Into the Extrapolation Machine

The Kaiser Family Foundation (KFF) recently released a study that showed that 42% of Americans are unaware that Obamacare (the Affordable Care Act) remains the “law of the land.” News like this seems to us, to act as a Rorschach test on how observers feel about the law. Considering 50% of Americans can’t identify New York on a map we tend not to read too much into these polls. However, according to the logic of extrapolation, since we know that the ACA remains law, we are in the elite 58% (it’s about time we made it into the elite of something).

In almost parallel to the KFF news, the New England Journal of Medicine published a follow-up study of the “Oregon experiment.” For those who haven’t been following closely, the study found that previously uninsured people who were enrolled in Medicaid did not see an improvement in clinical measures when compared to those who remained uninsured. The study did seem to show a reduction in the amount of financial distress for the insured however.

Another contentious study, another Rorschach test (example, example). The problem we see with the polarity of views is that both sides seem to be cranking up the extrapolation machine and use single studies/data points to draw broad conclusions to gin up opinions about ACA’s success or lack thereof. In light of the fact that for most practical matters ACA doesn’t really get going until 2014, use of the extrapolation noise generator approach smacks of a lack of analytical rigor in our view. We will know soon enough how the program is doing… exchanges start enrolling on 10/1.

As investors, we should state upfront that we tend to give more weight to financial returns than what the philosopher-kings might call the political context. So what caught our eye in the Oregon study was that Medicaid recipients had higher healthcare utilization rates (and associated costs) than the uninsured. The connection between gaining insured status and healthcare utilization should not come as a surprise since there is a very extensive literature elucidating this connection.

We have noticed that many in the political community, especially where it pertains to entitlements, have taken an absolutist utilization=bad view of the world. At Poliwogg we take a more nuanced approach. We view healthcare expenditures not as a cost but as the creation of an asset. In our framework, utilization can be categorized as generating either a positive return or a negative return depending on the outcome. If the expenditure creates additional value it is good, if it does not than it is bad. We would take a good pharmaceutical over a bad diagnostic test any day despite the former often being more expensive upfront.

To be fair, it is not always easy to determine a priori whether utilization will have a positive or negative consequence longer-term. And it should be noted that the system does have significant inertia hard-wired into it (prior mis-allocation of resources, a mal-incentive based reimbursement structure and asymmetry of information, among other things) which needs to be overcome. Nevertheless, there are signs that some things have changed recently to offer some glimmers of hope of a possibility of value creation and system enhancement. Like it or not, ACA is a change-agent and it will alter the status quo. The impact of improvement in technology and telecommunication, which allows for more real-time observation of events and predictive analytics, is first being felt. And, given the downward trajectory of utilization connected to the recent economic downturn it’s a pretty good supposition that healthcare costs (after many decades of rising) are finally starting to alter behavior on the part of patients and providers.

With respect to investing, uncertainly caused by media/pundit noise often depresses valuations in the short run. We suggest investors avoid the extrapolation machine and stick to observing fundamentals and fact. In the context of healthcare, as we see it, this means focusing on a double bottom line: companies with financial return potential (at the appropriate risk level) and companies who create value through innovation either by improving quality or meeting a need (especially true in the case of therapeutics for many currently poorly-treated diseases). If investors do this they can let the extrapolation machine work for them rather than exacerbate the problem.

Les Funtleyder is the managing director of Poliwogg, where this post originally appeared.

3 replies »

  1. I posted this on my REC blog in response to the Oregon “findings” –

    “I’m now a Medicare bene. 67 years old, with the usual litany of manageable (and managed) afflictions that come mostly with having lived this long. I’m a 99214, a “moderately complex established patient.”

    I see my doc twice a year. Lab pulls annually (blood and urine, the usual broad panels).

    On a couple of generic maintenance meds. You can likely guess what they are.

    I’m an uncontrolled, marginally “comparative effects research” experiment. Statistically df=3 at n-1 (4 visits in 2 years). Were I seeing my doc four times a year, n-1 df=15. For those few who took anything away from undergrad stats courses, this is “Inadequate Sample Size 101.” (See “Standard Error of the Mean,” “Power Analysis.”)

    Medicaid hating “conservatives” are all over this NEJM article. It proves that Medicaid is “ineffective.”

    What crap. The Instant Gratification Culture, Policy Wonk Division, Run Amuck.”

    Aggregating a bunch of patients to pump up the “n” is meaningless given the short time frame here.

  2. Better title for this post would have been “The Wild Extrapolation Machine.”

    This is perhaps the single best one line description of ACA I’ve come across.

    “The problem we see with the polarity of views is that both sides seem to be cranking up the extrapolation machine and use single studies/data points to draw broad conclusions to gin up opinions about ACA’s success or lack thereof”