Clinical Variation in Your Medical Organization?

My Folder

Examining clinical variation in medical practice is an important step to measuring efficiency and effectiveness in care delivery. Dr. Jack Wennberg and other health service researchers have documented extensive variation in the delivery of healthcare in many parts of the world, and this information on practice variation is important for examining the relationships between policy decisions and clinical decisions. Variations in healthcare delivery and utilization can indicate potential opportunities to reduce costs and improve the value of healthcare delivery without compromising patient care.

Additionally, variation in healthcare spending across the United States has been well documented by federal and state agencies. The Congressional Budget Office shows total per capita healthcare spending ranging from $4,000 in Utah to $6,700 in Massachusetts. Spending variations across smaller geographic units have also been documented using Medicare data. County-by-county analyses by the National Center for Policy Analysis show Medicare per capita spending in 2005 varied from just over $5,000 in Nobles County, Minnesota to $8,578 in Rice County, Kansas. Similarly, researchers with the Dartmouth Atlas Project found that among 306 hospital referral regions, Medicare spending per patient ranged from more than $13,000 in some areas to $6,900 in others.

Policymakers want to know why healthcare spending is higher in some areas than in others. More specifically, they want to know if there are some efficiencies in low-spending areas that could be replicated in higher-spending areas, thus reducing healthcare costs overall.

In evaluating practice variation, clinical care can be grouped into three categories with different implications for patients, clinicians, and policymakers:

  1. Effective care is defined as interventions for which the benefits far outweigh the risks. In the case of variation, the right rate of treatment is 100 percent of patients defined by evidence-based guidelines to be in need. Unwarranted variation is generally a matter of underuse.
  2. Preference-sensitive care is when more than one generally accepted treatment option is available, such as elective surgery. Here, the right rate should depend on informed patient choice, but treatment rates can vary extensively because of differences in professional opinion.
  3. Supply-sensitive care comprises clinical activities such as doctor visits, diagnostic tests, and hospital admissions, for which the frequency of use relates to the capacity of the local healthcare system. Among older Americans, most of these services are used in caring for chronic illness. However, regions with high rates of use of supply-sensitive care do not have better overall outcomes as measured by mortality and indicators of the quality of care, suggesting that the problem in the U.S. is overuse of this category of care.

Due to unique patient and/or care-setting characteristics, there will always be a degree of appropriate variation in the practice of medicine, even for patients with the same diagnoses. It is clear, however, that through the use of evidence-based and data-based approaches to clinical decision-making, hospitals and other providers across the country can do much more to reduce inappropriate or unwarranted variation.

Inappropriate variation in clinical practice occurs when non-evidence-based care is provided, or the care lacks wide acceptance, and the high level of variation cannot be supported on a quality or outcomes basis. Such care is often driven by nonclinical factors, such as legal, financial, operational (hospital or other care unit processes), or other considerations that providers bring — consciously or unconsciously — to the process of making decisions about how patients are treated.

Inappropriate variation can lead to reputational problems for healthcare providers, whether physicians, other clinical staff, or affiliated organizations, and often leads to disparate outcomes for patients — either unanticipated or suboptimal outcomes — and higher utilization, costs, and waste. The more healthcare providers base their care on good evidence and good data, and the more they standardize their care on best practice, the more they are likely to avoid these pitfalls.

The Top Four Sources of Variation in Clinical Care

Inappropriate variation is a known cause of poor quality and outcomes. Based on a detailed review of the literature, Dr. Brent James and colleagues (see reference 3) have identified a long list of reasons for practice variation. Here are the top four on the list:

1)          An increasingly complex healthcare environment. Over the last 50 years, we have witnessed huge changes in how care is delivered, with massive growth in complexity. In the 1950s, physicians had a small number of medications to choose from. Now, according to the Institute for Safe Medication Practices, there are more than 10,000 prescription drugs and biologicals — and 300,000 over-the-counter products — available in the United States. There have been equally profound changes in care delivery options and environments, including modern imaging techniques, highly sophisticated intensive care units and surgical suites, catheter-based procedures, transplant services, minimally invasive techniques, and a host of other complicated options. Under the current system, care providers are being overwhelmed with complexity. As stated by Dr. David Eddy, “The complexity of modern American medicine exceeds the capacity of the unaided human mind.”

