Infection fighters in hospitals start with one hand tied behind their back. They're always on the lookout for patterns of laboratory results and clusters of patients with the same microbial affliction, but the possibilities are so vast that they're reduced to targeting only certain areas of the hospital that are most likely to harbor and spread infections.
This targeted approach, sanctioned and recommended for practical reasons by the Centers for Disease Control and Prevention, concedes that some hospital-acquired infections-and their preventable causes-will go unnoticed until they've had a chance to spread unhindered and become an obvious threat, infection-control professionals say. Some subtle outbreaks just fester and run their courses without ever being recognized.
But a new class of analytical software is moving in with a breakthrough remedy: the ability to track every lab test from every part of a hospital; merge the findings with details of how and where a patient is being treated; and alert infection-control professionals to likely problems without the need for labor-intensive detective work.
"It picks up things you wouldn't normally notice," says Lance Peterson, director of microbiology and infectious disease research at Evanston (Ill.) Northwestern Healthcare. "It goes through so much data that you can't review manually."
In addition to broadening the field of vision for vigilance, the software's ability to identify nearly all hospital-acquired infections and calculate the costs of treating them is making a business case, and a big one, for committing the resources necessary to reduce the incidence of infection in hospitals.
At the three hospitals in the Evanston Northwestern system, six months of surveillance by a software system called MedMined detected a 5.8% infection rate and $5 million in unreimbursed costs of treating hospital-acquired infections, Peterson says. During the next six months, the hospital system reduced infections by 223 cases, cutting the infection rate to 5% in a program of infection control that included responding to the alerts fed to the system by analysts at the Birmingham, Ala.-based software company.
Besides a better track record for patient outcomes, the hospital system saved a total of $618,000 in avoided expenses-$3 for every dollar it spent on infection surveillance, Peterson says. That's a welcome bonus for the bottom line when other pressures are threatening hospitals' financial health, he says. "In Illinois, medical malpractice (insurance expense) is such a big problem that a lot of hospitals are teetering between the red and the black. So a half-million dollars in six months is a lot."
At Florida Hospital, a health system with seven hospital campuses in the Orlando area, infection-control professionals were able to identify 956 cases of hospital-acquired infections in 2003, which added $10 million in costs to treat affected patients, says Bill Wing, senior finance officer and senior vice president of performance measurement.
Those numbers have put infection control on the list of top budget priorities as talks begin for 2005, Wing says. "Being able to identify (all affected) patients and begin to create strategies to prevent infections is key," he says. "The real question is how far can you reduce your infection rate, and that's what we hope to find out during the next few years."
The Florida system already achieved results by focusing early on one type of problem-bloodstream infections in adult patients, often transmitted though catheters. That source of infection alone accounted for 382 hospital-acquired infections in 2002. In 2003, the incidence of that type of infection was reduced by 31%, to 263 cases.
The bloodstream infection incidence was a small fraction of the 103,000 admissions recorded at all seven facilities in 2003, Wing says, but the impact was considerable: $3.9 million in added costs. That total, however, was $1.8 million less than in 2002, down 32%, demonstrating the quick return on computer-assisted surveillance, he says.
When patients contract infections while hospitalized, "they cost more, they stay longer and you don't have the offsetting revenue gains," says G.T. LaBorde, co-founder and chief operating officer of MedMined. "You literally turn a profitable patient into an unprofitable patient."
Completing the cost equation
Until recently, the healthcare industry had only a vague idea of what hospital-acquired infections cost, based on anecdotal evidence. Then last year a study published in the journal Clinical Infectious Diseases highlighted the extra expense and hospital days that could be saved by preventing such infections. The study calculated that the infections added an average of 10 days and $15,275 to the hospital care of each affected patient in the intensive-care unit of Cook County Hospital in Chicago during 1998.
MedMined did its own analysis recently using comprehensive hospitalwide data gathered from 50 hospital clients, LaBorde says. Its results: Patients with infections stayed an average of eight extra days at $14,000 in additional cost.
But that per-patient impact, though helpful in quantifying the financial consequences, is only half the equation for establishing the total hit to hospitals. Because facilities generally target only selected classes of infections in certain high-intensity areas, there's no way they can capture the total number of cases in a hospital and multiply them by the per-patient cost to come up with a definitive dollar impact, experts say.
