape; New England Journal of Medicine; CDC; Forbes; Journal of the American Medical Association; Institute of Medicine; Quality and Safety in Health Care; RAND; Midwest Business Group on Health.
 
For years fixes at LDS and other Intermountain hospitals were piecemeal. That changed when a new executive vice president, William Nelson, was appointed in 1995 (he became chief executive in 1999). An accountant, he realized that higher reliability would save lives and money. Nelson corralled quality director Brent James and pressed him to delve into the unglamorous nitty-gritty of changing the entire hospital chain.

The first step was to revamp Intermountain's database to better link patient records at its hospitals and clinics to track how well each patient was faring in various measures of care, such as whether heart patients were getting the right drugs. This task took 20 programmers and statisticians over a year.

The harder step was persuading physicians to buy in. Intermountain hired 21 respected doctors part-time and an equal number of full-time nurses to act as peer reviewers for doctors at Intermountain's far-flung operations. They issued monthly reports to the hospitals on how various departments were doing. When data showed that many expectant mothers at its hospitals were being induced two weeks or more early simply because it was more convenient, even though national studies hinted it is safer to wait until 39 weeks, Intermountain clamped down in 2001. Some obstetricians objected, resenting the intrusion. "They tried to sabotage it," says Dr. Bryan Oshiro, who led the program. "They said the studies are wrong. ‘Our patients are different. This is just administrators trying to tell us what to do.'" Oshiro's team reanalyzed Intermountain's own data and showed that deliveries even a week early had higher complication rates. The critics faded.

Results have been rolling in. Annual deaths from congestive heart failure plummeted 22% by adding (and enforcing) a checklist to ensure all patients get the right medicines before they leave the hospital. A mandate to test all newborns for high bilirubin levels, which can lead to brain damage, slashed the number of infants requiring readmission by 24%. Various childbirth measures have helped reduce Intermountain's cost per delivery by $350 in four years, saving $10 million annually.

"What we are doing is simple but had never been done because there was no process and no accountability," says Nelson, Intermountain's chief executive. General Electric's health IT unit is so impressed with Intermountain's algorithms that it is opening a center in Utah to adapt them for mass use.

Touting Teamwork

Intensive care units specialize in stopping gravely ill patients from slipping over the edge. But what happens to patients in other wards who deteriorate? All too often a coordinated response doesn't occur until the patient's heart stops--and by then the patient has only a 15% chance of surviving. Signs of distress, such as breathing difficulties, usually appear six to eight hours before cardiac arrest, but nurses and junior residents may lack the expertise to stabilize patients on their own, and it can take hours to get an experienced doctor's help. "You have a universal problem: People who get sicker in the hospital don't get urgent care quickly," says Rinaldo Bellomo, an intensive care specialist at the University of Melbourne in Australia.

The solution: 24-hour emergency response teams that nurses can call, bypassing the chain of command, when they perceive a patient is in trouble. The teams bring together up to eight doctors and nurses with expertise in critical care and resuscitation. Some 1,200 hospitals now work with Harvard's Institute for Healthcare Improvement to implement these teams.

The University of Pittsburgh Medical Center adopted the teams in the late 1990s and has seen unexpected deaths drop 30% since then, says Michael DeVita, a critical care specialist at the hospital. DeVita says he hears from nurses at other hospitals who say they tried to get emergency attention for deteriorating patients only to be lectured never to call the cardiac-arrest team until the patient arrests. "It is unbelievable, but it is true," he says. "This should be mandatory for all hospitals." Bellomo's Austin Hospital in Melbourne was among the first to study the impact of the roving teams. It started using them in November 2000, and soon cardiac-arrest alerts over the hospital's public address system, once routine, stopped happening for weeks at a time. In the four months after implementing the teams, cardiac arrests plummeted 65%.

Other hospitals are trying to avert breakdowns in information sharing. In medical mistakes "the same theme comes up time and time again: lack of communication and confusion among team members," says Harry C. Sax, chief of surgery at Miriam Hospital in Providence, Rhode Island. In one published case, a name mix-up resulted in the wrong woman getting an invasive heart procedure; at least five nurses and three doctors disregarded the woman's objections or failed to verify her identity.

To fix the situation researchers hope to borrow from aviation, running checklists before operations and doing simulator training to practice teamwork. Last October surgeons at the Michael E. DeBakey Veterans Affairs Medical Center in Houston started doing preoperative briefings with nurses and anesthesiologists to verify the plans for the surgery and make sure all necessary preoperative drugs have been given.

Taming Infection

Each year 2 million hospitalized americans develope additional infections during their stays; 90,000 of them die, according to the Centers for Disease Control. "It is staggeringly outrageous that this problem is allowed to continue and proliferate," says patient advocate Michael Bennett of Baltimore. His 89-year-old father, hospitalized for a respiratory infection, died last June after contracting six bacterial infections in five hospitals over four months. Bennett plans to sue one hospital (he won't name it) and is pushing for laws requiring hospitals to make infection rates public. "I know that my dad would still be alive today" if he hadn't gone to the hospital, Bennett says.

For years doctors have assumed that certain rates of infections were simply the cost of doing business. But now hospitals are finding they can slash rates of killer infections by 50% or more by changing how they do things. One big source of deadly infections is surgical wounds, which hit 2% of surgical patients. Another rampant problem is ventilator-associated pneumonia, which afflicts anywhere from 10% to 25% of patients on the machines. The longer you remain on a ventilator, the greater the odds that germ-laden gunk from the back of the throat will get sucked into the lungs. Some embarrassingly low-tech steps can make a difference, if only hospitals would make the effort to do them. One is tilting the bed to 30 degrees, so secretions are less likely to bubble into the lungs. Another is minimizing the time patients are on ventilators, removing sedation once a day in order to assess whether patients are ready to be weaned.

The final killer is blood infections from central IV lines installed into veins near the heart; these finish off 28,000 patients a year. Installing a central line is a tricky procedure that is often done by inexperienced residents. Sterility guidelines, such as draping the patient in a sterile sheet and wearing a mask and gown, are often forgotten. To make it easier to do it right, some ICUs now gather all the equipment needed on a single cart ahead of time and give nurses permission to halt a central line procedure if they see a doctor doing it wrong. When Johns Hopkins Hospital did this and took other steps in a surgical ICU, rates of catheter infections dropped to nearly zero.

Sidestepping Drug Toxicity

Patricia Unger had worked at Rogue Vally Medical Center in Medford, Oregon for 18 years before being diagnosed with lymphoma and undergoing treatment there. In a high-dose chemo regimen, caregivers allowed blood levels of the toxic chemo drug methotrexate to soar to a thousand times above the safe level when a doctor misread a lab test, says the late Unger's son,Mark Unger. Soon there was nothing left to do but watch her burn to death from the inside out.

"Bedding was dripping on the floor from all the fluid coming out of her," he recalls. Unger settled a malpractice lawsuit against the hospital and doctor for an undisclosed amount. "There was no system to catch the error, no safety net at all," he says. The hospital says that it has added multiple new safety systems including processes to prevent overmedication.

Drug toxicity is the single most common threat, hitting 6.5% of hospitalized patients. Information technology can help prevent errors and catch drug interactions before they do much damage. Brigham andWomen's Hospital has been a pioneer here. Computerized doctor order systems can reduce harmful errors by 55%, says one study by David Bates at Brigham, which introduced such a system in 1993. But the order systems are expensive to install and tricky to get right, involving a big change in work flow.

In the meantime Dr. Berwick and IHI are working with hospitals on simpler measures that can slash drug-related injuries. Many drug mix-ups occur when a patient is admitted, discharged or transferred; carefully reconciling a patient's drug list every time he is moved can avoid this problem. OSF Healthcare in Peoria, Illinois has reduced drug-related injuries by almost 75% since 2001 with this and other techniques.