Prognostic Models Abound, but How Useful Are They?

The Hippocratic physicians of ancient Greece prized the skill of prognostication above all others. As cited by medical historian Roy Porter in The Greatest Benefit to Mankind, they deemed it “a most excellent thing for the physician to cultivate Prognosis; for by foreseeing and foretelling … the present, the past, and the future, and explaining the omissions which patients have been guilty of, he will be the more readily believed to be acquainted with the circumstances of the sick, so that men will have confidence to intrust themselves to such a physician.”

Some two millennia later, after a relatively brief period of decline, prognosis is enjoying a resurgence of interest. The tools wielded by today's practitioners, mathematical predictive models, would surely have baffled Hippocrates—but perhaps no more than they confound many contemporary physicians.

Nicholas A. Christakis, MD, PhD, MPH, formerly of the University of Chicago and now in the Department of Health Care Policy, Harvard Medical School, Boston, Massachusetts, traces what he calls “the ellipsis of prognosis” in modern medicine in his 1999 book Death Foretold: Prophecy and Prognosis in Medical Care. Christakis contends that prognosis played a central role in medical practice throughout most of history but waned in prominence as effective treatments emerged. He illustrates his argument by analyzing the entries for lobar pneumonia in selected editions of Osler's Principles and Practice of Medicine from 1892 to 1988, showing how the space devoted to prognosis dwindled as antibiotic therapy became available in the mid-20th century.

Throughout history, prognosis has almost exclusively assessed the risk for death from a disease rather than the course of the illness or any residual morbidity, Christakis writes. Before the advent of effective treatments for most diseases, prognosis was viewed as being specific to the individual. Patients with unhealthy “habits of life,” …