A recent report
found an 8% drop in hospital admissions since Part D was implemented, leading the authors to estimate an annual savings of $1.5B to the system. The report sounds promising, but it doesn’t answer some questions. Declines in hospital rates for dehydration, COPD and heart disease were found. Let’s think about this assertion for a minute. These decreased rates of admission were “associated” with the start of Medicare Part D, but an association doesn’t prove causality. What other factors might be happening at the same time that could also be involved?
First, look at specific diseases mentioned. Mortality rates for coronary artery disease (CAD) have been declining in the U.S., so I would expect to see decreased admissions for CAD events, and for its common complication, heart failure. Most experts would assert that CAD is on the decline due to better treatments and better control of risk factors, notably cholesterol and smoking. COPD is a chronic complication of smoking. For at least 90% of those afflicted by COPD, the cause is cigarette smoking. As smoking rates decline, admissions for COPD will also decline, but not because of Part D. As a physician, it’s difficult for me to speculate why better drug coverage would lessen dehydration admissions.
Next, think about the overall effect of the severe recession starting in 2008. This recession put a temporary brake on healthcare costs, including hospitalizations. Many feel this played a major role in slowing the cost trend.
Lastly, consider that overall mortality rates didn’t fall in this group during the study period. The authors were expecting to see this. It may be plausible that better use of medications might impact mortality – the report addressed the better cost coverage, but didn’t address whether or not medication use was optimized. Medicare Part D addressed drug cost, not optimization of drug therapy. That would be an interesting study.
My main point here is “associations are not causations,” and as a nation, we can’t make pricing and policy decisions based on associations. We’ve learned this the hard way for many years and need to embrace a rigorous analytic approach when reading this type of report.
Michael L. Taylor, MD, FACP
Chief Medical Officer