Informed consent for prostate-specific antigen (PSA) testing? No. No, but the issues are complex, and the idea cannot be dismissed casually. In the end, informed consent in this instance would benefit neither the medical provider nor the patient. My negative response relies on a few assumptions or definitions that I should explain. First, by informed consent, we usually mean written, as opposed to verbal consent. We explicitly mean a deliberate, formal process that culminates in a signature indicating consent. Second, we need to make a distinction between informal consent in the clinical as opposed to the research context. In the clinical context—where PSA belongs—informed consent is essentially a contract between provider and patient in which the benefit for the provider is some degree of protection from the accusation of wrongdoing, and for the patient the benefits are a guarantee of information and preservation of personal rights and autonomy (considered a value in our culture). In the research context, the contract is similar, but the patient is being offered a procedure that, by definition, does not have established clinical benefit, and this radically changes the risk/benefit equation. We require (written) informed consent in the clinical setting when the physician recommends a procedure with enough risk that the physician would like, or needs to have, some medicolegal protection in case of an adverse event, and—from the patient’s point of view—enough risk is present to consider the option of refusing the procedure and therefore a need to receive enough information about the risks and benefits to make the decision. It is important to note that the notion of “sufficient risk” implies a continuum involving subjective judgments and perceptions by all parties concerned. For example, we perceive auscultation of the heart in an asymptomatic pa-tient by a primary care physician as a procedure that does not require informed consent, even though the predictive value might be very low and the procedure conceivably could do more harm than good for the patient.
The risk of a procedure is a combination of the probability of an adverse event and its severity. For PSA testing, there are three types of risk: two common, but relatively less severe ones—anxiety and unnecessary biopsy (which entails cost, risk, and discomfort); and one less common but very severe injury—unnecessary treatment. Interestingly, unnecessary treatment is a matter for lawsuits in many situations (eg, the patient requiring a gallstone removal who receives a leg amputation), but, in this situation, unnecessary treatment for prostate cancer cannot be a matter for a lawsuit, because unnecessary treatment can rarely or never be identified in an individual case. Performing radical prostatectomy on an 87-year-old man with end-stage heart disease and dementia might possibly be actionable, but even in the unlikely event that a surgeon is willing to do this and the patient or his family are in favor of it, third-party payors would certainly have something to say. Overzealous or unnecessary intervention is a real threat in our affluent medical system, but patients are not accustomed to thinking of excessive treatment in itself as an adverse event, unless there was clearly no indication to do the procedure and something goes obviously wrong during the procedure. Because there is little risk of action against the provider for doing unnecessary treatment for prostate cancer, the benefit to the provider of informed consent for PSA testing is minimal.
For the patient, the potential benefits of informed consent are different, and the right to decline exposure to unnecessary treatment could be very important. Indeed, a recently published analysis estimated that PSA testing results in overdiagnosis of prostate cancer (meaning cancers diagnosed by screening that would otherwise never have been detected during the patient’s life) by 29% in white men and 44% in black men.1 This should be expressed to the patient contemplating screening as the absolute risk of unnecessary treatment (eg, 2 in 1000) relative to the absolute risk (ie, probability) of having an extended life. In fact, a 50-year-old man with an average prostate cancer risk who is thinking about undergoing initial screening should know the probability of experiencing any of the risks (anxiety, unnecessary biopsy, unnecessary treatment) of PSA screening, not just for a single initial screen, but for an entire program of serial testing between the age 50 and approximately 70 years. These risks are cumulative with each succeeding test, although the risk of harm in later years might be reduced conditional on negative preceding tests. It was recently calculated that the average woman undergoing nine serial mammograms at recommended intervals has a 43% chance of eventually experiencing a false-positive test with its associated reductions in the quality of life.2 We have some data on the cumulative risk of a false-positive result with annual PSA testing, but only in selected populations and thus not enough to comfortably extrapolate to many real-world situations.3
Assuming that we could estimate the cumulative risk of a false-positive test that would inevitably cause emotional distress, we would still need to estimate the probability that this will lead to an unnecessary biopsy. At the moment, this depends on whether and how the clinician uses information in addition to the total PSA level, such as free PSA, patient age, PSA velocity, and gland volume. Finally, we have to estimate the risks involved in attempts at cure, and the benefits in terms of longer or higher quality of life. Needless to say, we are still a ways off from being able to estimate many of these important risks and benefits for the patient. One could say that not every positive biopsy has to lead to invasive therapy, so an agreement to undergo screening is not equivalent to acceptance of surgery or radiation. However, we cannot yet identify early-stage cancers that will not progress with or without treatment. So the first problem with informed consent for the patient is that we do not know enough to really inform him to the same extent as in other written consent situations—the situation is far more complex than relating the complication risks of a single surgical procedure.
The second problem with informed consent for the patient is that the current medical system environment simply does not support doing it properly. Third-party payors do not reimburse physicians for spending time explaining complex uncertainties to patients, physicians are not adequately trained to do this type of counseling, and even those who are would have difficulty finding the time to do it in a typical, busy practice.
Currently, we do not require informed consent for cancer screening procedures that have proved benefit, such as mammography, colonoscopy, and Papanicolaou testing. I would argue, along with others, that PSA testing also has benefit, and that it has already contributed to a decline in prostate cancer mortality in the United States.4 The real question now is whether the risks of any of these screening tests outweigh the benefits, and it is vital for us to find ways to reduce the risks of false-positive tests (including the simple solution of repeating PSA testing at longer intervals) and unnecessary treatment. Two of the rare situations in which we require informed consent for a noninvasive diagnostic test are human immunodeficiency virus testing and testing for most high-penetrance genetic mutations. Informed consent is perceived to be necessary for these tests in part because of the threat of discrimination against patients who test positive. Interestingly, human immunodeficiency virus, genetic mutation, and PSA testing are similar—and distinct from tests such as mammography and colonoscopy—in that they can be done without the patient being aware of the testing. I would argue that this possibility increases the need for clinicians to inform patients before testing, but does not mean that informed consent is required per se in every case, for the reasons given above.
In conclusion, informed consent for PSA screening will impose additional regulation on the delivery of medical care without doing much, if any, good. It will provide the doctor with legal protection that is not needed, while failing to provide the patient with information and autonomy that he certainly does need. Instead of looking for a quick fix, we should work on fixing what is wrong with doctor-patient communication in our current medical care system, a system that is characterized by over-stressed providers, lack of continuity, excessive legal interference, and distorted financial incentives.
FULL TEXT
