By Christopher Guadagnino, Ph.D.

Published April 1998

• React to this article in the Discussion Forum.

Are these guidelines merely attempts to legitimate cost cutting measures or are they attempts to improve outcomes and reduce costs? Are they unjustified assaults on the professional prerogatives of physicians, or do they provide support to bolster medical practice? Is physician resistance based on professional integrity or professional arrogance?

However you come down on these questions, the number of guidelines, the number of institutions applying them and the measures taken to support compliance are unquestionably growing.

Guideline Rationale

The majority of information on which physicians base their clinical decision-making is of poor quality, with only 20 percent of decisions based on scientifically-derived studies, according to David B. Nash, M.D., M.B.A., associate dean and director of health policy and clinical outcomes at Thomas Jefferson University Hospital.

The principal goal of practice guidelines is to standardize the medical profession’s approach to care by limiting unexplained clinical variation, says Nash, and not principally to cut costs. Successful guidelines, says Nash, would reduce to 30 percent the variation in physician clinical practice patterns for cases that are uncomplicated and identical; the amount of variation currently in the system nationwide across many different diagnoses is 200 to 300 percent for some procedures. The 30 percent variation goal, adds Nash, would permit physician flexibility and account for complicated patients who do not respond to a guideline’s therapy and fall outside its parameters.

Guidelines could speed up treatment regimens by flagging therapies at appropriate intervals, such as ordering rehabilitation on a Friday rather than on a following Monday. Guidelines could include formularies that rein in costs attributed to varying prescription habits across physicians at an institution for the same condition, such as brands of medication or prostheses.

A standardized approach to medical care allows physicians to track what was done for a given illness and locate areas of improvement by observing how deviation from standardized treatment affects outcomes, says Richard L. Simmons, M.D., medical director of the University of Pittsburgh Medical Center Health System and head of Pitt’s Department of Surgery.

Sources of Guidelines

Clinical practice guidelines are generated by a variety of sources, including the Institute of Medicine, HCFA, hospital associations, the Joint Commission on Accreditation of Healthcare Organizations, physician specialty societies and a variety of private sector sources, notes Nash. Some 1500-2000 widely available practice guidelines exist currently, from at least five or six dozen agencies, says Nash.

Some of the guidelines that are disseminated by these agencies are based on statistical data derived from evidence-based studies of specific disease and treatment processes, while most are generated by consensus panels of physician experts who conduct literature reviews of clinical studies. The quality of the guidelines is not uniform because of the heterogeneity of their sources, acknowledges Nash, and even those developed by medical specialty societies are difficult to assess because of the closed-door nature of the process. Selection of consensus panel members is also a source of acrimony among physicians in a given specialty. As a result, such guidelines are not widely accepted and implemented.

The University of Pennsylvania Health System currently implements 30 critical pathways at its four hospitals for conditions such as hysterectomy, asthma, bone marrow transplant, carotid endarterectomy, colectomy, craniotomy, cystic fibrosis, GI bleed, hip and knee replacement, low birth weight infants, pneumonia and renal and liver transplants, says David J. Shulkin, M.D., the system’s chief medical officer and chief quality officer. With over 275 of Penn’s physicians currently involved in guideline implementation, the guidelines indicate specific roles for each person involved in a patient’s care and "variance collection" is performed, making a record of what was or was not done vis-a-vis the guidelines, notes Shulkin.

Penn’s outpatient guidelines are implemented with the help of Intranet computer systems installed in 93 primary care physician offices, assigning disease-specific case managers and support personnel, according to Shulkin.

Shulkin notes that there is no set way the guidelines are developed; some involve evidence-based studies and literature reviews, but, he says, in many cases no scientifically uniform studies exist upon which to base a pathway. The Penn system does not rely on a consensus process for protocol development, but on the recognized local expertise of its physician leaders, Shulkin adds.

Physician Resistance

Guidelines have not been successful in achieving a high degree of physician compliance, Nash indicates, because more energy has been spent on their construction and dissemination than on implementation and efficacy. Shulkin believes that compliance is low because most physicians believe that they are using an optimal approach to practice and feel little imperative to change.

Cultural obstacles exist whereby physicians resist perceived pressure to conform to guidelines as professionals who object to encroachments on their autonomy. Many may view guidelines as cookbook medicine designed to cut costs. Some guidelines, Nash concedes, may preclude the need for physician participation in certain aspects of care, for example, administration of therapies.

Physicians should also be wary of the dangers of overcompliance to guidelines, applying them inappropriately to patients with co-morbidities when they were designed for non-complex cases, says Joseph Giordano, M.D., past president of the Mercer County Medical Society. Guidelines may offer physicians a more systematic approach to treat infrequently encountered illnesses, but could lead to a shirking of responsibility for decision-making, Giordano believes, especially in an environment where their use is encouraged and deviation is sanctioned.

