Should Doctors Protect Us from Data about Medical Risks?
| June 22, 2011
Sara had a pain in her side that she attributed to using a new ab machine at the gym. But over the next couple of days, the pain increased and made her short of breath.' On the third day, she consulted her primary care doctor, who examined her and found nothing untoward.' But he recommended that she go to the Emergency Department to get the pain checked out.' At the ED, she had a blood test and a chest x-ray, which were both normal.' 'Do you want a CT scan?' she was asked by an ED physician.' She replied, 'Well I've already been here almost three hours.' I might as well.'
The CT scan showed a pulmonary embolism.
What I find interesting about this story is not the near miss of a catastrophy but rather the fact that Sara was asked to make the decision about whether to have a CT scan or not.' ' She was given no information about possible diagnoses; there was no discussion of guidelines; no risk statistics were provided; the price of the scan was never mentioned.' It was left to her to choose, based on what?' Her schedule for the afternoon?Sara's was the third story about patient choice I've heard about in the past three days.' 'Do you want to take a statin?' (in response to a high LDL reading) and 'Do you want a chest x-ray' (in response to a persistent, violent 8-week-old cough).' No risk information provided.' No trade-offs explained.' Just the question.
Such accounts are increasingly common from people describing their experiences with health care.' Many are asked to make decisions about their treatment with little or no objective information to guide them. ' Others talk about how they have unwittingly made important decisions simply by their choice of physician; for example, an oncologist versus a surgeon. Only in hindsight realizing that this determined the course of their treatment.
Does asking patients what they want reflect how some clinicians understand shared decision-making? Is it an attempt to improve patient satisfaction?' Is it some weird expression of support for patient autonomy? I don't know.' I do know, however, that many people report being faced with health care-related decisions they feel unprepared to make.
It was thus interesting to read a recent paper from the Hastings Center that challenges the proposed requirement ' ' advocated by proponents of shared decision-making ' ' that we should always be given evidence-based risk information when we make decisions about preference-sensitive care.[1] Peter Schwartz, a physician and ethicist, argues against what he calls 'the quantitative imperative' on the grounds that most of us struggle with innumeracy and many of us are subject to cognitive biases that might lead us to make irrational decisions. His critique is consistent with a more general argument against 'mandatory' (as opposed to 'optional') autonomy[2].' Rather than require that such information be provided to all, Schwartz recommends that clinicians judge each individual's interest and readiness to make use of a decision aid that arrays numerical risks and trade-offs and offer aids only to those who ask for one or whom the clinician believes might make good use of such information.
I appreciate Dr. Schwartz's recognition of the burden placed on us by having to incorporate difficult-to-grasp risk information into our decisions about health care.' He's right; it's not easy.' And I welcome any expression of concern ' however tangential ' about the shift of responsibilities from clinicians to patients who are often ill-prepared to fulfill them.' But his general argument seems overly protective in light of the rushed, confusing demands of health care today.
Dr. Schwartz elegantly describes the general innumeracy of the American public with a particular emphasis on our inability to understand the difference between relative and absolute risk.' He neglects to mention that we share this deficit with many of our clinicians.' Ah, the blind leading the blind.
He also does a nice job of summarizing the relevant literature on cognitive heuristics but again exempts clinicians from discussion.' Does he believe that clinicians are immune from these same biases such that they could, for example, ' judge whether I can make good use of a decision aid and whether providing one will lead to a rational decision based on a 15-minute consultation with me? What looks rational to my doctor may be a quick resolution to a cumbersome process, a fast buck or a good outcome. How does this affect my doctor's judgment of my capacities?
More puzzling is the importance Schwartz assigns to risk information in the decisions we make about our health and health care, preference-sensitive or not.' Empirical information is always only one of many factors that influence our choices. In fact, it is often inconsequential relative to the burden most people face in finding and using health care even if they understand it well. There is little relevant quantitative information available for most of the health care decisions we make now. ' We are winging it based on anxiety, our neighbor's experience and our sense of what the right choice is today, which can be influenced by the mood and experience of our doctor just as easily as it can by the state of family and work events taking place outside the exam room.
I have a good grasp of relative versus absolute risk, and as a person who has been treated for four different cancers, I have made many decisions about tests and treatments, both with and without the help of my clinicians. I always struggle to understand how to apply population statistics to myself when they are available. I always think that whatever decision I make at the time is rational, the best one I can make, given what I know and feel. And I have often ' a week, a month, a year later ' looked back and thought 'What could I have been thinking?'' Are these decisions rational or irrational?
It seems unconscionable to withhold well-presented risk information about treatment in the relatively few instances when it is available for preference-sensitive care, as it is for treating a herniated disc and early stage breast cancer..' In their most general form, decision tools communicate that we have choices ' that multiple treatment options are possible, there are trade-offs to be considered, and no guarantees exist regardless of our choice.' These are sobering but important messages for us to grasp as all of us, regardless of our numeracy skills and cognitive biases, are routinely forced to make critical decisions about our health care.' I would trade any day the harm imagined by Schwartz' of over- or under-estimating the risk element in my personal algorithm of choice for a) making a decision based on my gut feeling alone, or b) having my clinician make that decision for me based on her assessment of my preferences and capabilities.Schwartz's argument against mandatory autonomy and the provision of risk information to all comers is an academic exercise:' We already have mandatory autonomy by default, as illustrated by Sara's encounter in the Emergency Department and the two other experiences of everyday patient decision making above. Most of us make those choices with little guidance and support.' To argue that our clinicians should selectively protect some of us from empirical risk information is a wasteful distraction from the larger challenges we face in getting the most out of our health care as part of our effort to live for as well and as long as we can.
[1] Preference-sensitive care is 'the treatment of conditions for which two or more valid treatment choices are available for most patients, e.g., treatment for herniated disk, early stage breast cancer; whether a' PSA test or screening for colon cancer is right for an individual
[2]Mandatory autonomy holds that 'patients have a duty to themselves, to society and to the medical system to make decisions on their health care independently' in contrast to optional autonomy in which clinicians 'seek to ascertain patients' decisional preferences and to understand wider limitations on their freedom to choose.
