Under The Knife The health-care industry's archaic information technology is a major cause of runaway costs. One doctor's struggle to slash red tape at his hospital shows how IT can save dollars and lives--and how far reforms still have to go.
By Melanie Warner

(Business 2.0) – Emergency room physician John Halamka is on one of his evening shifts at Beth Israel Deaconess Medical Center in Boston. "Got a new admit for you, John," a nurse yells, as Halamka watches a heavyset middle-age man heading toward him on a stretcher. The patient, Joe S., complains of labored breathing and occasional dizziness, but insists he's fine--his wife made him come in. "OK, let's get his chart," the doctor commands, and a young resident wheels over a cart bearing a laptop computer. Halamka logs in to the hospital's network and discovers that Joe had an EKG done at the hospital a year earlier.

The physician clicks open the old EKG, and a trace of Joe's heartbeat blooms on the screen. Halamka and other doctors peer closer, eyes darting from last year's scan to the squiggles scratched out by the electrocardiograph to which Joe is now hooked up. Halamka utters a new command: "Let's get him into the CCL. Now." Despite Joe's protestations of health, he's having a heart attack. CCL is the cardiac catheterization lab, where physicians will race to save his life.

Joe is in deep trouble, but in one sense, he's a lucky man. He's at Beth Israel. Most hospitals don't have computer systems that let doctors instantly view a patient's past records, saving life-or-death seconds. According to trade publication Health Affairs, some 80 percent of hospitals and 95 percent of doctor's offices use the same methods for storing and accessing patient data that they did 50 years ago--which is to say, sheets of paper and film buried in huge metal cabinets. At those hospitals, a doctor who wants to see an EKG or other patient data must dispatch a request and wait anywhere from a few hours to an entire day for someone to retrieve it.

Even people who work for the federal government marvel at the inefficiency of such a system. "It's a brain-dead way to do things," says Carolyn Clancy, director of the government's Agency for Healthcare Research and Quality. Clancy calls the current state of affairs the "Marcus Welby system," after the '70s TV show about a doctor who writes everything on paper in a code no one else can understand.

In the real world, this situation carries a harrowing toll. Health-care costs are locked in a runaway spiral--up 20 percent since 2001, a time of near-zero inflation in the rest of the economy. While no one thinks technological backwardness is solely to blame, the dearth of IT is clearly a major factor. Mind-numbing inefficiencies add hundreds of billions of dollars to employers' and employees' health-care bills, while medical errors, many of them preventable with even rudimentary IT, kill up to an estimated 98,000 people a year.

Any attempt to improve the pitiful state of health-care IT faces formidable obstacles. Money is perennially tight in an industry with puny profit margins and many high-profile basket cases, such as HealthSouth and Tenet Healthcare. As things stand, health-care IT investment runs at just 3.9 percent of revenue, which pales in comparison with that of other industries; telecom, for instance, spends an average of 7.9 percent. A deeply rooted technophobia among doctors and administrators further complicates any reform attempt. For example, in late 2002, Cedars Sinai Hospital in Los Angeles installed software it had spent years developing with Perot Systems, only to rip it out six months later when doctors refused to use it. And finally, hovering over everything is the politically explosive issue of patient privacy. One chief information officer actually received death threats for spearheading a plan to share medical records electronically among hospitals in Indianapolis.

Against that backdrop, Beth Israel stands out as something of a medical freak: a hospital that not only deploys the latest in high-tech medical machines but also manages information in a sophisticated way. Its online database, called CareWeb, contains records on 9 million patients. A computerized system automates orders for all prescriptions, lab tests, and IV drips. Physician requests are checked against patient data to make sure there are no drug interactions or allergies, then are routed automatically to the pharmacy or lab. Illegible prescriptions, the bane of pharmacists everywhere, are a thing of the past. So are many medication-related errors, which have dropped 50 percent since the system was installed in 2001. "This system is designed to take out the guesswork," says Halamka, who should know--he designed it. In addition to being an emergency room physician, he's the CIO at CareGroup, the company that owns Beth Israel and four other Boston-area hospitals.

A boyish 41-year-old, Halamka is part of a growing group of tech-savvy health-care insiders who are finally starting to change a stubborn industry. More and more hospitals are looking to places like Beth Israel as a model for how to provide better care and rein in costs. Indeed, Gartner Group predicts that the industry will be the fastest-growing sector of the economy for IT spending during the next four years.

