Study of How Doctors Treat And Diagnose Shows Cultural Values, "Common Sense" Play Role In Decisions

BUFFALO, N.Y. -- A University at Buffalo researcher's study of how physicians work has found that current models of expert thinking do not adequately explain the complex factors involved in how physicians diagnose and treat patients.

The study, which challenges current models of expertise accepted by cognitive psychologists and artificial-intelligence researchers, has implications for management and staffing decisions for all occupations, its authors say.

The findings also imply that proposed changes in the health-care system, such as managed care and increased patient loads for physicians, will likely increase errors.

"We often study workplace design and staffing as though technical expertise were the only issue," said Valerie Shalin, Ph.D., UB assistant professor of industrial engineering and co-author of the study. "Our research shows that at the very least, effective workplace performance is a combination of technical expertise and other factors, such as this other not-so-common 'common sense.' We need to broaden our understanding of the expertise that a competent physician has."

Shalin added: "There is much more cognitive work going on in a doctor's mind than you could ever begin to imagine. We need to develop an appreciation of this as we increase workloads on people and we certainly don't want to wait until the system breaks down to discover the true cognitive demand on physicians.

"Given the current lack of scientific understanding about medical expertise, which our study underscores, it disturbs me that anyone is being asked to formulate broad policies about physician decisions or whether a medical procedure is warranted or not."

The paper, "Functions of Expertise in a Medical Intensive Care Unit," will be published in a forthcoming issue of the Journal of Experimental and Theoretical Artificial Intelligence.

Shalin, who is affiliated with UB's Cognitive Science Center and has studied cognitive performance among workers including airline pilots, co-authored the study with Dennis Bertram, M.D., of the HMO Group, a Buffalo research organization that evaluates medical treatments and technologies.

The study was based on their viewing of 21 1/2 hours of videotapes of physicians working during a five-day period in an intensive care unit in a teaching hospital in Buffalo.

According to the authors, many attempts to characterize physicians' cognitive expertise have focused primarily on whether or not a physician has come up with a correct diagnosis.

Their study shows that physicians' decisions about diagnosis and treatment are influenced by multiple factors, including cultural assumptions about the quality of life a patient should have after treatment and even "common sense."

"This research is evidence of the broad range of goals that have to be simultaneously considered in treating a patient," said Shalin.

She explained that current models of physician expertise do not take into account how a doctor bases decisions on "common-sense" knowledge and cultural values.

According to Shalin, the findings demonstrate that because medical reasoning involves far more complicated cognitive processes than were previously understood, recent efforts to increase physician workload are likely to result in more errors on the part of medical professionals.

"As we compress the amount of time an individual has to deal with a patient, we're obviously squeezing those complex reasoning processes into shorter periods of time," she said. "It's a well-established scientific fact that when the speed of reasoning goes up, so do the number of errors."

At the same time, she said, existing research shows how well physicians are able to process information. In the overwhelming majority of cases, mistakes do not occur and the complex cognitive processes that physicians use to diagnose and treat patients have the appropriate outcome.

The videotapes cover some 20 patient cases and five physicians: three were residents, one was pursuing advanced training in a specialty and one managed the unit while training the other physicians.

Shalin said that the videotapes, which include detailed conversations between physicians, allowed the researchers a rare opportunity to observe how a particular conclusion about a patient is reached.

A particularly important component of that expertise is the ability to integrate technical knowledge with '"common sense" knowledge, which, she noted, is not all that "common."

"That kind of 'common sense' is what tells a physician when to make a certain tradeoff, such as leaving the bedside of one patient to visit another patient whose condition is unstable," she said. "That's what is learned in an apprenticeship, like a residency."

Shalin noted that previous attempts to define physician expertise have not taken into account cultural expectations about cost/benefit analyses that are a basis for certain medical decisions.

For example, the paper cited the example of hip dysplasia, a surgical procedure that usually is prescribed because it eliminates a permanent limp, and therefore the surgery has significance in Western culture. It noted, however, the case of a Navajo who underwent the surgery and found afterward that he could no longer participate in important activities in his culture that involved sitting on the floor.

"We are not suggesting that the surgery is inherently bad," the researchers wrote, "but that its value is not a purely medical issue."

Shalin said that her conclusions, particularly regarding cultural expectations, are relevant to workplace issues in other occupations as well.

"Workers make value judgments all the time in all kinds of jobs," said Shalin. "We have a weak foundation for workplace design and staffing decisions when worker functions are evaluated as though expertise in any field were simply a technical matter. "