Avoiding Low-Value Care
Panel: Atul A. Gawande, M.D., M.P.H., Carrie H. Colla, Ph.D., Scott D. Halpern, M.D., Ph.D., M.Bioethics, and Bruce E. Landon, M.D., M.B.A., M.Sc.
The New England Journal of Medicine, April 3, 2014
(Excerpts here are not in continuity.)
Atul Gawande: When a service is widely recognized as providing little or no benefit, or maybe even harm, what should be done to reduce its use?
I want to start with a seemingly simple question. How do you define low-value care?
Bruce Landon: It turns out that there are very, very few services that are low-value in all clinical situations. So I actually think it’s really important to use a clinical lens, and particularly for, as I’m taking care of patients, it’s a lens that really focuses on that patient in front of me.
Scott Halpern: Well, I think it’s incredibly difficult to draw lines, particularly because all our definitions are predicated at the population level. And population-level estimates don’t apply very naturally to individual patients.
So when we describe something as low-value, I think what we’re typically trying to do is to distinguish it from something that is no-value. But it turns out that no-value interventions, first of all, are probably very few and far between. There are very few things we do in medicine that truly could not help any patient to which we might consider applying it.
Landon: I think it’s going to be hard to address the problem of low-value care by having payers and policymakers make rules, because there’s this clinical heterogeneity story.
Halpern: The right rates at which we utilize these quote-unquote low-value services is not zero. We don’t want to practice so frugally that we’re missing opportunities to provide benefits to patients by not intervening. So I think at some level, physicians should be comfortable that they can make clearly well-thought-out choices that, although there are recommendations not to do things for the overwhelming majority of patients who fit a particular description, that there may be exceptions where the service is in fact a reasonable choice.
Carrie Colla: I think that there’s a danger that blunt payment instruments will reduce the high-value care as well, and so I think to some extent that’s why thinking about it at a broader level while also monitoring outcomes, but thinking about it at a broader level in terms of payments makes more sense.
Gawande: What would you say that the policies of the government ought to be, or of insurers ought to be, in order to make sure decision making more effective for both patients and physicians under these circumstances?
Halpern: If we had one health insurance coverage system, all the prices would be a lot easier to keep track of, for physicians and patients alike. And it would be much easier to have a set menu at the bedside as these conversations are unfolding, of all the types of information that we would want. I recognize that may be a long way away. But it is one of the sort of unintended consequences of our variegated reimbursement system as it exists today.
Video (30 minute) and link to transcript: http://www.nejm.org/doi/full/10.1056/NEJMp1401245?query=TOC
Comment:
By Don McCanne, MD
Diagnostic and therapeutic interventions that are of low value remain a dilemma. In this age with an emphasis on containing costs, should interventions that have a high cost in relation to an anticipated minimal or negligible benefit be avoided simply to help “bend the cost curve”? Or should such interventions be offered since even the smallest potential benefit should not be withheld from the patient if the patient desires such?
The easiest decision to be made would be about interventions that clearly provide no benefit under any circumstances, and may even potentially inflict harm. This is not low-value care, but rather it is no-value care. Obviously such interventions should be abandoned. For the few health care professionals using them who fail to respond to educational processes, discipline should be considered.
What about interventions that have a significant risk of major harm but could provide a small benefit that is not commensurate with the potential harm? Clinical judgement begins to enter here, but it would be a rare circumstance where other factors may warrant proceeding with the intervention.
What about the intervention that is very expensive but potentially provides only minimal benefit? Although some might use measures such as anticipated increase in quality-adjusted life years (QALY), there are levels of spending that common sense tells you are far beyond the value of the potential benefit. Rejecting such interventions risks being labeled as rationing, but such a charge does not mean that common sense should be abandoned.
A variation of this category would be lucrative procedures in widespread use for which only a paucity of conflicting data provides a rationale for these practices – high cost but low benefit. Sometimes these correlate with excess capacity in the system, a problem that separate budgeting of capital improvements could improve. Also, administered pricing could lower payments to more closely match the extent of the benefit.
What about the expensive diagnostic intervention that has a very low probability of of turning up a disorder for which therapeutic interventions could be of great benefit, perhaps even life saving? This is where clinical judgement and being sure that the patient is well informed play a crucial role. This is also where those citing the Dartmouth studies hope to reduce health care spending. But if a low-yield test has a real chance of leading to an intervention of potentially great benefit, then the payer should not intervene.
A frequent criticism is that such low-yield tests are done too often to reduce the risk of a malpractice lawsuit, and that we could reduce the costs of malpractice if we did away with these “unnecessary” tests. Since such tests are low yield, frequently nothing significant is found and therefore no lawsuit was prevented. But the judgement should not be based on the cost per lawsuit prevented, but rather on the clinical benefit to the patient. This is why attacking low-yield tests is not a productive way of reducing malpractice costs.
What about the patient who demands an intervention when it is clear that there is no value in what the patient wants? It is the health care professional’s responsibility to inform the patient why such an intervention should not be entertained. Most patients will appreciate informed advice. For the rare ones that do not, physicians should never conspire with a patient to do wrong, even if it results in the patient seeking care elsewhere.
In all of these situations, the interests of the patient must come first. Clinical judgement is required for most of them.
So how do we address the costs? The current leading approaches are to erect financial barriers to care and to impair access by using narrow provider networks. These interventions are inappropriate because they save costs by preventing the patients from receiving appropriate care that they should have.
There is a far better method of reducing inappropriate spending, and that would be to enact a single payer system – an improved Medicare for all. Such a model dramatically reduces administrative waste and improves pricing of health care services. Under a single payer system it is much easier to match payment with value.