2)         Exponentially increasing medical knowledge. In 1998, Mark Chassin published an article tracking the publication of randomly controlled trials (RCTs) between 1966 and 1995. One look at the figure below and it is apparent that there has been an explosion in the production of published trials. The number of randomized clinical trials has grown to over 20,000 per year in 2010. In 2004, the U.S. National Library of Medicine added almost 11,000 new articles per week to its online archives. That represented only about 40 percent of all articles published worldwide in biomedical and clinical journals. In 2009, it was estimated that this rate of production had grown to one article every 1.29 minutes. A 1991 study published in the Journal of the American Medical Association found that approximately three to four years after board certification, general internist, and internal medicine subspecialists begin to show “significant declines in medical knowledge.” He estimated that 15 years after initial board certification approximately 68 percent of internists would not pass the American Board of Internal Medicine certification exam. He went on to estimate that to maintain current knowledge, a general internist would need to read 20 articles a day, 365 days a year. Clearly, maintaining current knowledge has become a near impossible task for all clinicians.

randomly controlled clinical trialsPublication of randomly controlled clinical trials

3)          Lack of valid clinical knowledge (inadequate evidence for what care providers do). There have been three published studies looking at the percentage of clinical care that is based on published scientific research. These studies have concluded that only between 10 and 20 percent of routine medical practice has a basis in scientific research. Thus, much of what we do in routine clinical practice is based on tradition or opinion. That doesn’t necessarily mean it is wrong, as much of it has likely been shown to work over time. However, it does suggest that healthcare delivery organizations should use their own data to determine the efficacy of clinical practice and to determine how to improve it over time. This implies the need to create a data-driven continuous learning environment (see references 1, 2, and 4).

4)         Overreliance on subjective judgment. Dr. David M. Eddy noted in HealthAffairs that he and others found that the beliefs of experts with respect to a given clinical condition can vary over a very wide range and that subjective evaluation is notoriously poor across groups over time. For example, Dr. Eddy once reported on a group of experts who were asked what overall reduction in colon cancer incidence and mortality could be expected from the routine use of fecal occult blood testing and flexible sigmoidoscopy. Their answers varied between almost 0 percent and more than 90 percent, with a completely random distribution. These were just some of the findings that caused Dr. Eddy to conclude that people can find anything they want.

The underlying fragmentation of the healthcare system also contributes to poor quality. It impedes the flow and integration of the data necessary for healthcare providers to provide the best possible care. This fragmentation is not surprising given that healthcare providers do not have the payment support, incentives, or other tools they need to communicate and work together effectively to improve patient care.

Variation in care should be examined in order to improve patient care, while reducing healthcare costs. Pinpointing the variations show opportunities for improvement and increased efficiency and effectiveness.

Why role do you see variation playing in your health system? Is it easy for you to measure and analyze in a meaningful way?

Read more about healthcare transformation in the free ebook: Healthcare: A Better Way.

Learn more about why change is necessary for hospitals at: Rising Healthcare Costs: The Case for Change.


  1. Ferguson JH. Research on the delivery of medical care using hospital firms. Proceedings of a workshop. April 30 and May 1, 1990; Bethesda, Maryland. Medical Care. 1991; 29(7 Supplement): 1-2.
  2. Institute of Medicine. Assessing Medical Technologies. Washington, D.C.: National Academy Press; 1985
  3. James BC. Quality improvement in health care: Making it easy to do it right. Journal of Managed Care Pharmacy. 2002; 8(5): 394-397.
  4. Williamson J, et al. Medical Practice Information Demonstration Project: Final Report. Office of the Asst. Secretary of Health, DHEW, Contract #282-77-0068GS. Baltimore, MD: Policy Research Inc., 1979.
Loading next article...