For Prof Tutor Only (others will be deleted)

4 H A S T I N G S C E N T E R R E P O R T January-February 2012

To the Editor: The traditional “in- formed consent” process for medical treatment is badly broken. As patients face fateful medical decisions, they of- ten do not know the basic “gist” of their options or the likelihood of the possible outcomes, good and bad. The shared decision-making movement aims to im- prove that sorry situation by having pa- tients and clinicians work more closely together when there is more than one reasonable medical option, as is the case for many if not most situations. By en- suring both that patients are informed about their choices and that clinicians are informed about patient preferences, the quality of medical decisions should be improved. Patient decision aids are not themselves shared decision-making; instead, they are tools to help make shared decision-making practical in the busy world of clinical medicine.

In his provocative article (“Question- ing the Quantitative Imperative: Deci- sion Aids, Prevention, and the Ethics of Disclosure,” March-April 2011), Peter Schwartz seems to acknowledge the need to improve the current informed consent process but worries that pro- viding quantitative information about possible outcomes as part of a shared decision-making process might be mis- leading to or unwanted by patients. He proposes the “default” option should be to withhold this quantitative informa- tion unless and until a patient asks for it. His primary arguments are that many patients have poor numeracy skills and might not understand—or might even be misled by—the quantitative facts, and that outcomes have not been shown to be better as a result of providing pa- tients with quantitative information.

Problems with statistical numeracy— which have been shown to be an issue

for clinicians, too—do make it more challenging to communicate risks and benefits. However, just because a task is difficult doesn’t mean it shouldn’t be done. Most people in the United States don’t want to play a passive role in their health care, and the proportion isn’t conditioned by assessed numeracy (M. Galesic and R. Garcia-Retamero, “Do Low-Numeracy People Avoid Shared Decision Making?” Health Psychology 30 [2011]: 336-41). Moreover, as noted in Peter Ubel’s companion editorial (“The Experimental Imperative”), research on the communication of quantitative health information is revealing that new strategies and technologies can help overcome these barriers. For example,

visual aids, commonly used in decision aids, can help even people with few nu- meracy skills better understand health statistics (R. Garcia-Retamero and M. Galesic, “Who Profits from Visual Aids? Overcoming Challenges in People’s Understanding of Risks,” Social Science Medicine 70 [2010]: 1019-25). Many of the interventions for which informed consent is necessary in health care, such as open-heart surgery and organ trans- plantation, had to overcome numerous barriers to enter the medical armamen- tarium. It’s time to bring a similar in- tensity of effort to overcome the barriers of literacy and numeracy in communi- cating important health information to the people who must live with the con- sequences of their health decisions.

Holding informed consent to an outcome standard is an interesting ar- gument. Schwartz acknowledges that decision aids that present quantitative outcome probabilities have been shown to give patients more accurate percep- tions of their health risks. Is putting the “informed” in informed consent not an important goal in itself? Would the ethical and legal imperative of informed consent hold up under scrutiny in a trial where patients facing surgery were randomized to an informed consent process versus none? Would the whole notion of informed consent be scrapped if functional status scores were no differ- ent in a second trial? I think not.

The key issue here is whether the de-

fault option in informed consent should be withholding quantitative informa- tion unless patients ask for it or provid- ing it unless they say they don’t want it. For too long, the medical system has kept patients largely in the dark about what clinicians have planned for them. Given this history, perhaps it’s time to make giving more information—in- cluding quantitative information—the default position, and to work much harder at doing it well.

Michael J. Barry The Foundation for Informed

Medical Decision Making

To the Editor: Every day people tell me about the challenges they face in finding safe, decent health care and making the most of it. Facing tough

letters

Too Much Information?

It’s time to overcome the barriers of literacy and

numeracy in communicating important health

information to the people who must live with

the consequences of their health decisions.

DOI: 10.1002/HAST.4

January-February 2012 H A S T I N G S C E N T E R R E P O R T 5

oN tHe WeB n Bioethics forum http://www.bioethicsforum.org

Administration Reveals Lack of CLASS By Peter S. Arno, Michael K. Gusmano, and Deborah Viola Just as the baby boomers are entering retirement, the first real step toward a national long-term care policy in forty-five years has been cast asunder.

