by Tom Robertson
Executive Director, Vizient Research Institute


It’s interesting to think about the industrial engineering and almost fanatical insistence on consistency that goes into the production of something as inconsequential as a hamburger or a can of soda, and to contrast that with our tolerance for variation in the treatment of cancer or the use of high-cost imaging in the diagnosis of back pain. Two of the top five worldwide brands outside of the technology sector are McDonald’s and Coca-Cola, valued at $39 billion and $58 billion, respectively, by Forbes.

A large part of that brand equity rests on reliability and predictability. Walk up to a McDonald’s counter in downtown Philadelphia or drive through a franchise in South Dakota and you know exactly what you’re going to get. With nearly 40,000 locations, McDonald’s has consistency down to a science. In 1985, Coke changed the formula for its signature soft drink. It took only 79 days to realize the magnitude of that mistake, and they immediately reverted to their original formula which, in a masterful public relations stroke, is kept in a bank-like vault in Atlanta.

McDonald’s and Coca-Cola know the value of consistency; neither would tolerate a franchise or a regional bottling partner straying from the norm. So why do health systems, which carefully cultivate brand identities in their local markets, seem so tolerant of extraordinary intrasystem variation in the care of clinically similar patients? And should they be?

When they initially form or when they expand, health systems typically promise patients “the right care in the right place at the right time.” Implicit in that promise is the notion that two identical patients who walked into two provider locations within the same system would receive the same course of treatment. Our latest research discovered that intrasystem variation is a persistently prevalent phenomenon. In fact, intrasystem variation is often larger than the variation observed between health systems in different markets. 

We compared the occurrence of “marker events” across hospitals within 209 health systems nationally to assess the relative variation in utilization both within and between health systems. Marker events included post-acute care (PAC) facility use following uncomplicated joint replacements, major imaging in the emergency department for patients presenting with back pain, repeat imaging within 90 days for selected diagnoses, and two end-of-life utilization measures. Differences between health systems were noteworthy; it was not uncommon to see four-fold variation between systems in the occurrence of marker events.

The variation within health systems, however, was even higher. Two patients with virtually identical clinical profiles who walked into two providers within the same health system had markedly different probabilities of experiencing a marker event. Intrasystem variation was the norm, not the exception. Rare examples exist where variation within a system is low, but for each marker event studied, the overwhelming majority of health systems exhibited wide variability across their facilities. Where low variation was observed for a given health system, it was limited to one or two marker events; no health system studied exhibited consistently low variation across all marker events.

Acute bundled pricing, typically anchored around a surgical event, is not a new concept. Hospital and physician payments for heart surgeries were bundled together in the late 1980s and the practice resulted in a precipitous decline in market prices for the targeted procedures. Organ transplantation has been paid under bundled prices for more than 20 years. What is new about the latest incarnation of bundled pricing is the inclusion of pre-admission and post-discharge services, such as PAC facility use following joint replacement surgery.

Prospective bundled pricing penalizes avoidable variation. As bundled pricing expands to include longitudinal episodes of care, such as newly diagnosed cancer cases or back pain, we should expect to see observed intrasystem variation in utilization drop sharply, in response to financial disincentives for avoidable resource consumption. An interesting question to ponder is why health systems have not considered internal variation to be a serious threat to their brand equity while McDonald’s and Coca-Cola are so intolerant of even the slightest variability.

A major difference between McDonald’s or Coca-Cola and a health system is the consumption behavior of the customer/patient. McDonald’s and Coke are heavily dependent on repeat business. Consumers develop expectations and the products have to meet them every time. In health care, chronically ill patients become serial consumers, but most acute medical episodes occur only once for any individual patient. Two similar patients may receive completely different services for the same condition within a health system, but neither of them has the experience of the other to compare to their own. Nor do most patients have the expertise to judge whether the battery of services that they received were necessary and appropriate.

Unlike health care, consumers of hamburgers and soft drinks know what they like, and they are very likely to consume again. Consistency and predictability are essential to brand equity for McDonald’s or Coca-Cola. By contrast, most patients have no basis on which to assess appropriateness, they lack the points of reference to compare their care to that of others, and once an acute episode has passed, they are far less likely to buy the same services again. As a result, health systems have been largely insulated from the risks to brand equity associated with avoidable variation in other industries.

Payers – both governmental and private-sector insurers and employers – are in a position to compare utilization patterns across multiple patients, and to identify intrasystem variation that is not obvious to individual patients. To date, employers and commercial insurers have tended to focus on unit prices for high-cost single events and have been slow to challenge intrasystem variation in utilization patterns, particularly when the variation occurs within a longitudinal episode of care. As CMS expands bundled pricing to include longitudinal episodes, to the extent that commercial payers follow suit, there may be some brand risk arising from avoidable variation, but the more immediate threat to health systems will come in the form of shared financial accountability. Prospective payments will make intrasystem variation economically intolerable.

It is unclear when (or even whether) variation in care processes will threaten the brand equity of health systems. It seems likely that risk-sharing in the form of expanded bundled pricing, and its associated economic pressure to reduce variation, will come more quickly than brand vulnerability. But early movers may change that dynamic significantly.

Patients trust health care providers to do the right thing. They have little or no idea about intrasystem variation; they believe that providers are practicing consistently. And everything that health systems say in their promotional materials – the right care at the right place at the right time – reinforces that belief. Health systems that deliver on that promise, every time in every setting, can translate consistency into brand strength and differentiate themselves from competitors who are slower to optimize episode efficiency.

About the author and the Vizient Research Institute™. As executive director of the Vizient Research Institute, Tom Robertson and his team have conducted strategic research on clinical enterprise challenges for 20 years. The groundbreaking work at the Vizient Research Institute drives exceptional member value using a systematic, integrated approach. The investigations quickly uncover practical, tested results that lead to measurable improvement in clinical and economic performance.

Published: March 9, 2017