Wellsteps claims to have dramatically reduced the total cost of the Boise School District’s health spending. Their Koop Committee colleagues gave them an award for it, as they typically do for their fellow board members and sponsors. (Yes, Wellsteps’ CEO, Steve Aldana is on the committee that grants the award, but, in accordance with Koop Committee tradition and Wellsteps ethical standards, there is no mention of this possible conflict of interest in the announcement from Wellsteps. This is not a violation of the Wellness Industry Code of Ethics, because there is none.)
Unfortunately for Wellsteps, three completely distinct observations from Wellsteps’ own data invalidate their analysis separately. In combination, these observations create a level of impossibility demonstrating that whoever invented the English language was unfamiliar with the wellness industry, or they would have come up with a word meaning: “Impossible doesn’t begin to describe it.”
First, as we saw in the last posting, employee health deteriorated over the course of the program, measured both subjectively and objectively.
There is a concept, covered in Health Services Research 101, called “causation.” Wellness vendors love taking credit for everything that happens during their program. Hence we can conclude that Wellsteps’ program caused employee health to decline.
Therefore, no reduction in costs due to a healthier employee population can be attributed to Wellsteps’ program.
Second, also covered in Health Services Research 101, is the concept of “fifth-grade arithmetic.” Potentially Preventable Hospitalizations (PPH), as described in the official Health Enhancement Research Organization (HERO) outcomes measurement guidelines (a report on which Wellsteps’ CEO claims to have collaborated), account for a very small percentage of all spending. Specifically, as described in the HERO report and reproduced below, these events account for 2.62 hospitalizations per 1000. Boise’s 3284 employees would therefore suffer about 9 PPH’s. If each PPH cost Boise the HERO-assumed cost/admission of $22,500, that’s about $202,000. Not enough to cover even the out-of-pocket fees for the Wellsteps program, assuming the Wellsteps did a perfect job. And as we’ve learned in the past, a beam of light leaving “perfect” wouldn’t reach Wellsteps for several seconds.
Still, we’ll never know because despite their endorsement of the HERO report, Wellsteps decided not to disclose the rates of Boise’s PPH’s, likely to obscure the fact that they didn’t reduce it.
And as the HERO guidebook says, any reduction in PPH’s is offset by more spending elsewhere. In Wellsteps’ case, “more spending elsewhere” is annual biometric screenings. Curiously, Wellsteps admitted annual screenings are a stupid idea four whole days before they announced their Koop Award for doing exactly the opposite.
Yet Wellsteps reported annual savings ultimately exceeding $5-million, or about a third of Boise’s total spending. This would be equivalent to wiping out every hospitalization unconnected with childbirth plus every ER visit. Adding yet another layer of impossibility to this narrative, at the end of this, we show the rank order of the top 25 reasons people visited the ER or went to the hospital, which basically have nothing to do with corporate wellness programs. Hence the cost of these visits and admissions couldn’t be dented, let along wiped out, by massive overscreening or even by appropriate screening.
Compounding this impossible outcome (and wellness is one of the few industries in which there are degrees of impossibility, since a typical wellness vendor makes at least a dozen impossible claims before breakfast) is the surprising health of the Boise population to begin with. People rated their health as 7.98 out of 10, and only 2.5% reporting smoking (vs. 20% for Idaho as a whole) and only 20% reported drinking (vs. 70% for the US as a whole). These ridiculously low levels did not strike Wellsteps as suspicious, so we will assume they are accurate.
Further, most Boise employees had fairly normal blood pressure, glucose and cholesterol — at least before they got sucked into the program.
So, with impossibly smoking and drinking, excellent reported health status, and largely normal biometrics, how is there room to improve health enough to save money, especially when health status is declining and the population is being overdiagnosed?
Speaking of misunderstanding the concept of arithmetic, third and most important is the data Wellsteps suppressed between their initial report and their Koop Award application. Normally Koop Award Committee Chairman Ron Goetzel suppresses the invalidating data after the award is announced, as with Health Fitness Corp/Eastman Chemical, and then the state of Nebraska (technically actually incriminating, not just invalidating). That’s because in the past I haven’t predicted who would win the award. Rather I just pointed out all the obviously incorrect data after the fact. By predicting Wellsteps would win a Koop Award and pointing out exactly why their data was sufficiently fictitious to merit it, I gave Wellsteps itself the opportunity to suppress their own data, so Ron wouldn’t have to do it for them.
Contrast below the trend they reported for total spending for Boise against the claims cost per person for Boise. The former goes up. The latter goes down even though the number of employees stays the same. Obviously, this is an impossible coexistence, as mentioned both in the first installment of this series and in every elementary school in the world. This time, we are going to transpose the bar graph, which separated participants from non-participants, onto the line graph, which included both cohorts, for the baseline and the first two years of the program. The assumption, as Wellsteps states, is that their participation rate is 80%, largely because of the massive $870/year incentive, which in classic wellness fashion is not included in the savings calculation. The number of employees seems to bounce around a bit between reports, but we’ll go with 3284 for this valuation.
By way of review from the first installment, here is total spending:
Also by way of review, here it total spending, participants vs. non-participants, going exactly the other way:
Now let’s overlay the second set of figures onto the first. In addition to trending in opposite directions during the wellness program years, the total spending on these dueling slides doesn’t even coincide in the baseline year. Since the whole point of the exercise is to look at the trend subsequent to the baseline, we will add about $3-million to each year for the bar graph figures, so that the 2011 starting points coincide. That let’s us focus on the difference-of-differences in the program years. Starting at the baseline, costs increased about $1.7-million by 2013, putting the 2013 red endpoint almost exactly in line with their prediction.
In all fairness, let’s continue the analysis, by looking at the Wellsteps performance in 2014:
As you can see, costs did fall below the “prediction” in 2014, though still way above “Wellsteps Begins” and “Actual.” The only problem? More than 100% of the entire decline from the previous year was due to non-participants’ costs plummeting, while participant costs increased.
So where do we stand? Wellsteps, by their own admission, overscreened and overdiagnosed this population. And their own admission, the population was quite healthy to begin with. By their own admission, the health of the Boise employees deteriorated. And, by their own admission, the only successful annual performance is attributable to a large improvement in non-participants.
To paraphrase the immortal words of the great philosopher Samuel Goldwyn, the title of this post, Wellsteps has raised the Koop Award standard for outcomes invalidity to a new low.
Top hospital discharge codes:
Top ER visit codes: