|
||||||||
|
HEALTH CARE MANAGEMENT Researchers Argue that Fluctuation in Care Delivery Wastes Hospital ResourcesTheir slogan could be, It's variability, stupid! In a paper that analyzes the main source of inefficiency in hospital operations, two Harvard researchers are challenging health care leaders to adopt a fresh approach to the ongoing debate about cost and quality in their field—and to reap huge savings in the process. Excessive variation, compounded by a lack of understanding process flows in a complex system, are the basis of waste and probably errors in medicine, according to Eugene Litvak, adjunct lecturer in the Department of Health Policy and Management at HSPH, and Michael Long, assistant professor of anesthesiology and deputy director of operating room services at Massachusetts General Hospital. They present their view in the March 22 American Journal of Managed Care. The authors have developed methods that enable hospital executives to identify the different types of variability and to quantify how each affects cost and quality at their institution. The authors have applied tools from the field of operations research to specific examples of poor hospital management, such as low operating room utilization, and have co-developed user-friendly software to fix the problems.
Mike Long (right), a senior anesthesiologist at MGH, recalls that when he met Eugene Litvak (left), an operations researcher from HSPH, five years ago, he "felt like I had finally met my long-lost twin. Eugene knew the answers to questions about inefficiencies that had been sitting around in my brain for years." Long contributed his knowledge about clinical medicine to an ensuing collaboration that produced a new approach to cutting health care cost while simultaneously increasing its quality. "I think this work is of profound significance to health care. It is extraordinarily important that we attempt to apply it," comments Don Berwick, president and CEO of Boston's Institute for Health Care Improvement. Berwick is also HSPH associate professor of health policy and management and HMS clinical professor of pediatrics and health care policy at Children's Hospital. A Game of Catch-UpAll major industries except health care have during the past 30 years perfected techniques to better manage variability, says William Pierskalla, a leader in operations research, Litvak's area of expertise."In that sense, this is a matter of technology transfer," says Pierskalla, former dean of the Anderson Graduate School of Management at the University of California, Los Angeles. Previous attempts to transfer such techniques to health care largely failed, in part, because apparently unlimited cash flow allowed the field to focus exclusively on improving quality while ignoring cost. Perhaps more importantly, operations research demands that decisions be made with consideration to how they ripple through all interconnected units of a complex system, such as a hospital. By contrast, "the nature of decision-making in health care has been more like a mom-and-pop shop. Doctors work independently, and looking to reduce variation is not part of their thinking process. They do not know about these technologies," says Pierskalla. The time for operations research in health care may have come, however, as managed care imposes new cost constraints and problems like waste and even medical errors are increasingly viewed as system breakdowns that occur despite individual efforts. Even so, most health policy research still treats cost and quality as separate subjects. "You need a general approach that ties the two together. Variability analysis can do that," says Litvak. The Hidden WasteSo where is this waste in medicine?Much of it is hidden in the cost of delivery. Suppose you want to order lunch. Pizza A and pizza B both taste good. Shop A charges $7 for the pizza and $1 for delivery. Shop B asks $6 for the pizza and $3 for delivery. Naturally, you choose A for its lower overall price. In medicine, delivery costs depend on the efficiency of process flows at provider institutions. This is difficult to measure, in part, because few people study this area. Most research focuses on the cost of the clinical procedures themselves. HMOs generally have no handle on runaway delivery costs either. Instead, they control overall cost through capitation and overseeing physicians' clinical decisions. That is like pizzeria owner B ordering the cook to scrimp on pepperoni because he does not know how to cut delivery costs, say the authors. Providers and HMOs currently talk past each other. Cost constraints have compromised quality and will do so further, but at the same time, the HMOs' charges about waste are also true. "I think there is so much waste in the system that it is difficult to discuss it," Berwick says. Cutting certain kinds of cost will actually increase quality, Litvak and Long argue. Using variability analysis to eliminate management waste would bring providers to a point where they could make an evidence-based case against further cost cutting. In turn, HMOs could give employers precise options for how