Persistently high-cost Medicare patients are more likely to be younger, come from an ethnic minority group


Persistently high-cost patients are younger, more likely to be members of racial or ethnic minority groups, eligible for Medicare based on having end-stage renal disease and dually eligible for Medicaid compared to their transiently and never high-cost counterparts, a study of more than 5.5 million Medicare beneficiaries published in Health Affairs found.

A research team led by José F. Figueroa, MD, an instructor of medicine at Harvard Medical School in Boston, established that 28.1% of patients who were high cost in 2012 remained persistently high cost over the subsequent two years.

“Persistently high-cost patients had greater relative spending on outpatient care and medications, while very little of their spending was related to preventable hospitalizations,” they wrote. “Healthcare systems and policymakers can use this information to better target spending reductions and care improvements over time.”

The authors set out to answer three main questions: What proportion of Medicare patients who are high cost in one year remain persistently high cost over three years? What are the key characteristics of persistently high-cost patients that differentiate them from transiently or never high-cost patients? And what are the differences in underlying spending patterns of persistently high-cost patients compared to transiently or never high-cost patients?

The investigators conducted their study against a backdrop of rising healthcare costs, noting moves by U.S. policymakers to invest in new approaches to reduce spending and improve health outcomes.

“Over half of spending for the Medicare population in any given year is concentrated among approximately 10% of the patients,” the investigators continued. “What is less clear is how often these Medicare patients remain persistently high cost over time. If health systems are to intervene with these patients, understanding the degree to which high costs persist over time and which patients are likely to remain high cost is critical.”

A vascular perspective

Thomas O’Donnell

The study bears some important implications for the vascular space, says Thomas O’Donnell, MD, professor of surgery at the Tufts University School of Medicine, Boston, and a former hospital CEO.

“It is interesting that the three risk factors identified in this study for persistent high costs in Medicare patients were ethnicity, i.e., if you’re African American or Hispanic; as well as dually eligible for Medicaid, i.e., younger patients,” observed O’Donnell. “And that latter factor highlights, if you will, a vascular problem and a major risk factor, end-stage renal disease. If you have end-stage renal disease [ESRD] in the U.S. you qualify for a special program under Medicaid. That greatly affects the allocation of where the money was spent.”

O’Donnell zeroed in on a central point from both an arterial and venous point of view. “What I thought was most interesting in the persistently costly versus the transiently costly group is the allocation of expenditures, where in the persistently group outpatient cost was greater than inpatient,” he said. “And that’s not what we usually see in the people with some form of arterial disease. The inpatient costs drive the total expenditures.

“For example in the transient group, the inpatient cost, when calculated in a standard way—per patient per year expenditure—was around $7,500 versus almost $13,000 for the persistently. But the outpatient cost for the transient group was one half of the inpatient cost, $4,000. By contrast, for the persistently outpatient cost was the higher area of expenditure at $16,000.

“How does this relate to peripheral arterial disease [PAD]? If you look at an older landmark paper published in Vascular Medicine in 2008, written by the late Alan Hirsch, they identified $4.37 billion on PAD-related treatments. But the most important part of that study was that 88% of the expenditures were for inpatient care. That was 2008 so you know that’s increased, but that gives you a magnitude to compare to this.”

With vascular disease, O’Donnell explained, both the volume of procedures and the procedural costs are dually important factors when considering Medicare expenditures. “If you look at aneurysm surgery volume in the U.S. from 2003 to 2013, there’s been a 36% decrease overall in aneurysm surgery,” he said. “That includes both the open approach and EVAR [endovascular aneurysm repair]. If you then move to carotid disease: there has been a 54% decrease in CMS [the Centers for Medicare and Medicaid Services] carotid procedures from 1999 to 2015.

“So you’ve got two arterial procedures that have seen volume decreases.”

How are such costs controlled? This current study, said O’Donnell, demonstrated “a tremendous increase in costs. It’s not sustainable if we have to expend this much.” He continued, “What has been proposed for controlling Medicare costs starting with the [Barack] Obama administration is the Alternative Payment Model, or value-based programs.”

This involves the use of Accountable Care Organizations (ACOs), in which costs are calculated on a population basis rather than a disease-specific basis. O’Donnell pointed out success achieved by a specialty-oriented ACO for ESRD, which focuses specifically on access to dialysis. “The importance of this approach is that there are some data now that show how you can control ESRD by adopting a disease-specific ACO type of structure. With this they have seen a 4% decrease in hospitalizations and a reduction—just in the first phase—of $68 million, or 2%, in spending by Medicare.”

Yet, it remains an elusive method to control costs for the vascular space at present, O’Donnell explained. “My assessment is none of the therapies for vascular disease currently would satisfy either the volume requirements or the cost requirements to do a specialized ACO or bundled payments.”


The researchers used a 20% sample of Medicare administrative claims data from 2012‒2014, excluding patients for attributes such as lack of a valid beneficiary identification number or sex designation and those who died during the period from which the data were drawn.

Of the 5,507,218 patients in the sample who were continuously enrolled in Medicare, 2.8% remained persistently high cost for three years, while 7.2% were transiently high cost. “Of the top decile of patients by costs in 2012 (n=550,722), 28.1% remained consistently high cost for three years, while 71.9% were transiently high cost,” the authors wrote.

Regarding age, the high-cost patients were younger than the transiently high cost, the investigators found: a mean age of 66.4-years-old to 73.3. Similarly, persistently high-cost patients were also more likely to be black (20.1%) and Hispanic (4%) than their transiently high-cost counterparts (9.2% and 1.8%).

In their findings, the investigators laid out the impact in dollar terms. “On average, in the first year persistently high-cost patients spent $64,434, compared to $45,560 for transiently high-cost patients and $4,538 for never high-cost patients,” they continued. “In subsequent years persistently high-cost patients spent slightly more per year than in the index year, while the transiently high-cost patients, by definition, spent substantially less in subsequent years. Persistently high-cost patients spent more across all categories of spending, with the most spending occurring in the outpatient and inpatient settings and for drugs.”

In conclusion, the investigators wrote: “We found a modest degree of persistence of high costs in the Medicare fee-for-service population. Patients who were persistently high cost were younger, more chronically complex, and more likely to be a member of a racial or ethnic minority group, compared to transiently and never high-cost patients. The greatest differences in spending between persistently high-cost patients and the other two groups were for outpatient services and drugs, with a smaller difference in spending for preventable acute care use. Policymakers and clinical leaders should consider these issues when developing programs to improve care for high-need, high-cost patients.”

Source: doi: 10.1377/hlthaff.2018.05160


Please enter your comment!
Please enter your name here