But change won't come easily. There are no magic bullets for health care's condition. Indeed, as shown by any number of IT deployment disasters like Cedars Sinai's, hospital systems seem particularly resistant to costly, one-shot-cures-all tech fixes. Halamka's experience suggests a better, if no less arduous, course of treatment: dragging health care into the modern IT age, doctor by doctor, hospital by hospital. And even then, it can be touch and go.

When Halamka became CIO of CareGroup in late 1998, he walked into a hornet's nest. Beth Israel and Deaconess Hospital had merged two years earlier, and the marriage had left the company in financial straits. There was political infighting (in the spring of 1998, the entire 40-member anesthesia team from Deaconess left in a management dispute) and two completely different, largely incompatible computer systems. Staff at one facility had to print out files and then fax them or walk them across Brookline Ave. to the other hospital. Patients who'd been to both hospitals had different ID numbers, often causing enormous confusion.

James Reinertsen, CareGroup's CEO at the time, turned to Halamka to clean up the mess. Normally, naming an ER doctor CIO would be an odd choice, but Reinertsen knew that Halamka was no ordinary physician. At 12, Halamka had taught himself analog and digital logic, the basis of semiconductor design; at 18 he started a software company, which he ran all the way through med school. Halamka had also written two books on technology and was involved in technical issues at Harbor UCLA Medical Center in south-central Los Angeles, where he worked before moving to Beth Israel.

One night in December 1998, Reinertsen called Halamka on his cell phone. "Hi, this is Jim Reinertsen," he announced. As it happened, Halamka was in the process of resuscitating a 95-year-old man. "Who?" Halamka asked. He'd never spoken to the CEO before. "We'd like you to come by tomorrow morning to talk about being CIO," Reinertsen said. Halamka was dumbfounded--he hadn't exactly campaigned for the job--but he jumped at the chance.

His first problem: capital. Like other hospitals, CareGroup was suffering from dramatic reductions in health maintenance organization and Medicare reimbursements. In 1999, Halamka's first full year as CIO, CareGroup lost $100 million. What money was available was earmarked for new medical equipment, an investment far more popular with both doctors and board members than IT.

So instead of buying applications from a big-name vendor like Siemens or Philips--potentially a $50 million splurge--Halamka proposed keeping all the old systems and simply attaching a Web interface so the data could be accessed through a browser. This would not only help reconcile the merged hospitals' two incompatible systems but create an intranet for electronic medical records. Moreover, Halamka proposed plugging all CareGroup facilities into the new network. Most hospital execs would have balked. "In 1998 people thought using the Web for electronic records was crazy. They didn't think it was a stable technology," Halamka explains. But Reinertsen, a big believer in reducing medical errors, gave Halamka the go-ahead. There was a catch: Halamka would have less than $500,000 to spend.

With no cash to hire software jocks from outside, Halamka and his team of three developers started designing the system themselves. Every day, Halamka arrived in his office at 5 a.m. and started writing code; there were nights when he never got home at all, instead curling up for a few hours of sleep on a camping mattress he stored in the office. But he and his coders completed the system, which he dubbed CareWeb, in six months at a cost of $250,000.

Halamka initially kept things simple: CareWeb at first was a modest version of a medical records system--it had no capacity to access radiology scans, for instance. Yet Halamka was still nervous about how doctors and nurses might react, and indeed, when the system was rolled out widely in August 1999, some people hated it. Many older doctors thought computers were inherently intrusive to the doctor-patient relationship. They also feared that the system would suck up valuable time. Richard Parker, a 44-year-old internist at Beth Israel, confesses that many years ago he was one of these technophobic doctors. "I had a computer on my desk, and for years it never got turned on," he recalls. "When I finally turned it on, I had 1,000 e-mails. I was horrified."

Gradually, though, resistance waned as doctors and nurses saw how CareWeb helped avert errors and save time. "Now I don't think anyone would ever want to go back to using paper," Parker says. "I don't see how they could stand it."

Halamka feared that he might encounter greater resistance to his next project--creating a computerized physician order-entry system, known in the trade as a CPOE. Such systems are tougher to build and install, and they involve changing doctors' behavior more dramatically. To avoid calamity, Halamka solicited input before a single line of code was written. On a Monday night in February 2000, he gathered 35 doctors, residents, nurses, and technicians in a spacious 10th-floor boardroom at Beth Israel. "This is going to be the hardest project we've ever executed," he warned them. During the next year, groups of doctors, nurses, and technicians hunkered down several times a month to debate every aspect of medical ordering, no matter how mundane. (One heated discussion: Should orders for the anticoagulant drug heparin be done by patient weight, and if so, in kilograms or pounds?)