much quality they want to buy. What, then, is variability analysis? It classifies the different kinds of variability into "clinical" (patients have different diseases, treatment responses), "flow"(they arrive randomly, few on some days, many on others), and "professional" (providers differ in their skills, teaching takes time). Then it distinguishes between "natural" variabilities that are inherent to health care, and "artificial" ones, which arise from poor management. Consider operating rooms (ORs) in a city hospital. They routinely juggle competing demands for the same time slots from two directions: on one hand, urgent and emergent cases appear randomly at the hospital's door. (Urgent cases, like trauma, demand immediate attention while emergent cases arise and worsen within hours, such as an inflamed appendix.) On the other hand are scheduled cases. Both categories include different specialties and case severities. The question is how to allocate resources most efficiently. The variabilities in urgent and emergent cases are natural and cannot be eliminated. Yet to make their management optimal, operations research offers the techniques of queuing theory, which matches random demand to the fixed supply of ORs. Based on an analysis of MGH's OR operations, the authors worked out a redesign plan in 1996 that the hospital since has begun to implement. Specifically, MGH has allocated special ORs to urgent and emergent procedures in orthopedic and general surgery to ensure that these patients need not wait long and that they do not disrupt the schedule set for elective surgeries. To work out decisions such as these, Litvak has written the algorithms for software that calculates the critical factor for urgent cases—their waiting time—for a given statistical patient flow and other variables. It allows hospital executives to play what-if scenarios, calculating how many ORs they need at standby if patients were to wait, say 10 minutes, or an hour. This tool makes possible rational cost–quality decisions that work for a given hospital's patient population, specializations, financial resources, and other constraints. Unnecessary Ups and DownsThe critical factor for elective procedures is OR utilization, which should be high and smooth. But in analyzing such data, Litvak and Long made an astounding discovery. Although elective procedures generally contribute substantially less to the total number of beds used in a hospital on a given day than does the emergency room, both frequently add equal amounts of variability to the hospital's census. "In essence, we are as likely to predict when someone will break their leg as when a surgeon will do a scheduled surgery," says Litvak.Long says that surgery scheduling is sometimes dysfunctional because individual surgical services have control over certain blocks of time. When his time goes unused, a surgeon produces a dip in the number of cases the OR handles one day and an uptick on other days when he catches up. Many surgeons doing this simultaneously can produce big troughs and peaks in OR utilization. "We experience this on a routine basis," says Long. "This is artificial variability, the worst kind," says Litvak. "Queuing theory does not work for it. It must be eliminated by conventional management reforms." It is expensive, since ORs—which cost roughly $400,000 apiece per year—lie fallow some of the time. But worse, it adds variability to the entire hospital by feeding more cases to the floors on some days than others. In recent years, many hospitals have attracted more patients, often operating at 70 to 80 percent capacity. Natural flow variability alone—around 30 percent—can bring the hospital to its limit, even without the artificial variability caused by inefficient scheduling. This is where Litvak and Long believe that excessive variability can set caregivers up to make mistakes. Operations theory describes how miniqueues—to admitting, to X-ray, to MRI—can add up in an overloaded hospital, eventually turning otherwise inconsequential errors, such as a misplaced file, into system damage, which manifests itself in longer stays or even tragic medical errors. "We suspect that management may have as much to do with lowering the medical error rate as physician certification and technological fixes like computer entry of pharmacy orders," says Long. Block booking remains in place at most large hospitals. The changes MGH has made in OR allocation are working, the authors say. Yet as long as there is no research in this area, hospital leaders cannot draw on a body of knowledge to convince medical staff that this approach will benefit all. The authors worry that some leaders at hospitals and HMOs agree with their approach but take no action. "This is what I call cutting sour cream, when I have no resistance and no slice," says Litvak. "Everyone knows Litvak and Long are right," Berwick says. "If we are able to engage the issue they are challenging us to, then the potential for creating lots of winners is huge because we will be able to mobilize resources that can be reinvested where they are needed." —Gabrielle Strobel |