What is Humane? A Plea for Plain Language in the Debates on Animal Experimentation By Joel Marks “Humane” is implicitly defined as meeting accepted standards of care and use according to legal and institutional guidelines. What this means in practice is that anything can be done to a laboratory animal, provided it is necessary to carry out an experiment or other procedure that a committee has deemed worthy on scientific and perhaps humanitarian grounds. I submit that this is an illegitimate use of the term “humane.”

Also: michael K. Gusmano points out misinformation about U.S. poverty; Carol Levine suggests that we’d all better start saving now for our babies’ future long-term care; franklin G. miller and robert d. truog explore uncertain- ties prompted by the latest research on patients in the vegetative state; and Cameron Waldman visits Zuccotti Park to talk with clinicians occupying Wall Street.

n the health Care Cost monitor http://healthcarecostmonitor. thehastingscenter.org

Global Competitiveness: How Other Countries Win By Daniel Callahan and Elizabeth H. Bradley Nearly every country that leads the world in international economic competitive- ness also has a strong government-run or regulated universal health care system and a comprehensive welfare policy. The one exception is the United States.

decisions about treatments and tests with little or no objective information to guide them often tops their list of dif- ficulties. I have frequently experienced this myself as I recover from treatment of my fourth cancer diagnosis.

I appreciate Peter Schwartz’s recog- nition of the burden placed on patients and loved ones to incorporate complex risk information into decisions about our care. And I welcome any con- cern—however tangential—about the shift of responsibilities from clinicians to patients, who are often ill-prepared to fulfill them. However, I find his ar- gument overly protective in light of the rushed, confusing demands of health care today, increased public access to health information, and the shared decision-making policies imbedded in the Affordable Care Act.

Schwartz describes the general innu- meracy of the American public with a particular emphasis on our inability to understand the difference between rel- ative and absolute risk. Yet he neglects to mention that we share this deficit with many of our clinicians. He also summarizes the literature on cognitive heuristics but again exempts clinicians from discussion. Does he believe that clinicians are immune from these same biases? My doctor might withhold a decision aid because she doesn’t have time for this cumbersome shared deci- sion-making nonsense or she believes she knows what I should do to realize the best outcome. She would view each of these as rational choices.

More puzzling is the importance Schwartz assigns to risk information in decisions we make about our health care, preference-sensitive or not. Em- pirical information is always only one of many factors that influence our choic- es. Scant relevant risk information is available for most of the health care decisions we make now. We just wing it, based on anxiety, our neighbor’s experience, and our sense of what the right choice is today, which can be in-

fluenced by our doctor’s mood as easily as it can by family and work events.

It is thus oddly shortsighted for Schwartz to recommend withholding decision aids for some patients (based on the clinician’s assessment of our competence) in the relatively few in- stances where these aids are available for preference-sensitive care, as they are for decisions about early-stage breast cancer treatment or getting a prostate- specific antigen test. Even if we don’t fully understand what’s at stake, well- presented risk information powerfully communicates that we have choices: that multiple treatment options are possible, that there are trade-offs to be considered, and that no guarantees ex- ist, regardless of our choice. These are sobering but important messages for us to grasp as we, regardless of our nu- meracy skills and cognitive biases, are routinely forced to make critical deci- sions about our health care.

It’s too late to argue that our clini- cians should selectively provide de- cision aids: the ACA provisions for shared decision-making will likely eventually tie clinician reimbursement to providing them. And the impetus for that argument—that providing this in- formation imposes mandatory autono- my—is akin to discussing the benefits of closing the barn door after the cows have wandered away: Our autonomy is already mandated by default.

Jessie Gruman Center for Advancing Health

To the Editor: In his article, Peter Schwartz eloquently discusses the ben- efits and potential harms of providing patients with numeric risk informa- tion. He describes how—despite our best efforts to inform patients about the risks and benefits of screening tests and preventive treatments and to im- prove understanding of probability— people “have persistently irrational responses to quantitative information about risks and benefits,” regardless of their level of numeracy. For decades,

6 H A S T I N G S C E N T E R R E P O R T January-February 2012

decision scientists, economists, and psychologists have struggled to under- stand why even the most knowledgeable and numerate people make suboptimal health decisions.

