In my first year of surgical residency, a mentor told me, “You’re going to spend years learning to cut well, but one of the most useful skills you’ll have is knowing when the knife isn’t the answer.” I found myself reflecting on this during a recent policy seminar that highlighted regional variations in surgical utilization. The first projector slide showed a smattering of blue dots from the “Dartmouth Atlas” database arranged like a long, thin diamond. Each dot represented rates of total knee replacements performed per thousand Medicare beneficiaries in a different state. While many of the dots clustered toward a rate of 8–9 per thousand Medicare beneficiaries, the long spread of the data points astonished me. At the top of the diamond, one dot represented Utah, with a total knee replacement rate of 13.2 per thousand Medicare beneficiaries. At the bottom of the diamond was Hawaii, with only 4.6 total knee replacements per thousand Medicare beneficiaries. A chart showing rates by hospital referral region revealed a wider chasm. In Idaho Falls, Idaho, 15.2 total knee replacements were performed per thousand Medicare enrollees in 2015. In that same year, in the Bronx, New York, the rate was 4.4.
It wasn’t necessary to cross state lines to find such marked variation. Highlighting hospital referral regions in Texas, represented by a dot within the blue diamond, showed rates of utilization for total knee replacements ranging from 12.6 per 1,000 in Victoria, Texas to 6.9 per 1,000 in El Paso in 2015.
As the charts whirred by, others in the class were still trying to make sense of the data — but its fiendish implications were obvious to me: Surgeons were either performing too many procedures for their own financial gain (in the regions with high rates), or were depriving patients of a useful intervention in other places for unknown reasons (in the regions with low rates). As the only budding surgeon in the room, my brow furrowed and I felt myself brace for the uncomfortable discussion sure to follow. How could there be such tremendous disagreement on when and for whom to perform these procedures?
In medical school, the appeal of surgery was straightforward for me. I gravitated toward the objective and methodical means by which surgeons solved patient problems with surgical intervention. These findings called that objectivity into question. It wasn’t just knee replacements — chart after chart showed drastic variation in surgical volume by region for different procedures. Even when regions had similar patient demographics and prevalence of disease, which might account for variations in surgical procedure volume, idiosyncratic patterns of utilization were observable.
I considered many plausible explanations for the variation in surgical procedure rates. The most concerning theory sprung to mind first: Financial incentives are different for surgeons in region A versus region B, driving volume where profit follows. Yet this data was derived from Medicare and Medicaid populations, so reimbursement rates for procedures should have been similar.
Theory two: Patient factors are fueling utilization rates. If there are more obese and elderly people in a region, there are going to be more total knee replacements. Obesity rates in Indiana and Texas were certainly higher than in New York City or Hawaii, where the rate of total knee replacements per thousand Medicare beneficiaries were substantially lower. But, Utah’s rate of total knee replacements was high despite its relatively low obesity rate….
The paradox isn’t a phenomenon unique to the United States. Several studies have described substantial regional variation in utilization of total knee replacement in the United Kingdom, also. The criteria surgeons use to identify patients for total knee or total hip replacement is variable. However, rather than nefarious profit-seeking or overutilization, I believe now that it is diverging physician beliefs about the efficacy of a given procedure, coupled with differences in training, that may contribute to variation in utilization.
I believe that the “correct” rate of total knee replacements is somewhere in the middle of the extreme cases represented by Utah and Hawaii. The knee-jerk reaction is to place limitations on the number of procedures by clamping down on surgeons and creating more hoops for them to jump through before a patient is eligible surgery. The existence of drastically different rates suggests that there is an opportunity to reduce expenditures and curtail wasteful overutilization on one end of the spectrum. That said, accepting this line of thinking could have unintended consequences for vulnerable and underserved patients. I have worked in various practice settings, and I have seen how resource limitations and access issues can prevent underinsured, uninsured, and incarcerated patients from having timely arthroplasty and elective sports procedures done that would elevate their quality of life. This is true despite an availability of surgeons willing and eager to cater to these patients. Financial barriers, lack of social capital, and biases rooted in the structural deficiencies of the health care system perpetuate disparities. There is literature, for example, demonstrating disparities in rates of recommendations for total knee replacement for minority and women patients. And, if we swing the pendulum too hard in the direction of restricting elective procedures, it is likely the most vulnerable patients who will be impacted first.
It is worth noting also that patient-induced demand may also influence surgical utilization rates. I have seen some patients come to an appointment aggressively demanding surgical intervention despite recommendations against it — and though my mentors have often stuck by their clinical decisions during these encounters, where there is patient demand, supply will often follow. Assessing direct-to-consumer advertising for these different procedures by region may offer further clues about drivers in variations of surgical volume.
In recent years, a paradigm shift has occurred. As volume-based care, also known as fee-for-service, has been slowly phased out by value-based reimbursement structures, programs like the Comprehensive Joint Replacement Bundle and other value-based care options have been implemented. It will be interesting to see how rates in variation continue to change. The verdict on whether bundles succeed in incentivizing physicians to focus less on volume of care (and, by extension, revenue), and more on quality of care is not out yet. Training medical students and resident physicians who are cognizant of utilization rates and literate in health policy will do much to create leaders who seek the root causes of variation patterns and may aid in the creation of uniform guidelines to identify eligible patients for these procedures.
Dr. Christian Pean is a third-year Orthopedic Surgery resident physician at the NYU Langone Orthopedic Hospital. He is on the board of directors for the global health non-profit organization Orthopedic Relief Services International and a member of the Speaker’s Bureau for the Physicians for Criminal Justice Reform organization. He is a 2015 recipient of the American Medical Association Excellence in Medicine Leadership Award, a previous National Medical Fellowship Recipient, and a member of the Gold Humanism Honor Society. His research interests include ethical analysis of health policy, addressing racial health care disparities, global surgery, and violence as a public health issue.