Another problem, notes Giordano, is the inability to track patients after a guideline-prescribed early release from the hospital. As an example, he notes, Medicare guidelines say it’s okay to release from the hospital a patient with a temperature under 101 degrees. That patient may be readmitted earlier and more often, despite being inaccurately tracked as an optimal outcome.

A minority of patients may actually do worse if treated according to guidelines. For example, 10 percent of those given oral antibiotics at home may have done better with a more intensive treatment, says Giordano. He cites a 1996 study published in the New England Journal of Medicine in which a protocol for managing acute stroke got patients out of the hospital faster, without adverse outcomes, but 6 percent had hemorrhaging—representing a ten-fold increase over patients not treated by the protocol.

Another potential danger exists if smaller hospitals adopt protocols developed for larger institutions with residents, neurosurgeons and more expensive technology, warns Giordano.

Pressures by insurance companies to induce compliance to their own utilization criteria may inappropriately affect the way critical pathways are drawn up, putting a greater emphasis on cost control than would otherwise be justified. Providers will be responsible for their own protocols when they engage in risk contracts with managed care companies.

Guidelines can be more easily implemented than evaluated, believes Wanda Young, Ph.D., president of the International Center for Health Services Research and former president of the Pittsburgh Research Institute and corporate vice president of Blue Cross of Western Pennsylvania. Many hospitals are not well-equipped to evaluate practice guidelines, a task that requires linking administrative and clinical data that is often coded differently, as well as utilizing advanced information systems to track and synthesize data, notes Young.

Valid evaluation of physician compliance with guidelines should be systematized instead of done on a case-by-case basis as is typically done, she believes, because patient-specific comparisons are often made that are not clinically comparable. Costly hardware and personnel are required for valid data comparison, and misuse of data must be avoided, she warns.

Compliance Regimes

Efforts to boost compliance are focused primarily on providing periodic comparative performance feedback to physicians, comparing their adherence to guidelines with those of their colleagues. At Penn, physicians are given patient-specific feedback reports as frequently as every week, says Shulkin. Penn’s "capture rate," or the percentage of patients eligible for protocol use that is actually enrolled and followed according to protocol, varies from 96 percent compliant for some services to less than 25 percent compliant, says Shulkin.

Penn’s five-part strategy for changing physician compliance behavior includes the use of education, clinical opinion leaders, feedback of data, administrative interventions (more paperwork for a physician who deviates from protocols) and incentives (financial and nonfinancial), says Shulkin, who adds that physicians should not be penalized for deviating from protocols, but that deviation should be a deliberate decision if and when it is made.

At some institutions, monthly medical staff meetings are held to compare protocol compliance. If outlier physicians do not offer explanations for their deviation from the practice patterns of their colleagues, some physician groups are applying economic sanctions and hospitals are looking into revoking privileges, according to one medical director at a hospital in Pennsylvania.

Of course, slavish adherence to the protocol could make a physician an outlier as well, the medical director notes, adding that physicians still control the process at this point "and can bury a hospital that applies protocols in a heavy-handed manner."

UPMC’s Simmons acknowledges that sanctions are difficult to apply to physicians whose outcomes and practice patterns deviate from those of their colleagues. The UPMC system engages in "report-carding," says Simmons, whereby physicians who are not doing well with respect to their peers are advised on how to make changes. "You deal with deviation by instruction, by data, by convincing people that this is now the way to do it, that there’s a better way to do it," says Simmons, who cautions that the goal is not for the standard to be blindly applied, but for the outcome to be optimal.

For Simmons, optimal outcome includes a consideration of cost. As for equal outcomes with one approach costing more than the other, "That’s an outcome that would not be optimized under those circumstances," he says. Physicians providing the more expensive treatment would have to convince their colleagues that their approach is justified.

Perhaps the most fascinating outcome of the increasing usage of practice guidelines, protocols, formularies and critical pathways is the re-definition of medical quality. Nash defines quality as the best possible outcome at the best possible price, and adds that research bears out the assumption that high-quality medical care costs less, by definition. The assumption is premised on the industrial maxim that if you do something right the first time, you will improve quality and lower costs.

Penn uses what Shulkin calls a balance scorecard approach to defining quality and evaluating its critical pathways. The approach merges four factors into a single index: service (patient satisfaction measures), clinical quality (mortality, morbidity and infection measures), access (ease of getting care) and value (efficiency and cost measures). Shulkin says such a definition of quality accounts for the non-algorithmic components of the practice of medicine and preserves the physician-patient relationship.