It seems that in our attempts to educate patients about probability, we sometimes fail to appreciate that un- derstanding numeric facts and figures is not an exclusively cognitive effort; rather, it is often heavily influenced by affect, which in turn influences one’s ability to reason. A well-known experi- ment conducted in 1994 by Veronika Denes-Raj and Seymour Epstein illus- trates how affect can trump rational- ity, even for well-educated people. In it, subjects won if they drew a red jelly bean from one of two bowls. The small bowl contained one red and nine white beans, and the large bowl contained five to ten red beans and a total of one hun- dred beans in all. Despite the fact that each bowl was labeled with the percent of red beans it contained, the majority of subjects drew a bean from the large bowl, which was clearly the inferior choice. Subjects reported that they “felt” they had a better chance of winning by selecting the bowl with the greater abso- lute number of “winning” beans.

The interaction of affect and numer- acy in health decisions has been dem- onstrated in several studies examining choices of medical treatments that in- clude potential side effects. In one 2006 study conducted by Jennifer Amsterlaw and colleagues, people were presented with two surgical scenarios: one had a 20 percent mortality rate, and the other had a 16 percent mortality rate along with a 1 percent chance of four unpleas- ant side effects (colostomy, chronic di- arrhea, intermittent bowel obstruction, or wound infection). Most people chose the surgical option with the higher mor- tality, presumably because of their affec- tive response to the side effects and their tendency to overweight low probability events. Indeed, the mere presence of a small side effect may decrease willing- ness to undergo treatment, even if the

treatment offers substantial benefit. Erika Waters and colleagues found in 2007 that side-effect aversion occurred regardless of how probability was pre- sented or whether graphic formats were used to convey risk information, sug- gesting that decisions were guided by an affective response to the possibility of side effects, rather than by numeric risk.

There are, of course, potential dan- gers associated with providing patients with too much quantitative informa- tion. In this era of shared decision-mak- ing and unprecedented access to health information, it is easy to experience data overload. Providing people with all available information can actually hin- der decision-making, and often, less nu- meric information is more when helping people make quality health decisions. Among the potential dangers of provid- ing too much information is that pa-

tients may not be able to discern useful information from the merely relevant or altogether irrelevant. However, there are also potential dangers associated with providing too little information. As An- thony Bastardi and Eldar Shafir demon- strated in 1998, when people are faced with a preference-sensitive decision, they often seek additional information regardless of whether that information is critical to their decision. In a series of experiments, they observed that people who pursued missing information tend- ed to endow it with greater value than they would have if it had been available initially. Somehow the act of pursuing a missing but nonessential piece of in- formation lent greater psychological weight and salience to it. In a health context, such misguided information- seeking might lead patients to base im- portant health decisions on factors that may be relevant but nonessential to an

effective decision. Clearly, more work is needed to understand the possible unintended consequences of providing too much or too little risk information. At the very least, providers should keep in mind how affect can be attached to numbers and risks are perceived.

Wendy Nelson National Cancer Institute

To the Editor: There is a great deal of merit in Peter Schwartz’s important and useful article, and it will doubtless prompt considerable debate. I would like to add two brief comments.

First, Schwartz’s target—the quan- titative imperative—can be viewed as a specific instance of a broader target. In Rethinking Informed Consent (Cam- bridge University Press, 2007), Onora O’Neill and I, like Schwartz, were struck by the considerable evidence that

patients and research subjects often do not comprehend what is disclosed. As a result, “informed consent” is often considered obtained even when, rela- tive to contemporary standards, it is substandard or invalid. To improve this situation, we argued—amongst other things—that we need to be clear about the distinction between consent and informed consent. The former is a fa- miliar form of action that involves the setting aside of rights or the removal of certain kinds of prohibitions. Consent is of considerable, but not foundational, ethical importance for clinical actions. While those who consent need to know something of the action to which they consent, it does not require “disclosure” of large amounts of information about proposed actions or risks. Informed consent, in contrast, has its roots in negligence law in the clinical context and shares biomedical ethics’ particular

The quantitative imperative is simply part of an

unjustified informative imperative that places

unfeasible demands upon those consenting.

January-February 2012 H A S T I N G S C E N T E R R E P O R T 7

focus on the importance of individual decision-making. We argued that bio- medical ethics has, without sufficient justification, taken informed consent to be of key ethical importance. Matters are made worse by the prevalence of a wide range of metaphors that shape our thinking about knowledge and commu- nication, such that information is read- ily cast as a type of “stuff ” to be passed on, stored, disclosed, or picked up.