Before launching its computerized-tracking system in September 2002, Evanston Northwestern took "the classic current approach to infection control, which is targeted surveillance," Peterson says. The hospitals "were really not focused for surveillance purposes on the normal nursing units."
Evanston Northwestern and other hospitals using MedMined set up their hospitalwide data on lab and microbiology results to be fed to the IT company. Added to those details are particulars of each patient's movement in the hospital, including admission, length of stay and transfers to different departments and nursing floors-the better to trace infection patterns.
A sophisticated computer process sifts through thousands of clinical details to check for predetermined problems. It also picks up bacterial patterns on its own that just don't seem right. The findings are fed back to hospitals on a monthly schedule, though evidence of serious problems can be flagged quickly, LaBorde says. Then infection-control practitioners take the leads and check them out.
"It's the things you would do if you could see the forest for the trees," Peterson says. "But the forest is so big you can't notice it unless you have something like this to help you." It would take five full-time people reviewing charts to find the hospital-acquired infections that are automatically flagged by the computer system, he says.
Isolating infection outbreaks
Although a big attraction of such information systems is their ability to do a wider sweep of surveillance than was possible before, they also can pinpoint the source of problems down to a single cluster of rooms.
For example, Florida Hospital picked up the strong presence of a type of mold that causes lung problems, especially in patients with compromised immune systems, says Christine Kapur, director of infection control. That problem, detected in a single department of one hospital, previously would not have been identified, Kapur says. It eventually was traced to a filter in a ventilation duct that wasn't properly seated.
Infection-control software also flags problems stemming from outside the hospital. At Great Plains Regional Medical Center, North Platte, Neb., early recognition of an influenza outbreak through patterns a computer detected in the emergency department alerted the ER to gear up for it, breaking out masks and gowns and cutting back on visitor access, says Teresa Nowak, infection-control coordinator at the 122-bed facility.
Computers at the Evanston system identified patterns in a salmonella outbreak that helped public health officials trace it to a restaurant near Highland Park (Ill.) Hospital, one of its facilities, Peterson says.
Pinpoint accuracy also is a plus on the business side, where the computerized tracking can differentiate the cost of each type of infection to better calculate the total expense of a particular problem.
The 263 bloodstream infections at Florida Hospital in 2003 were found to add $15,000 per case, for a total of $3.9 million. Urinary-tract infections, by contrast, cost an additional $7,000 per case. But the problem was equally high in priority from a cost standpoint because the tracking software was able to identify 548 urinary-tract cases in 2003, each adding an average of 9.2 days to a patient stay.
The upshot financially: a total infection-related expense of $3.8 million, plus lost revenue from hundreds of extra hospital days that took away beds for additional admissions-a problem on the same level as bloodstream infections.
Meanwhile, the cost of hooking up the MedMined data feeds and installing the system ranges from $10,000 to $50,000 depending on the size of the hospital organization, LaBorde says. In addition, MedMined charges a fee ranging from $5,000 to $25,000 per month.
For infection-control pros, the service saves hours of time previously spent flipping through lab results and other reports just to find a problem to solve, says Jayne Bassler, Florida Hospital's assistant vice president of clinical-performance improvement. That frees them up to "stop sitting at a desk looking at paper and (instead) spend time with physicians," she says-time spent preventing future infections by educating clinicians on proper hygiene and adherence to procedures that might not have been followed.
At Great Plains, an automated surveillance system from TheraDoc, Salt Lake City, has shaved six to 10 hours per week from the preliminary investigation process, Nowak says. The system assembles findings taken during the previous 24 hours or the past week, with patient census and transfer data added, and presents them to Nowak at the beginning of the day. That capacity "has allowed me to significantly control and reduce infections," she says. "In the past I have had to manually analyze information a month or two behind."
By issuing a list of leads to infection-control professionals, "We help them move from a hunter and gatherer to an interventionist," says Stanley Pestotnik, TheraDoc's co-founder and president.
What do you think?
Write us with your comments. Via e-mail, it's [email protected]; by fax, 312-280-3183.