"Why Nobody Believes the Numbers" - Reviewing a New Book on Disease Management Results

Submitted by jonpearce on Sun, 2012-10-14 23:41

In the current health care environment (or what might be termed as the "second coming" of managed care), disease management initiatives designed to reduce healthcare costs are springing up in record numbers, claiming high levels of effectiveness in reducing healthcare costs. Casual readers of these articles may come to the conclusion that implementation of several of these programs together could turn the trend of healthcare inflation into negative territory.

Al Lewis isn't a casual reader, however. He's a national leader in disease management healthcare outcomes analysis, with impeccable credentials to understand and evaluate these types of initiatives. His new book, "Why Nobody Believes the Numbers – Distinguishing Fact from Fiction in Population Health Management" reviews the math behind some of the more outlandish overstatements of cost effectiveness of recent disease management initiatives, exposes consistent flaws in the methodologies used in these calculations, and attempts to drag the reader into an understanding of the realities of cost effectiveness of these programs. With chapters having such titles as "Actuaries Behaving Badly" and "Case Studies that Flunk Every Plausibility Test Known to Mankind" (CSTFEPTKTM), Mr. Lewis pulls no punches in criticizing methodologies that he believes deceptively overstate effectiveness of such programs.

In "Actuaries Behaving Badly" Mr. Lewis deals with issues regarding separating populations into a "pre" group that was not affected by disease management program, and a "post" group that participated in the program. He points out the mathematical problems in trending forward healthcare costs among groups and using the trended amounts as the metric against which actual results are measured. He also notes that disease management programs will necessarily increase costs in certain areas (for example, for drugs used to treat chronic conditions), and that for results to be valid these costs must be increasing as the targeted costs decrease. If these cost of implementing the program do not increase, then the cost decreases must not have been the result of the program.

In the "Plausibility Testing" chapter, Mr. Lewis shows the relationship between the relatively narrow cost and utilization elements involved in the disease management program as they compare to broader health care costs and trends. He notes that it's invalid, for example, to attribute the decrease in chronic care costs to a specific program if the same costs have been similarly decreasing in settings where the disease management programs not present. He gives a detailed description of how to perform a plausibility test that is arithmetically correct.

The CSTFEPTKTM chapter introduces the "50% savings rule" which states that "in a voluntary health management program with no incentives, declines in excess of 50% in any given resource category are the result of invalidity, not effectiveness". This criterion should be applied to each component utilized to calculate savings. For example, Mr. Lewis describes a result reporting a 19.0% net savings cost that was attributed to a chronic disease management program. However, to achieve that level of savings, a reduction in hospitalizations of more than 50% would be necessary since these programs achieve savings primarily through admission management. Since achievement of a 50% reduction in hospitalizations in any category is considered to be impossible, the results of this analysis should be questioned.

Similar issues are discussed in several other chapters of the book. But the book ends with several "how to do it right" chapters, including descriptions of "Case Studies of Where, When and How Wellness Programs Have Actually Worked", and the final chapter of "Contracting/RFP Checklist of Do's and Don'ts (Mostly Don'ts)".  These chapters apply the test described previously and show how some programs can create cost savings that don't violate plausibility and mathematical rules.

This reviewer is not an expert in disease management or statistics, so I can’t opine on the accuracy of the books statements (although Mr. Lewis offers a $10,000 reward to anyone finding inaccuracies in the book). However, it provides fascinating reading for reader such as myself who may believe that these claims of cost effectiveness appear unreasonable, but who has insufficient knowledge to challenge them. It also provides an interesting analytic structures to use in areas other than disease management.