From the Bedside to the Desktop
How Risk and Information Technology Are Remaking Medicine
Eighteen years ago, when I was a first-year internal medicine resident, one of the first patients in my outpatient practice was a 77-year-old man with emphysema. My exam discovered a blood pressure of 180/70, the top number well above the “normal” 120’s. My preceptor asked me what I should do about the pressure of 180.
“Nothing,” I answered.
He nodded in agreement.
At the time, blood pressure of 180 was the body’s physiologic compensation to assure a steady flow of blood through the old man’s hardened arteries. It was normal aging.
One year later, however, this elevated pressure became a disease called isolated systolic hypertension—and I was prescribing blood pressure medications to patients like him. Within four years, when I was a fellow in geriatric medicine, two-thirds of my patients were on these medicines. What had been a routine part of aging had become a minor pandemic.
I began treating isolated systolic hypertension in older adults because a study showed that treatment reduced the chance, or probability, of suffering a stroke or congestive heart failure. This study was the last step in a new approach to discovering disease. First, large longitudinal studies identify factors that predict risk. Next, a clinical trial shows whether changing a factor, such as blood pressure, reduces the risk. As a result, what had been just natural variations in human physiology became diseases.
In the years to follow my residency, I felt like I was becoming a small-time insurance salesman. I was ordering a host of tests, such as bone scans and blood cholesterol levels, to decide whether a patient was at enough risk to warrant still more tests and drugs. I was experiencing the transformation of medicine from the care of the sick patient at the bedside to the management of the at-risk patient at the desktop.
And I was ill-prepared to practice this new kind of medicine. I wrote my notes in longhand in a paper chart. I had no training in probability theory and statistics. My last math class was high school calculus. I had not a clue about the psychology of risk.
For much of the modern era, medicine worked at the bedside. “Go to the bedside” was a senior physician’s command to my colleagues and me, a command indicating where to discover, diagnose, and treat disease. The patient’s chief complaint and the detailed history and physical that followed were the foundation of the medical encounter.
But today, medicine now occupies a new space. Physicians discover diseases and diagnose and treat patients at the “desktop.” In a recent issue of the Journal of the American Medical Association, I presented this new model of medicine.
Desktop medicine describes how risk assessment and information technologies are transforming medicine. The desktop with a networked computer—and that computer with its own virtual desktop as well—are where researchers examine large databases to discover risks and where clinicians and patients meet to assess a patient’s risk factors and decide whether the patient needs treatment.
A key actor in these events is technology. An online search for “health risk calculators” reveals the fruits of this technology. Pages of user-friendly websites where physicians and patients can input information, click “submit,” and receive immediate risk results: your 10-year risks of bone fracture, heart attack, or death from colon cancer to name just a few. In developed nations, desktop diseases such as hypertension, osteoporosis, and dyslipemia are prevalent; occupy a substantial portion of a physician’s practice; and are among the leading causes of morbidity and mortality.
And researches are discovering more of these diseases. Among the headline stories this past summer was the announcement from an international conference that Alzheimer’s disease was being redefined from a bedside disease characterized by disabling declines in cognition to an abnormal biomarker signifying the risk of future cognitive decline.
Desktop medicine relies on the mathematical sciences such as statistics and epidemiology, and psychology’s discoveries about how people perceive risk and discount the future. Behavioral economics, a hybrid of psychology and economics, shows that small and periodic monetary incentives can address one of the most vexing problems in desktop medicine—a lack of adherence to long-term risk reduction, such as taking a daily medication.
Unfortunately, medical training and the health care system have not caught up to desktop medicine. Though much has changed in the 18 years since I sent the elderly men with elevated blood pressure home, we still deliver an education suitable for an aspiring bedside physician circa 1975. Premedical requirements have no expectation of proficiency in the basic sciences of risk such as statistics, psychology, and economics.
The board exams that medical students must pass to enroll in a U.S. residency program scarcely address desktop sciences and practice. Expertise in diagnosing and treating anxiety and worry—the principal symptoms of desktop diseases—is relegated to underpaid mental health counselors and the knee-jerk prescription of antidepressants. Many of the problems in “health care reform” come from trying to jerry-rig a bedside health care system to care for patients with desktop diseases.
The result is that it is still possible to practice medicine with little expertise in how to interpret the complex multivariate models that define disease, how to talk to a patient about risk and its management, and how to set up a clinical practice that wisely uses information technology and incentives to diagnose and treat patients with desktop diseases. Physicians often blindly act upon the results of online risk calculators with little attention to the validity or logic of their results, even if those results compel a potentially wasteful test, or yet another prescription.
Just as doctors practicing bedside medicine need a certified, hospital-based laboratory in order to scientifically diagnose, treat, and discover diseases, doctors practicing desktop medicine need a national electronic medical record that functions as a database to track trends in risks and the successes and failures of interventions, and to identify new risks. The FDA needs to consider the value of public review and approval of the technologies that calculate these risks as many of them are owned by private and for-profit interests.
The bottom line is we need to be training our future physicians in the judicious use of information technology and analysis of risk and redesigning our health care system in order to fulfill the promise of a healthier society offered by desktop medicine. Failing this, we should not be surprised if the numbers of students training in primary care, the field at the frontline of desktop medicine, continue their present nose dive. We should also not be surprised to find ourselves less masters of our own technology than mastered by it and its owners.
Jason Karlawish is an associate professor of medicine and medical ethics at the University of Pennsylvania.
Comments on this article