Putting these elements together and aligning with Schwartz’s terminology, there is a ubiquitous informative im- perative that pervades biomedical eth- ics. This informative imperative rests upon a range of distortions and confu- sions. Once the ethical arguments are clarified, the informative imperative can be seen as much less demanding in its scope than is typically assumed. We would thus agree with Schwartz’s con- clusion—that quantitative information of certain kinds need not be disclosed— but for a different set of reasons. The quantitative imperative, on our view, is simply a part of a widespread but unjus- tified informative imperative that is too broad in its scope and places unfeasible demands upon those consenting.

Second, it is important to note that there may be reasons other than en- suring validity of consent for disclos- ing quantitative information. Given the legal context of informed consent, Schwartz’s proposals might raise wor- ries for clinicians. Suppose information about certain risks is merely made avail- able, rather than being communicated (and acknowledged as such by, say, sign- ing a consent form). Suppose that one of the risks not communicated, but made available, happens. The patient sues on the basis that had she been informed of the risk, she would not have consented. The quantitative imperative can thus be seen as a precautionary strategy for cli- nicians and their institutions.

Schwartz focuses on the role that in- formation plays, or is purported to play, in enabling patient decision-making. In a recent paper in the Journal of Medical

Ethics (“Why Do Patients Want Infor- mation If Not to Take Part in Deci- sion Making?”), I focus on the (good) reasons that patients have for wanting information other than for decision- making purposes. Patients want infor- mation because they expect it, because it allows them to establish trustworthi- ness and credibility with their doctor, and because the very process of com- municating information can be a signal of respect and faith in a patient’s com- petence. So even if we agree that the quantitative imperative lacks firm ethi- cal support, we might still believe that such information should be disclosed.

Neil C. Manson Lancaster University

Peter Schwartz replies: I must correct Jessie Gruman’s sug-

gestion that my article supports with- holding decision aids or questions their importance. Like Michael Barry and Gruman, I agree that patients often want or need more information than they receive, and I believe that decision aids can help address this problem and will most likely play a growing role in medicine. But the question is whether decision aids can best help patients by always providing quantitative infor- mation—in particular, complex data framed in multiple ways, as the Interna- tional Patient Decision Aids Standards and many experts recommend.

Research has not established that such disclosure improves patients’ un- derstanding or decision-making in the range of situations where decision aids might be used. There are many possible negative impacts, mostly stemming from innumeracy and heuristics and bi- ases in human thought, as described in my article and in the excellent addition- al examples provided by Wendy Nelson in her letter. Given all this, I argue that the quantitative imperative must be subjected to more careful testing, in the spirit of evidence-based medicine.

Barry’s letter and Peter Ubel’s editori- al that accompanied the article support

research to investigate innovative ways to provide quantitative information to inform patients, including innumer- ate ones, without creating confusion or engendering irrational responses. Such research is certainly important, but my article emphasizes that the question is not just how to present certain types of data, but whether to present data at all, and in what situations. Researchers should not assume that all relevant data should be given all the time: it may turn out that in at least some cases, less is more (for example, see B.J. Zikmund- Fisher, A. Fagerlin, and P.A. Ubel, “A Demonstration of ‘Less Can Be More’ in Risk Graphics,” Medical Decision Making 30 [2010]: 661-71).

My article is thus a call to keep an open mind as research goes forward. It aims to play a role that theoretical ethi- cal and philosophical analysis should play: identifying and critiquing assump- tions that guide behavior or research in unrecognized or unexamined ways. The assumption that all quantitative information should be provided all the time is exactly that sort of philosophical commitment.

I agree with Gruman that patients too often cannot get information they want, but I believe that decisions aids may best address that problem by mak- ing the information available to those who want it, rather than presenting it to everybody. She raises the excel- lent question of how to choose who will receive additional data, and she is right, of course, to reject the idea that the decision should be based simply on a clinician’s impression of the patient’s competence.

But what about a system where pa- tients are offered additional information

Letters to the editor may be sent by e-mail to [email protected], or to Managing Editor, Hastings Center Report, 21 Malcolm Gordon Road, Garrison, NY 10524; (845) 424-4931 fax. Letters ap- pearing in the Report may be edited for length and stylistic consistency.

