Just a few years ago, if I had been asked to comment on variation in healthcare, I would have said it needed a fundraising event for awareness, or even a respected celebrity patron—maybe Sandra Bullock decrying unnecessary variation on Oprah. Fortunately, more socially influential forces evolved. In a relatively short (from a cultural perspective) span of time, variation has emerged to become standard water-cooler talk amongst physicians and politicians alike.
Although the first analysis of medical variation surfaced in 1938, it wasn’t until Wennberg and Gittelsohn’s seminal paper that our collective medical consciousness emerged.1 Wennberg noted that if his children had simply gone to school in the neighboring district of Stowe, Vt., they would have had a 70% chance of having a tonsillectomy, as opposed to a 20% chance in their chosen district of Waterbury. Decades later, that work is the foundation of the Dartmouth Atlas project, which has turned its lenses toward unexplained variation in the costs of healthcare.
Meanwhile, in a parallel—nonmedical—universe, two engineer-statisticians were busy refining quality-control theory in the late 1930s. Shewhart developed the PDCA (plan-do-check-act) cycle, and Deming took it to Japan, revolutionizing that country’s manufacturing industry. They recognized that unwarranted variations were key quality constraints in any process, and that sustained improvements in outcomes could be attained only through careful analysis and control of this variation.
Hospital Variation and Application
Following the Institute of Medicine’s landmark report a decade ago, these fields of study explicitly converged, and variation began to emerge as a key player in healthcare quality discussions. Sprinkle in a few more ingredients—such as the looming cliff that is Medicare insolvency, a failing economy, and Atul Gawande’s uncloaking of McAllen, Texas—and the transformation of the Kool-Aid is now complete.
A fortunate (or unfortunate, depending on your perspective) byproduct of these analyses has been that physicians are at the sharp end of the most important yet variable decisions in medicine. Why are doctors so different in their practices? The short answer is that we’re human; the long answer is that, well, we’re human.
In complex settings, the literature on medical decision-making tells us that we humans simply are not wired to process more than three to five different options at any one time. Even rocket scientists might disagree if they regularly encountered large boluses of clinical data in the face of an ever-exploding body of knowledge. When we dissect the more straightforward daily decisions, the complexity of the human persona then becomes an overlay with as much variability and heterogeneity as our own genetic makeup.
The Simple Life
The late John Eisenberg, in a book titled Doctors’ Decisions and the Cost of Medical Care, lists a dizzying array of reasons behind physician decision-making: experience, risk tolerance, practice style, incentives, and concept of social good, to name a few.2 Each of these domains could be a unique area of study—just for each individual human practitioner. In an era of genomic medicine, the strongest predictor of the phenotypic quality of care might simply be the genotype of the physician.
This is not a revelation for anyone who has ever questioned another physician’s care. My guess is that it’s been less than a week for most of us. After all, we’re hospitalists, perfectly perched as second-tier providers to judge other physicians’ care. We are air-traffic control for doctors’ decisions, and it’s quite a scene: thousands of independent physicians practicing on isolated islands. Like “outside EDs,” some of these habitats appear quite a bit more aboriginal and remote than others. Now, I will admit that I’ve often dreamed of practicing on an isolated, single-palm-tree island. Armed with only a coconut (my patient) and evidence-based medicine, this would be an overdue retreat from the chaotic morass of illogical (i.e., different from my own) medical decisions.