Taking issue with the logic of your last paragraph. The cost graphic makes no mention of condition, age, or total duration of care/# of years kept alive with a chronic illness. What if the Hopkins care is keeping people alive longer? No way to tell since this information only charts cost 2 years prior to death. Hopkins could be getting penalized for keeping people alive for a full 2 years or more of chronic illness care, and a hospital that can only keep people alive for 1 year of chronic illness care would have a lower 2 year look back from death cost figure for providing care a shorter period of time.

The information provided is not very valuable and absent additional info, can't be used to draw the conclusion you're drawing.

It seems as though hospitals are defending their high prices by noting that their patients are soon dead.

Hi Josh: Yes, you make a good point. But the principle of Ockham's razor (the simplest explanation is often the right one), the sheer size of the study (researchers looked at 4.7 million cases) and Dartmouth's own analysis point to the conclusion that some hospitals are delivering "too much" care given the medical outcomes they achieve. Dartmouth found a high correlation between medical spending and medical capacity: ie., the more doctors, nurses and hospital beds a region had, the more care was delivered relative to a diagnosis. As Dartmouth's David Goodman told Jamie Smith Hopkins and Kelley Brewington: "We as physicians use whatever is available to us - we help by doing. If there's more that allows us to do more - more physicians, more consultants, more availability of ICU beds - we use it."

Crunching numbers looking at a spread sheet it's easy to say they are delivery "too much care." When it's your mom or dad who needs the care that's something entirely different. From a moral philosophy standpoint it is a decision that should be as decentralized as possible as no single man or small group of men are qualified to make it.

Post from Jeff Nelligan, Senior Director of Strategic Communications, Marketing and Communications, Johns Hopkins Medicine

Jay, we always read your column and thought we needed to respond to this recent piece. Appreciate you giving us the space.

There is no doubt that the Mayo Clinic is a world-class medical center. And there is little doubt The Dartmouth Atlas is valuable in purely assessing Medicare hospital costs. But the Atlas — and the Mayo comparisons — haves limitations that cause us, as well as others, to question their results.

First, the Dartmouth Group began its study years ago focusing on geographic variations. Baltimore is vastly different from Rochester, Minn., home of the Mayo clinic. Our surrounding population also differs dramatically from Mayo’s in terms of income, education, disabilities, crime rates and employment status. For example, 35 percent of our patients are African-American vs. 2 percent for Mayo; 34 percent are below poverty level vs. 5 percent for Mayo; 29 percent are disabled vs. 10 percent for Mayo; 3 percent have a bachelor’s or higher degree vs. 35 percent for Mayo. Another example is security. Unlike Mayo, we have to maintain a security force of over four hundred professionals to assure a safe campus for our patients, visitors and staff. The most striking factor in terms of geographic variation, which is not accounted for in the Dartmouth model, is cost of living. The cost of living index (including costs for food, shelter, tax rates and transportation among others) for Baltimore is 121 compared to 98 for Rochester. This alone would explain 50 percent of the difference between our costs and Mayo in the Dartmouth data.

Second, Dartmouth has a very limited risk adjustment for differences in the underlying conditions. Our patients also tend to have more associated comorbidities, such as diabetes, hypertension, drug abuse, heart disease, etc., than those seen by many other hospitals. Much of this is attributable to the population difference described above. The Dartmouth model does not account for any of these factors, each of which can have a significant impact on health management strategies and their related costs. The Dartmouth authors suggest that you do not have to account for severity of illness among patients because “the study only focused on patients who died so we could be sure that patients were similarly ill across hospitals. By definition, the prognosis of all the patients in the cohort was identical — all were dead after the interval of observation. Therefore, variations cannot be explained by differences in the severity of individuals' illnesses.” We find this to be a problematic statement because it assumes that the experience of all patients in the cohort had similar experiences. This assumes that all end-of-life care is the same and the only thing that matters is how much it costs. It does not reflect care that appropriately extends life and/or improves the quality of life. If we provide more care to patients who thus have longer productive lives, this is not reflected in the data associated with Dartmouth.

Third, Dartmouth indicates that more hospital days and more inpatient consults are "aggressive" care versus conservative care. One might argue that patients and their families prefer to be at a place that provides more physician services during their hospital stay — one that doesn’t push them out of the hospital too fast.

Fourth, their premise that higher costs are not associated with higher quality is based on indicators that may not actually indicate higher quality. For example, indicators such as the percent of people who died in the hospital or in an ICU do not necessarily point to poor quality.

Finally, it must be understood that because of the all-payor system, unique to Maryland, private payors do not subsidize governmental payors in our state. The Medicare rate we receive in Maryland is higher than other states because our system requires that all payors pay the same rate for a given service at a given hospital. This prevents “cost-shifting” of Medicare and Medicaid due to low reimbursement rates to other hospital. Since the Dartmouth study looks only at Medicare patients, the costs for the Medicare patients in other states appear to be less, but all other patients — that is, all other patients not on Medicare — pay a higher amount than those in Maryland.

We hope that readers can appreciate that a simple comparison between the costs associated with care at Johns Hopkins, the University of Maryland Medical enterprise, Mayo and at many other institutions, does not even begin to tell the whole story. There are many factors that explain these apparent differences, and when these factors are included in the calculations, the differences begin to melt away.