Pamela Hartzband and Jerome Groopman have a recent New York Times editorial arguing that quality incentives for doctors are corrupting medicine:
[F]inancial forces largely hidden from the public are beginning to corrupt care and undermine the bond of trust between doctors and patients. Insurers, hospital networks and regulatory groups have put in place both rewards and punishments that can powerfully influence your doctor’s decisions.
This large claim deserves extensive discussion.
In a quality-based incentive, a doctor’s medical practice is measured based on the patterns of care recorded in her patients’ medical records and evaluated against a standard of good care. For example, diabetic patients should be screened periodically for retinopathy, so this element of quality might be measured by counting how many of the doctor’s diabetic patients were screened at the appropriate times. The doctor then gets an additional payment if she meets or exceeds a recommended standard. Quality-based incentives play an important role in many health care reform schemes, including Accountable Care Organizations.
Hartzband and Groopman have four arguments for their view that pay-for-performance is corrupting medicine. Confusingly, three of them are packed into this sentence:
These metrics [used to measure the quality of doctors’ practice] are population-based and generic, and do not take into account the individual characteristics and preferences of the patient or differing expert opinions on optimal practice.
Unpacked, these arguments are that: (a) experts differ on optimal practice. (b) Individual patients vary biologically such that what is good for the typical patient may not be good for this patient. And, (c) variation among patients in risk preferences and life goals are such that what good means for one patient may not be what good means for the next patient. According to (c), even identical twins with identical diseases may have different optimal treatment strategies.
Argument (a) is not compelling. Experts’ opinions vary and so what? Many treatments based on opinion alone have proved disastrous. This is why we seek evidence-based treatment guidelines.
Argument (b), however, points to an important challenge for evidence-based medicine. Ben Djulbegovic and Gordon Guyatt argue that too many ‘evidence-based’ guidelines underlying physician quality metrics are not based on enough evidence to be trustworthy.
But surely the solution here is to develop better guidelines while weeding out the untrustworthy ones. Are all guidelines untrustworthy? I would be surprised if Hartzband, an endocrinologist, does not believe in screening diabetic patients for retinopathy.
Suppose, however, that the guideline is based on enough data to provide sound guidance for the ‘average’ patient. Hartzband and Groopman are right: the guideline will be of limited use if it fails to track important dimensions of biological variation among patients. However, high quality guidelines for complex disorders are longitudinal: they enjoin doctors to carefully assess outcomes and to adjust treatment to address individual variation in patient response. And here again, the urgently needed solution is to develop the science to produce better algorithms that personalize treatment to the unique biology of the individual.
Argument (c) is that what is optimal for an individual patient may vary depending on the preferences of that patient. This is absolutely the case, it’s important, and in other writings, Hartzband and Groopman have deep observations about doctor-patient communication and decision making. But are personal definitions of human well-being so variable that guidelines are useless? Are there, for example, significant numbers of sighted diabetic patients who are indifferent to whether they lose their vision?
I agree with Hartzband and Groopman that a doctor could experience a moral tension when a well-informed patient wants a course of treatment that deviates from a guideline and would threaten his quality incentive. However, the solution to this problem is not to eliminate the guideline. Instead, the quality measure should provide the doctor and patient a mechanism to resolve the tension in the patient’s favor. The doctor should be able to document that she discussed the standard with the patient and that the decision reflects the patient’s preferences. The patient would then be excluded from the quality calculation, as is presently done when there is a medical contraindication to the recommended treatment.
Finally, Hartzband and Groopman have an argument (d), which is that a doctor’s duty is exclusively to the patient in front of her, not the population as a whole.
When a patient asks “Is this treatment right for me?” the doctor faces a potential moral dilemma. How should he answer if the response is to his personal detriment? Some health policy experts suggest that there is no moral dilemma. They argue that it is obsolete for the doctor to approach each patient strictly as an individual; medical decisions should be made on the basis of what is best for the population as a whole.
We fear this approach can dangerously lead to “moral licensing” — the physician is able to rationalize forcing or withholding treatment, regardless of clinical judgment or patient preference, as acceptable for the good of the population.
Here Hartzband and Groopman are arguing not only with insurance companies and anonymous health policy experts, but also with the Ethics Manual of the American College of Physicians (ACP). The Manual says that
Physicians have a responsibility to practice effective and efficient health care and to use health care resources responsibly. Parsimonious care that utilizes the most efficient means to effectively diagnose a condition and treat a patient respects the need to use resources wisely and to help ensure that resources are equitably available.
What the ACP is urging, however, is not that the individual patient be deprived to benefit the population. Rather, the ACP recognizes that physicians have a duty to practice with the awareness that medical resources are finite and need to be available to all, rather than selectively allocated to the affluent. This can occur only if physicians practice parsimoniously, that is, as Jon Tilburt and Christine Cassell (and see Austin here) write,
delivering appropriate health care that fits the needs and circumstances of patients and that actively avoids wasteful care—care that does not benefit patients.
The needs of the population, on this view, are legitimately addressed by actively avoiding wasteful care. This ethic is consistent with doctors being at the same time zealous advocates for their patients’ needs. Do Hartzband and Groopman think that physicians lack the moral discernment to combine the two?
Finally, Hartzband and Groopman write with a strange insensitivity to historical context. First, they write with an urgency that suggests that pay-for-performance incentives are rapidly deforming medical ethics. But as Aaron has noted, so far the evidence suggests that pay-for-performance schemes have little effect on physician behavior.
Second, I am baffled by their claim that “financial forces are beginning to corrupt care.” Fee-for-service payment systems give physicians an obvious incentive to overtreat. On Hartzband and Groopman’s logic, prior generations of physicians must have been deeply corrupt. Yet they express no concern about the effects of those incentives. My view is that fee-for-service does affect physician behavior, sometimes in ways that harm patients. But it would be hyperbolic to call traditional physicians ‘corrupt’. Similarly, we should carefully study the effects of quality incentives on physician behavior. But we shouldn’t panic. The management of moral tension is part of every complex human endeavor.
So it seems likely that present and future physicians can negotiate the moral tensions between incentives tied to well-supported professional guidelines, the need for stewardship of scarce medical resources, and the priority of helping patients make autonomous medical choices. The cause here must be advanced by improving quality measurement, not abandoning it.
For discussion the ACP Ethics manual, see Aaron (here and here), Ezekiel Emanuel, and Paul Kelleher. See here for more TIE writing on pay-for-performance and be aware that skepticism about the effectiveness of P4P is not the same as skepticism about evidence-based medicine.