Cutler D. What Near-Death Experience Can Tell Us About Medical Care. JAMA Forum Archive. Published online August 27, 2014. doi:10.1001/jamahealthforum.2014.0034
In 2005, a survey was conducted examining how aggressively physicians treat patients near the end of life. Some 1500 primary care physicians and cardiologists were presented a scenario describing an 85-year-old man with severe congestive heart failure. The fictional patient was described as receiving maximal medications, using supplemental oxygen at home, and not being a candidate for surgery or other procedures to treat blockages in blood vessels to the heart. When the patient experiences worsening shortness of breath (a symptom of fluid accumulating in his lungs) and makes an office visit, increasing his supplemental oxygen makes him more comfortable.
The survey group was then asked “How often (always/almost always, most of the time, some of the time, rarely) would you arrange for each of the following?” The choices were
A. Allow the patient to return home and receive increased oxygen and increased diuretics (to remove excess fluid from the body)B. Admit to the hospital (but not to the intensive care unit [ICU] or coronary care unit [CCU]) for aggressive removal of excess fluidsC. Admit to the ICU/CCU for intensive therapy and monitoringD. Place a pulmonary artery catheter to assess the severity of heart failure and to optimize drug treatmentE. Recommend a pacemakerF. Initiate or continue discussions about palliative care
A. Allow the patient to return home and receive increased oxygen and increased diuretics (to remove excess fluid from the body)
B. Admit to the hospital (but not to the intensive care unit [ICU] or coronary care unit [CCU]) for aggressive removal of excess fluids
C. Admit to the ICU/CCU for intensive therapy and monitoring
D. Place a pulmonary artery catheter to assess the severity of heart failure and to optimize drug treatment
E. Recommend a pacemaker
F. Initiate or continue discussions about palliative care
Recently, the responses to this survey were analyzed by me, Ariel Stern, PhD; Jonathan Skinner, PhD; David Wennberg, MD, MPH. Although the data are nearly a decade old, there is no reason to think that answers today would be any different. Our analysis of these responses reveals the troubled state of American health care—as well as the possibilities.
Options A and B are less-intensive ways of managing the patient and each finds support in the literature. Three options—C, D, and E—are much more intensive but have no evidence to support them (they were not recommended in the guidelines then and are not recommended now). Option F, a palliative care discussion, is clearly warranted; half of patients like this die within a year. Thus, a physician following practice guidelines would be expected to choose A or B, plus F.
But the answers that physicians gave were far from these recommendations. One-quarter of physicians recommend 1 of the 3 most aggressive options. Only about half would likely have a palliative care discussion. Primary care physicians and cardiologists differ slightly in their answers, but not enormously so. Very aggressive treatments were supported by 22% of the primary care physicians and 28% of the cardiologists, while only 49% and 45% would have a palliative care discussion.
This issue has great import for medical spending. Based on the responses to our survey, we called the physicians who consistently advise aggressive end-of-life care not supported by guidelines “cowboys,” and those who would have a palliative care discussion “comforters.” A physician can be both a cowboy and a comforter, but empirically, few are.
Cowboys and comforters are not randomly distributed in the population, and medical spending reflects this. Medicare spending for end-of-life care is 60% more in an area where physicians are mostly cowboys than in one with few cowboys. An area full of comforters will spend 27% less. Generally, spending in this arena negatively tracks patient preferences. Patients routinely express a desire for comfort over aggressive care at the end of life.
What makes physicians become cowboys or comforters? Many physicians say that the fear of being sued for malpractice drives them to do too much. But our survey evidence does not support this theory. Although 41% of cardiologists report that they have sometimes or frequently done a cardiac catheterization because of malpractice fears, there is no correlation between the response to that question and being a cowboy. Nor is there a relation between physicians who report responding to patient expectations and being a cowboy. Some characteristics are correlated with not being a cowboy: those who are younger or who practice in larger groups are much less likely to be cowboys than other physicians.
In reality, I think there are 2 fundamental factors at work, and each is important for the future of medical care. First, although the evidence of what is appropriate may exist in guidelines and medical journals, nothing compels physicians to follow it. A cardiologist can believe he has better outcomes than the literature suggests. And if the issue comes up, “my patients are different” or “my outcomes are better” usually ends the conversation. These statements cannot be generally true or the literature would reflect that. But who can say if they are true for any particular physician?
The interesting question is why these statements are rarely challenged. Sometimes, big systems do have these conversations, which may explain why there are fewer cowboys in hospital-based and group practices. Intermountain Healthcare in Utah is famous for the idea that internal standardization can lead to better care. Even today, though, Intermountain remains the exception.
Look behind the curtain, and one sees the importance of money. Standardization generally results in a greater reduction in overused services than an increase in underused services. Thus, in a fee-for-service system, standardization results in a revenue loss. Add to this the financial and psychic costs of physician profiling (to compare practice patterns) and standardization, and it is easy to see why even the largest systems avoid it.
As the payment system changes, this calculus will change as well. In an accountable care organization (ACO), money saved by avoiding unnecessary procedures becomes surplus to the organization. The same is true for organizations accepting a predetermined bundled payment for an episode of illness. There are more than 600 ACOs in the country today, and many hundreds of institutions are receiving bundled payments. Thus, internal analysis and standardization may become significantly more frequent.
Why do only half of clinicians discuss palliative care for patients at the end of life? A likely reason is that such discussions are painful and physicians naturally wish to leave them to someone else. This choice may be easy to rationalize because so many physicians are involved in patient care near the end of life. The primary care physician can say that she is not an expert in severe congestive heart failure and hence the cardiologist needs to discuss options with the patient. The cardiologist can say that the primary care physician knows the patient better and is better equipped to have the palliative care discussion. Each is right, but the outcome is still bad.
Again, what is needed is a system for seeing that such discussions take place. At institutions such as Gundersen Lutheran in Wisconsin, end-of-life conversations are routine as people age and end-of-life spending is correspondingly lower. In other areas, the conversations are avoided. Not surprisingly, the Dartmouth Atlas shows enormous variation in end-of-life care spending across the country.
But there is cause for hope: if all physicians followed the guideline recommendations, end-of-life costs would decrease by one-third, and overall Medicare costs would decrease by 15%. This outcome can be achieved without rationing: gathering data, having discussions with colleagues and patients, and setting standards would do an enormous amount.
My suspicion is that much of health care is like the case of caring for the fictional 85-year-old with congestive heart failure. No one is trying to do the wrong thing, but the consequences of unstandardized and often excessive care is wasted money and unnecessary pain. Establishing the right systems, realigning payments, and pushing for the right care can accomplish far more than many think is possible.
David Cutler, PhD David Cutler, PhD, is the Otto Eckstein Professor of Applied Economics in the Department of Economics and holds secondary appointments at the Kennedy School of Government and the School of Public Health at Harvard University...