The case study may no longer be on Pharos’ Innovations website by the time you read this, and for very good reason — the national health plan which is named may have had them remove it.

Note a few items. First, CHF admissions declined 79%. Most epidemiologists would call this result impossible to achieve — if not in general than certainly as a result of a few phone calls and monitoring. But I am not an epidemiologist. I don’t even play one on TV. I did, however, complete 8th grade math, which is all the education one needs to prove that this Pharos result, whatever the merits of their claim about admissions reduction, would require non-hospital costs to decline about 87%.

How? Well, total costs declined 85%. This means that — assuming the average avoided admission costs the same as an average admission — hospital costs declined 79%. For total costs to decline 85%, non-hospital costs would have to decline by noticeably more than 85%. If hospital costs are 75% of total costs, for example, non-hospital costs would have to decline about 87% to average an 85% cost reduction.

In disease management, the point is to substitute other resources for hospitalizations: compliance, preventive doctor visits, more frequent lab testing etc. If you listen to phone calls made by disease management nurses, a major theme is advising people to take their drugs regularly, and check in with their physicians more often. These two activities — taking more drugs and seeing one’s doctor more — cost money.

As a result, non-hospital costs do generally rise. If they fall, they do so by small amounts, not 87%.

Now here’s where the math gets interesting. Let us relax the assumption that a 79% decline in admissions equates to a 79% decline in hospital costs. Instead, replace it with an assumption that admissions which are preventable by a few phone calls are less costly than admissions that are not preventable starting 30 days after the program begins, like GI surgeries, phlebitis, infections, transplants, etc. So a 79% admission reduction in low-acuity admissions may generate a hospital cost savings of, for instance, 50%. Using that more realistic assumption, non-hospital costs would have to fall by more than 100% to generate an 85% reduction in overall costs.

While we are on the subject, note another biostatistical quirk that sheds light on the credibility of the vendor and the outcome:

The two months with the lowest admissions were the first two measured months of the program, February and March. In those months, admissions fell to about a level equivalent to about half the admission rate for the average Medicare member, even though these CHF members are much sicker than the average Medicare member. As the cohort continued in the program after that, the admission rate rose quite substantially in future months. This is the opposite of most disease management programs, where it takes a while to get patients to change their eating, exercise and compliance habits.

I am leaving my opinions out of this posting due to the possibility of a lawsuit (to which I say, “Go ahead. Make my day”), and so will leave the opinions up to the reader. This is just the facts, ma’am. (Dum-de-dum-dum.)

Next Intelligent Design Award