8 H A S T I N G S C E N T E R R E P O R T January-February 2012

of various sorts? This idea is far from radical: recommendations for discus- sions between health care providers and patients suggest that some information should be given initially, and additional information should be offered as an op- tion to those who want it (see, for exam- ple, R.M. Epstein, B.S. Alper, and T.E. Quill, “Communicating Evidence for Participatory Decision Making,” Jour- nal of the American Medical Association 291 [2004]: 2359-66). The amazing capabilities of computer-based decision aids may tempt designers to provide too much information up front, and to for- get the wisdom of tailoring disclosure to the patient’s interest and understanding.

I agree with Neil Manson that the quantitative imperative is part of a larger “informative imperative” in medicine that should be questioned and challenged as Manson, O’Neill, Carl Schneider, and others have done. Considering how to provide the right information, to the right patients, at the right time, by way of a decision aid or personal interaction, raises important ethical and empirical questions, as the article emphasizes.

Doctors and Torture

To the Editor: In “The Tortured Pa- tient: A Medical Dilemma” (May-June 2011), Chiara Lepora and Joseph Mil- lum raise the issue of whether a physi- cian may be justifiably complicit in torture and answer in the affirmative. Their argument is predicated on there being a litany of moral considerations, of which the wrongness of complicity in torture is merely one; this wrong- ness competes against other values and sometimes is outweighed. While I disagree with some of the authors’ as- sumptions—for instance, that torture is always unethical in the cases that physicians are forced to countenance, or that complicity in an immoral act is prima facie immoral—I agree with their conclusion. Surely those who trumpet deontological constraints would think

otherwise, but this conclusion naturally follows from a pluralistic set of moral values.

While they cite a wide range of dec- lamations against physicians’ involve- ment in torture, one that they leave out comes from section 2.067 of the Ameri- can Medical Association’s Code of Medi- cal Ethics. What makes section 2.067 interesting is not just what it says, but also the fact that it comes hierarchically nested under section 2.06, which speaks to physician involvement in capital pun- ishment. From the Code’s perspective, the issues pertaining to capital punish- ment and torture are isomorphic: what matters is merely that physician involve- ment could make the patient worse off. In the case of capital punishment, the upshot is obvious and, in the case of tor- ture, resuscitation in order to facilitate more torture is similarly depraved.

This argument fails in both cases, and the reason helps elucidate why Le- pora and Millum are on the right track. The question to ask is less what would happen if physicians were present, but rather what would happen if they were not. For example, imagine that physicians were disallowed from these settings and a prospective patient expe- rienced complications: the abolitionist would just settle for this person being worse off. A physician’s presence ensures that easily remediable situations be rem- edied, which is precisely what I would advocate. This is not to say that there are no capacities in which physicians could make people worse off, just that there are some in which those people could be made better off; therefore, a whole- sale abolition on physician participation misses the mark. (There’s also an open question about whether such agents should be conceived of as “physicians”

at all—as opposed to medically trained military personnel—but I shall not pur- sue that discussion here.)

To be sure, those opposing physician involvement in either capital punish- ment or torture are, almost always, not just opposing physician involvement, but rather those practices themselves. When the Code says that physicians must “oppose . . . torture for any rea- son,” it is clearly making a political claim and not one narrowly tied to medical ethics; it is for precisely this rea- son that I find such statements by the Code to be inappropriate. As Lepora and Millum acknowledge, some debate the appropriateness of torture in “narrowly specified, extreme cases.” It is a credit to their essay that such a debate is left open, rather than foreclosed by fiat.

From the perspective of medical eth- ics, the central question is whether phy-

sician involvement in torture makes the patient better or worse off. From a more thoroughgoing consequentialism of the sort that I would advocate, this ques- tion bears no privileged status. While Lepora and Millum would surely not agree with all of my arguments, they are to be commended for eschewing dogma and reaching a controversial conclusion. More generally, one would also hope that their paper portends increased at- tention to military medical ethics; this is an important area within medical ethics, and one that has received insuf- ficient attention.

Fritz Allhoff Western Michigan University

The question to ask is less what would happen if

physicians were present, but rather what would

happen if they were not.

Copyright of Hastings Center Report is the property of Wiley-Blackwell and its content may not be copied or

emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.

However, users may print, download, or email articles for individual use.