The Stoeckle Center

A more personalized approach to doctor's clinical guidelines can improve patient care while saving money.

Finding the Patient in a Sea of Guidelines

19/May/2011

Blog for The New York Times WELL column; Tara Parker-Pope on Health

Finding the Patient in a Sea of Guidelines

Every Friday morning, the patient, a homeless man in his 60s, lumbered into one of our exam rooms, slipped off the running shoes he wore like bedroom slippers and gingerly lifted his swollen legs so we could remove the medicated bandages and examine the raw wounds on his inner ankles.

Those baseball-size leg ulcers were only one of his many medical problems. He also suffered from obesity and poorly controlled diabetes and blood pressure. Like his other doctors, I encouraged him to lose weight, watch his diet and take his diabetes medications, and stay on his blood pressure pills while making mental note of our medical aims — lower his body mass index to under 25, his hemoglobin A1C to under 7 and his systolic blood pressure to 130.

The reality, however, was that even getting to our clinic was a small triumph for him.

One morning, the clinic nurse approached me after I finished examining him. “Do you know what the problem is?” she whispered, as we watched the man shuffle out to the waiting room. “We keep trying to treat the guidelines rather than him.”

Increasingly over the last two decades, doctors and insurers have turned to evidence-based clinical guidelines to guide treatment. These guidelines both simplify and codify a wide array of health concerns, ranging from blood pressure treatment protocols in diabetic patients (treat with drugs when systolic blood pressure is 140 or higher), to mammography schedules (every other year after age 50), to weight classifications (healthy weight, overweight and obesity I, II and III).

For many, clinical guidelines have become a proxy for quality; physicians who ignore or veer away from them do so at their own legal and financial risk. As a result, a booming cottage industry devoted to the creation of guidelines has emerged. The federal Department of Health and Human Services National Guideline Clearinghouse Web site lists more than 7,000 clinical guidelines created by governmental agencies and professional organizations from all over the world. That number will only continue to rise as politicians and policy makers increasingly focus on studies that cull out new and more cost-effective treatments.

A growing number of doctors and patients, though, have begun to chafe under the absolute one-size-fits-all view of disease inherent to guidelines. Most guidelines assume the presence of only a single condition or risk factor, recommend unwavering cutoff points for treatment and are based on averages derived from large groups of patients.

But real patients are rarely average.

Now a report published this month in The Annals of Internal Medicine has shown that it may be possible to create more personalized guidelines that not only improve care but save money as well.

Researchers from Archimedes Inc., a San Francisco-based health care company, applied complex mathematical models to the electronic medical records of almost 3,000 hypertensive patients who had been followed for over a decade. They found that those patients who had received standard care, following a commonly accepted current guideline, had 40 percent more strokes or heart attacks than those who were cared for according to a more specific, individualized guideline. Moreover, this more personally tailored guideline resulted in a nearly 70 percent greater cost savings.

Previous efforts to personalize and use less general guidelines have failed in part because of the limits of technology and inadequate research methods. In an era of primarily paper-based physician practices, clinicians often must resort to entering patient data and calculating guidelines by hand; more general guidelines are easier to use and remember. For researchers, inadequate mathematical modeling results in a logistical nightmare – the need for hundreds of thousands of patients in order to adequately study the outcomes of myriad combinations and permutations of multiple risk factors.

But by using electronic medical records and integrating them with clinical guidelines, clinicians experience few disruptions in their work-flow patterns. Moreover, with improved mathematical methods now at their fingertips, researchers can build and test guidelines and easily eliminate those that are less promising.

“It’s like designing airplane wings,” said Dr. David M. Eddy, lead author of the report and founder of Archimedes. “Boeing can’t build a hundred models and test them all. They have to design airplane wings by mathematical models.”

Dr. Eddy and his colleagues are currently studying an even more intricate set of guidelines that incorporates more than 30 patient-specific risk factors like cholesterol levels, weight, blood sugar levels, aspirin use and cancer status. Based on mathematical calculations of those risk factors, each patient receives individualized graphs and bar charts of the “benefit score” for every recommended intervention. Doctors and patients can review these results and decide together how to prioritize the interventions. “This kind of individualized guideline wouldn’t have been possible as little as five or 10 years ago,” Dr. Eddy said. “It’s just taken a while to bring powerful mathematical modeling to clinical guidelines.”

While it will most likely take a few more years before such individualized guidelines are widely available, Dr. Eddy believes they will one day become an integral part of health care. “When I was younger, banks would close at 3 in the afternoon, so clerks could count their ledgers. That’s unheard of now.”

He added: “Now we have the tools to move to the next generation of guidelines. It’s the way of medicine in the future.”