Are End-of-Life Care Costs Actually Decreasing?

06/08/18

It is well documented that health care expenditures are on the rise in the United States. Recent studies of end-of-life practice patterns suggest that increased hospice care costs more than it saves, end-of-life care intensity continues to increase, and end-of-life care costs may be accelerating.

However, a group of researchers from the Dartmouth Institute for Health Policy and Clinical Practice (Lebanon, NH) have found that per-capita end-of-life spending is actually decreasing and contributing to overall per-capita Medicare spending growth moderation. Researchers point to changes in both supply and demand factors for this trend.

To better understand the design and implications of this study, Journal of Clinical Pathways spoke with William B Weeks, MD, PhD, MBA, lead author and professor at the Institute. Dr Weeks discusses the data that led to this conclusion as well as plans for future research.


How did you come to the conclusion that spending for medical care at the end of life is declining?

Dr. Weeks: We used publicly-available data from the Dartmouth Atlas to examine overall and per-capita spending by year for Medicare fee-for-service enrollees who survived the year and those who died (and whose spending, therefore, was described as end-of-life spending). We examined data from 2004-2014 and found that per capita costs associated with end-of-life care increased until 2009 and declined rather sharply thereafter.

Do you believe that end-of-life quality of care is decreasing as a result of cost of care decreasing?

Dr. Weeks: Unfortunately, we were not able to look at quality of care issues in our study. That is the subject of a grant that we have proposed since discovering that per-capita end-of-life care costs have been declining. For this study, we did not have any metrics that could be used to evaluate end-of-life care quality. However, it is really an important question because a concern is that care quality might be decreasing as costs are decreasing. However, if end-of-life care quality is increasing or remaining constant while costs are decreasing, that would be a good thing. It is vital that further research determine what is occurring with end-of-life care quality. 

Do you believe that the transition from fee-for-service and traditional payment models to value-based care models has contributed significantly to the decreased end-of-life care costs?

Dr. Weeks: The transition to value-based care models might be an explanation for our findings. That is another important question that we have proposed in our grant. At this point, we do not know the contributing factors to the decline in per-capita end-of-life care costs among Medicare fee-for-service enrollees. It is possible that value-based care models are influencing overall costs and driving down care costs; however, one would think that uptake of value-based care models would also drive down care costs for Medicare fee-for-service survivors, who represent about 95% of the Medicare fee-for-service population.  However, we really did not find much change in per-capita care costs for survivors between 2009 and 2014, so value-based care models might not be driving our findings.

A different possibility is that conservative approaches to end-of-life care have become more popular. When I was a medical student and a resident, a hot-button topic was whether patients should necessarily have to accept life-sustaining food and fluids; now, it is accepted that if patients do not want foods and fluids, they do not have to receive them. The point is that there has likely been a cultural transition that might account for the trend that people are more accepting of less intensive end-of-life care, which could explain our findings.

Another possibility is that the growth of palliative care board certified physicians could account for our findings. Physicians who can better manage end-of-life care can perhaps provide more efficient care that does not entail repeat hospitalizations and tests, actions that might help keep costs down.

Another possibility is that the financial crisis of 2008 might have contributed to the downward cost trend in per-capita end-of-life care costs. As a result of the crisis, people may have become wearier of spending money at the end of life and more conscious of their out-of-pocket spending.

Clearly, there are a variety of possibilities that we are hoping to explore in further studies.

In future studies, how important will it be to track certain conditions or other potential factors that could help explain this trend in decreased spending? What are some of the potential factors you anticipate will be the most impactful?

Dr Weeks: Along with the aforementioned factors, it seems that the changes in end-of-life spending that we found could be due to other changes in supply factors or changes in demand factors.

For instance, it could be that when the Affordable Care Act was passed, an influx of new patient enrollment may have caused physicians to redirect their efforts; perhaps they did not put as much effort into taking care of people at the end of life or did not think that such patients should be admitted as often due to resource constraints. This is all speculation, but we believe those supply and demand factors might be causative for the changes that we observed. 

Are there any other points you would like to make that are important for this conversation?

Dr Weeks: It is important to note a potential limitation of our study: we examined only older fee-for-service Medicare enrollees. While some may believe that this population is not representative of the larger population, it is the population in which most deaths occur, particularly the kind of deaths that are anticipated (ie, old age, Alzheimer’s, and long-term cancer). This is a really important population to study in order to capture where the overall costs of death incidents are coming from.

It is also important to note that, although only about 5% of the Medicare fee-for-service population dies each year, those patients account for about 40% of annual Medicare Part A and Part B expenditures. A concern is that Medicare is headed for bankruptcy if health care cost growth continues at the current rates. If we can determine why per-capita end-of-life health care costs dropped so quickly between 2009-2014, whether that trend has continued, and whether the reduced care costs are associated with changes in care quality, that information might be quite helpful in trying to curtail overall health care costs in the future. Our findings are important and relevant to the nation as a whole, and it will be important to answer these questions in further research.