A Tale of Two Cities: Unwarranted Variation in Tonsillectomy Rates in Vermont

By Jill Arnold


A summary of a story told by Shannon Brownlee in her 2007 book, Overtreated.

Jack Wennberg is a physician, PhD in public health and has been a professor at Dartmouth Medical School for more than thirty years. Known early in his career for blowing the whistle on the use of the dangerous drug Orabilex in the interest of patient safety, Wennberg decided early on that he was more interested in improving the health of communities rather than treating one patient at a time.

In 1967, Wennberg took a position at the University of Vermont as director of a newly formed Medicare regional planning program. The NIH was giving block grants to medical schools to create regional programs for improving the treatment of certain conditions.

Wennberg used the grant to investigate what people actually needed in the way of health care in Vermont with a goal of making sure new facilities were created to fill in any gaps in access to care. Like his colleagues at the time, he assumed that the problem that needed resolve was that patients were not receiving enough treatment. The assumption was that if hospitals were full, it indicated that there were not enough beds to go around and another hospital was needed.

He and his colleague, statistician Alan Gittlesohn, went from hospital to hospital and offered small grants to each of Vermont’s sixteen hospitals to digitize their data.

When they reviewed them, nothing made sense. For example, in Middlebury, 7 percent of children under the age of 16 had their tonsils removed. In Morrisville, just two hours away, 70 percent had their tonsils removed.

Wennberg and Gittlesohn explored possible explanations for these small area variations. Maybe patient demand for procedures varied between regions? Perhaps patient characteristics were the reason for these geographic differences and people were simply healthier in certain areas of the state.

Yet it still didn’t make sense to Wennberg and Gittlesohn that the parents of one town would demand tonsillectomies for their children at 10 times the rate of a neighboring town. They just couldn’t believe that 10 times more children had swollen tonsils requiring surgery in Morrisville.

When they presented the data to physician colleagues at the University of Vermont, they were nonplussed. Brownlee summarized the attitude of their colleagues:

Of course differences between patients were driving the differences in the amount of surgery. Patients in areas where surgery rates were higher obviously needed more surgery. Either that or patients in areas where rates were low simply weren’t getting enough surgery.

Wennberg and Gittlesohn tried to get their results published in every major medical journal but they were turned down. Even when the paper was finally published years later in 1973, it was ignored.

Wennberg teamed up with Floyd Fowler, an MIT sociologist, and they randomly surveyed 4,000 people throughout the state of Vermont to measure illness and whether they were demanding medical care more or less often.

Not surprisingly, they found a homogenous population across the state with very little variation in relevant factors such as race, access to care, frequency of emergency room visits, etc. They hypothesized that the deviation was coming not from patients, but from doctors. Wrote Brownlee:

Little consensus existed in the early seventies about when a child really needed to have his tonsils removed. When Wennberg went to Morrisville and sat down with each doctor he learned that they were simply too quick to use the scalpel compared with their colleagues in other parts of the state. Morrisville doctors were yanking out tonsils when more-conservative doctors might have waited to see if their young patients outgrew their susceptibility to sore throats and ear infections—as most children do.

Morrisville doctors were shocked to see the results. Wrote Brownlee, “[t]hey had no idea how different their practices were from their colleagues, that they were subjecting children to unnecessary tonsillectomies.” In order to reign in their use of the surgery, they implemented a system of requiring that a second opinion be sought whenever a tonsillectomy was recommended. As a result, the doctors of Morrisville performed two-third fewer tonsillectomies over the next five years.



This post originally appeared on The Unnecesarean on April 22, 2011.