Costs and Wasteful Care: A Physician Perspective
Michael Chernew has reviewed the policy issues involved in controlling the growth of health care expenditures. He notes the pressure to control growth came at first from the payers, who have a limited set of policy tools. Because payments for physician services has been basically flat for years, insurance companies have relied on increasing payments from patients in the form of co-pays and deductibles in hopes of making them cost aware. But this has not been working—costs in the commercial segment have been growing at 6% annually as providers seek to cover losses from the Medicare book of business. Now patients are reaching to point of revolt. So, pressures are increasing on providers to lower costs, although I would suggest we have not reached critical mass yet. As Chernew notes,
“I do not believe that the payment system will drive us to success, necessarily, but I certainly think the payment system, if we are not careful, can be an impediment to success. We need to design a payment system that will allow organizations that can find efficiencies, to have those incentives to create efficiencies. And the notion is that we share the savings. Why? Because if you won’t share the savings there will not be savings to share…
We call those payment models, commonly, value-based payment, but in my view its just a misnomer. The word “value” is simply the sugar that makes the medicine go down. This is about risk transfer. Someone has to control what goes on and someone has to have the incentives to control what is going on. Frankly, as providers, you’ve drawn the short straw.”
One obvious target is “waste.” A Pro Publica report on NPR last year cited data from several studies under the title “Unnecessary Medical Care: More Common Than You Might Imagine.” The Washington Health Alliance did a study of 1.3 million insurance claims using widely accepted measures of “unnecessary” tests and concluded more than 600,000 underwent needless tests at a cost of $282 million dollars. The National Academy of Sciences has estimated the costs of such tests as $765 billion, about a quarter of all the money spent annually on health care. In response to such studies, the Washington state medical and hospital associations noted patients often demand such tests.
How should an individual physician think about such data, much less incorporate it into practice, where the increasingly computerized “power plans” include a lot of lab testing? I suggest we should re-train with emphasis on Bayesian analysis. While lip service is given to this early in medical training, I find very few physicians really understand the essence of the analysis, despite the fact Thomas Bayes published his theorem in 1763.
In brief, non-mathematical form, all laboratory tests have true positive, true negative, false positive, and false negative rates, but in practical terms, the likelihood a given result is in one of these categories depends on the prior probability of the disease being present in the population (or person) of interest. Suppose our test is 99% specific and 99% sensitive. It turns out the likelihood a positive test is a true positive is 50% or less when the prior probability of disease is less than 10%. By the same token, the same test has a 50% likelihood of being a false negative if the prior probability is greater than 90%. And, of course, most of our diagnostic tests are neither that specific nor sensitive. Said another, way, diagnostic tests are most useful when the prior probability of disease is in the middle range.
But that is not the way we use testing—we frequently do shotgun testing and react to the results and use them to form diagnostic hypotheses, a manner designed to increase errors. We also assign a higher reliability to lab tests than to clinical observations. How often has someone been diagnosed with “congestive heart failure” because of an elevated BNP in the absence of a history of congestive findings? When the literature is reviewed, it becomes clear the BNP is most useful for excluding CHF when the value is low in a patient with dyspnea, not in diagnosing CHF in a patient where it is elevated. And even in the patient with obvious pulmonary edema, we still like to follow the BNP rather than the chest exam, particularly when trying to decide if the patient is ready for discharge.
We also don’t distinguish between screening tests and those used to monitor progression or response to therapy. Very few tests are recommended for screening, but guidelines are full of recommendations for testing for all sorts of things. But that is a discussion for another day.
Another challenge is the pressure to make a diagnosis quickly. If someone is acutely ill, that makes sense, but in someone who is suffering from chronic illnesses and has a new symptom, it may not. Primary care doctors know more than half of all symptoms go away without anything being done other than to note them and follow along.
As a busy physician, thinking about the cumulative financial impact of lab testing is very difficult, but if we can come to recognize the implications of Bayesian analysis, we will adopt a more conservative testing strategy, knowing the dangers of both false positive tests, leading to yet more studies, and false negative tests, causing us to reject starting therapy as soon as we should. I think this makes more sense than focusing on costs, because I think we want to believe we are operating in the patient’s best interests. We need to recover the parsimonious approach to diagnosis and see if we can recover from our infatuation with “the numbers.” It has and continues to be true that the patient’s story is the determining fact for what constitutes good medical care. We need to recover that idea.
28 April 2019
 Chernew M. Curbing Health Care Expenditures. NEJM Catalyst. 25 October 2018. Accessed 20 February 2019 at https://catalyst.nejm.org/videos/curbing-health-care-spending-growth/?utm_campaign=Connect%20Weekly&utm_source=hs_email&utm_medium=email&utm_content=70063354&_hsenc=p2ANqtz-_nHbozl1gdrkmNKv3tiofKohRyQib9uELdglbhbeHWcvX3EwT3fDHIQuyCEKptOJtwzktkfaqrbkKaZVsWyBanM9N1_Q&_hsmi=70063354
 Allen M. Unnecessary Medical Care: More Common Than You Might Imagine. 1 February 2018. Accessed same day at https://www.npr.org/sections/health-shots/2018/02/01/582216198/unnecessary-medical-care-more-common-than-you-might-imagine.
Population health is a phrase that disguises some hard realities as illustrated by two recent reports.
A new study suggests only 5% of Medicare spending in 2012 was preventable, much of it in frail, elderly patients. Is this good news or bad?
The Anchoring Heuristic
Businessmen and health policy experts fail to recognize the limits imposed by the experiential nature of medical practice, both of which impact achieving the "triple aim."
Thoughts on Clinical Realities