When the CPOE went live in June 2001, there was instant grumbling. Under the old system, colored flags were affixed to patient charts to indicate orders awaiting execution. With the new system, some nurses complained that they would have to constantly run to a computer to check for orders. But the greater advantages of the CPOE won out. Nurses no longer had to transcribe orders written by doctors or fax prescriptions to pharmacies. "It streamlined so many processes," says Tricia Bourie, a nurse for 20 years.

By the fall of 2002, the new technology was starting to pay off. A McKinsey study found that CareGroup was saving about $1 million a year and eliminating 90 percent of certain kinds of errors, like a drug being administered in the wrong dosage. At Beth Israel, ER patients were being discharged an average of 30 minutes quicker, thanks to a huge electronic whiteboard in the ER that links to CareWeb and tracks the status of every patient.

It was starting to look as though CareGroup was on its way to becoming a fully wired hospital system. Then the thing Halamka dreaded most happened. On a rainy morning in November 2002, he was alone in his office when he noticed that the network was sluggish. Halamka called one of his network administrators, who told him there had been a surge in activity that would eventually work itself out. Halamka headed off to a meeting, but he couldn't get the problem off his mind. Justine Carr, CareGroup's director of clinical resource management, remembers the meeting. She'd heard rumors of network problems. At one point she glanced over at Halamka: "I saw a little bit of sweat on his brow, and I thought, 'Oh, my God.' John is unflappable, and when I saw him stumble, I thought, 'We're in for it.'"

Later that day Beth Israel's whole system crashed--CareWeb, the underlying applications, the CPOE, and e-mail. It stayed down for three days. At first there was chaos. Everyone had become so dependent on computers that they had forgotten how to do things any other way. Hospital administrators were running around looking for forms nobody used anymore, and everyone was trying to locate a printout of doctors' telephone numbers. The labs at each of the hospitals started dumping thousands of results a day into plastic bins, which were then delivered to doctors by runners who came by every 10 to 15 minutes.

Halamka and some Cisco Systems troubleshooters eventually got the system back up, and CareGroup emerged from the crisis without a major medical mishap. But the outage almost killed Halamka's agenda. At a board meeting several weeks later, one executive suggested that the company had fostered too much reliance on computers. A debate ensued, but in the end, the tech skeptics lost. Halamka pointed out that during the outage an additional 100 full-time people were needed to do work that had been automated. And he was able to convince everyone that the problem wouldn't recur. Since then Halamka has added more features to CareWeb and recently started installing CPOE systems at two other CareGroup hospitals, Mt. Auburn and New England Baptist.

Halamka has taken CareGroup a long way in the five years he's been ministering to its tech systems. Last year the American Hospital Association ranked CareGroup among the most wired health-care organizations in the country. In December, CareGroup announced fiscal-year net income of $13.8 million, its first annual profit in six years; IT initiatives, the company says, have played a part in the turnaround. Still, Halamka is far from satisfied. For one thing, patient records on CareWeb are still incomplete. If a patient has never before been to a CareGroup hospital, its physicians are back in the dark ages, calling up other doctors or relying on what the patient tells them.

Fixing this problem involves another, much more ambitious endeavor: getting hospitals and doctor's offices throughout the region to share electronic medical data. Halamka has been working on this for years, but concerns about patient privacy, among other things, have stalled the project. Next year, however, three hospitals around Boston, one of them Beth Israel, are scheduled to start sharing patient medication records among ER doctors. "You'll come to the emergency room, and with your consent, I'll be able to click a button and retrieve all the medications every doctor in the state has prescribed for you," Halamka says.

Moving beyond medication lists to swapping images like X-rays and EKGs and then looping in doctors outside ERs could take years. But to Halamka, the gains in efficiency are worth all the politicking and grinding IT work that lie ahead. He has scaled back his own hours in the ER to free up time for the task. "This is where I think I can make a big difference right now," he says. "I'm in this for the long haul."

For people like Joe S., that's good news. Back in the ER, the cardiac team has confirmed what Halamka saw so swiftly online: Joe has suffered major cardiac deterioration since the earlier scan; there is a blockage in his left interior descending artery. Within 10 minutes of Joe's arrival at the ER, doctors are administering clot-busting drugs and prying open the artery by emergency angioplasty. Within a day, Joe is back on his feet. At many hospitals, his condition almost certainly would have taken far longer to deal with--time he may not have had. Halamka says his team's ability to get the patient on the operating table so swiftly spared Joe severe heart damage at the least, and may have spared him death. "IT saves money, and that's very important," he says. "But in the end, the lives that it saves is why every single hospital system has to embrace this stuff."