Medical overtreatment is the inverse of former Supreme Court Justice Potter Stewart’s definition of pornography: while easy to define in concept, it can be hard to know it when you see it.
A treatment that is appropriate for one patient can also be unnecessary or even counterproductive for another, depending on the patient’s condition. This has been a major obstacle for studies seeking to pinpoint overused services, which by the most expansive estimates may account for as much as a third of the nation’s health spending.
Using a novel method, a study released Monday by researchers from the Harvard Medical School Department of Health Care Policy evaluated the prevalence in Medicare of 26 tests and procedures that have been found to offer little or no clinical benefit. The services were mostly culled from mainstream lists, including studies in medical journals, the Choosing Wisely campaign and the U.S. Preventive Services Task Force. The procedures included a form of back surgery in which collapsed spinal disks are filled with cement, as well as CT scans or MRIs on people with headaches.
The study in the journal JAMA Internal Medicine concluded that at least one in four Medicare beneficiaries received one of these 26 “low-value” services during 2009, and possibly significantly more. Two of the researchers, Aaron Schwartz and Dr. J. Michael McWilliams, said in an interview the total number of unnecessary treatments was surely higher, since their study encompassed just 26 tests and procedures.
“There are hundreds of other low-value services,” said McWilliams, a Harvard professor.
The researchers took a two-step approach. First, they analyzed Medicare billing records to isolate services provided to patients where they might not be warranted. For example, when tallying the number of surgeries to remove plaque from a carotid artery in the neck, they focused on patients who did not have any history of symptoms of mini-strokes.
The researchers calculated that there were 21.9 million instances of the 26 low-value treatments during 2009. Forty-two percent of beneficiaries received at least one such service, costing Medicare $8.5 billion, or 2.7 percent of spending.
To err on the side of caution, the researchers then analyzed the records a second time, using the most restrictive definitions they could devise. In re-examining the frequency of the carotid surgeries, for instance, they excluded operations that were associated with an emergency department visit, because that might indicate the patient was having a stroke. They also looked at only female patients, because there is evidence this procedure, known as a carotid endarterectomy, is less beneficial for women. With the more selective filter, the amount Medicare spent on that surgery dropped from $263 million to $110 million.
The frequency of other procedures dropped by an even greater degree. The researchers initially found that out of every 100 patients with general back pain, 12.4 received an imaging test. When the researchers limited their search to back scans of patients with no diagnoses in the billing records that might possibly justify the scan, such as a fever, cancer or drug use, they identified 2.5 instances for every 100 beneficiaries. The dollar figure those back scans cost Medicare shrank from $226 million to $82 million.
But even with the more specific definitions, the researchers calculated that 9.1 million low-value services were provided in 2009, with 25 percent of Medicare beneficiaries receiving at least one. Those services cost Medicare $1.9 billion, or 0.6 percent of overall spending.
The study found significant amounts of spending on low value-services even in parts of the country where they were least frequent. They wrote that “overuse may be substantial even among more efficient providers.” However, it is not clear how often those services were being provided in the four years since the time examined by the study, when overtreatment received more attention within the medical profession. The Choosing Wisely campaign, for instance, did not begin until 2012.
The researchers said that the wide gap between how many inappropriate services were detected using the broader versus the limited method illustrated how difficult it may be to try to police the appropriateness of specific treatments. Instead, they recommended that reforms focus on moving away from paying doctors and hospitals for each service they provide. Instead, they wrote, giving caregivers set sums to treat a patient’s overall health—known as global payments—or “bundled” payments to treat one particular episode, such as replacement of a knee and the subsequent rehabilitation, “could allow greater provider discretion in defining and identifying low-value services while incentivizing their elimination.”
There are several such efforts being tried out. Massachusetts is experimenting with global payments. The federal Center for Medicare & Medicaid Innovation has approved several hundred bundled payment experiments around the country.