The power of the C word

The words we choose to describe reality matter, perhaps especially with a disease like HIV/AIDS, as Susan Sontag argued in her book on AIDS and Its Metaphors. Sontag relies on Aristotle to define metaphor, as ‘consisting in giving the thing a name that belongs to something else.’3 By adopting the oxymoron ‘Cure Research’, we seem to be adopting a metaphor for research that may be pleasing for both physician and patient, but which is at best confusing, if not outright misleading. In this regard, it is another example of the ‘therapeutic misconception’, the belief that a research (following a protocol to obtain generalisable knowledge) trial is actually therapy (treatment solely for the benefit of the patient).4 President Obama was indulging in what we might call the ‘political misconception’ that it is better to say something the public will like than to be realistic (curing cancer is not like going to the moon—biology is much more complex than engineering, and unlike the moon, there is no fixed destination, and it is not simply getting one or two people to a destination). My primary interest in this short piece is to suggest that the words we use in the informed consent process (and in our politics) matter—and that there are words, arguably including ‘cure’, that are illegitimate to use in recruiting research subjects, and obtaining their consent to be a research subject. This will no doubt be controversial, especially in an age where informed consent is itself under broad attack, the ‘right to try’ unapproved drugs is seen as a legitimate claim by terminally ill patients, and Congress is drafting legislation to further restrict the FDA's safety powers under the guise of promoting cures.5

Stuart Rennie and colleagues rightly note that ‘HIV cure concepts’ are works in progress.6 In the medical literature, for example, it is more common to find terms like ‘functional cure’ or ‘sterilising HIV cure’. The press and the public, on the other hand, are more likely to simply use cure, and are likely (it seems to me) to use it in the way Rennie and colleagues tell us it is found as the first definition in Merriam-Webster Medical Dictionary: ‘recovery from a disease (“his cure was complete”)’.6 In this regard, cure can also be used as a short hand for permanent disease remission: HIV may still be detectable, but it has no health impact on an individual and requires no further treatment. These are all good points, but I will concentrate on saying more about one of their questions, “how might cure language ‘do wrong?’” Their answer isThe most obvious way is falsehoods through unsubstantiated claims of HIV cure … In the short term, concern about ‘harms due to false belief via cure language’ is likely to center on informed consent to HIV cure research … the therapeutic misconception is the well-documented phenomenon of research participants mistakenly believing that they stand to benefit …[a similar] curative misconception in HIV research would involve HIV-positive research participants falsely believing that they stand to be cured by the tested intervention. (Emphasis in original)6

It has been suggested that oncologists are too pessimistic, in that they mostly never use the word cure. Former head of the National Cancer Institute, Vincent DeVita, understands their reluctance but disagrees.7 In his words, “In my opinion, when there is less than a 10 percent chance of cancer recurring after a patient passes his or her cancer's critical period, then the person should be told that, in all likelihood, he or she is cured.”7 Of course, DeVita is hedging his bets two ways: requiring both a probability assessment (10%) and a verbal hedge: ‘in all likelihood’. Cure really is a powerful concept.

The Mississippi Baby's role in cure research

Cure is even more powerful when combined with the names of particular individuals, such as the ‘Mississippi Baby’, and it seems reasonable to conclude that these combinations could make the ‘curative misconception’ stronger and more problematic. I'm no expert on the HIV trials that include the adjective ‘cure’, but I have had some experience with informed consent in this context. My brush with cure research involved the ‘Mississippi Baby’. In July of 2014, it was discovered that the Mississippi Baby (now a toddler) had not been ‘cured’ as an infant as had been formerly thought but was again infected with HIV. Near the end of the day this was announced, I got a call from Richard Harris of National Public Radio (NPR) who told me about the recurrence of her HIV, and asked me what effect it might have on the HIV cure studies with infants. The story ran on NPR the following morning.

My sound bite was placed just after that of Anthony Fauci, who said that NIH was taking a ‘close look’ at it to ‘make sure we do the study in an ethically sound way…’8 Here's what I said (on radio you can't really claim to have been misquoted): “To the extent that the justification for doing the trial is this one HIV-free child, and now it turns out that the child does have HIV, the trial should be stopped. We can't put them essentially in harm's way without some very good reason to think that they're going to benefit.” Harris concluded immediately after that in his own words: “The study designer will have to grapple with that issue as they continue to look for ways to treat children facing a hard future with HIV infections.”8 I got a strong negative reaction from the HIV researchers at my own institution and learned much more from them about the trial. On reflection, ‘suspended’ would have been a better choice of words than ‘stopped’, but the basic message is (I think) clear: The references to the Mississippi Baby had to be changed.

I later reviewed a copy of the protocol and the consent form for the infants' study, IMPACT P1115.9 I was happy to see that ‘Cure’ was not in the title of the study, ‘Very Early Intensive Treatment of HIV-Infected Infants to Achieve HIV Remission: A Phase I/II Proof of Concept Study’. But there were specific references to the Mississippi Baby (the only success to date and virtually the entire rationale for the study) that would have to be changed. For example, the protocol referred to both the Berlin patient (not relevant to the infant study) and the Mississippi Baby. The Mississippi Baby was not described as ‘cured’ but rather as being in ‘remission’ (defined as ‘no viremic rebound within one year’). Obviously, this was no longer true. Later in the protocol, it was asserted that the ‘widespread media coverage’ of the Mississippi Baby had led to ‘rapidly evolving’ clinical practice in newborns, which seems likely. In the consent form, under the caption ‘Why is this Study being Done?’, it stated: ‘For newborn babies who start antiretrovirals (ARVs) very soon after birth, it might be possible to later stop taking ARVs and stay healthy. This has only been seen in one baby so far’ (Emphasis added).9 The one baby, of course, is the Mississippi Baby. After her recurrence, the accurate statement in the consent form should be: ‘This has never happened.’ Or better I think, “We thought this happened once, in a case known as the “Mississippi Baby,” but we were wrong.” It might be argued that the Mississippi Baby should not be mentioned at all, and although I have some sympathy with that argument (and would generally argue not to use specific cases in the consent context), I think that the case of the Mississippi Baby is now too well known not to refer to directly in the consent process.

The answer to the scientific (and ethical) question of whether the study should have been ended depends upon how much of the rationale for the study was the single case of the Mississippi Baby. Two weeks after the NPR story, Fauci wrote that he was unsure ‘if indeed a true cure is possible’.10 The NIH news release on the recurrence contained the following statements by Fauci: “Certainly, this is a disappointing turn of events for this young child, the medical staff involved in the child's care, and the HIV/AIDS research community … we still have much to learn … The NIH remains committed to moving forward with research on a cure for HIV infection.”11 All of these are fair comments.

HIV research, the FDA and patient demand

Anecdotes persist in driving at least some other Cure research projects. On NIAID's website under the entry ‘Finding a Cure for HIV/AIDS’, there is a list of 17 funding opportunities and 8 active grants. ‘Cure’ is used in the titles of three of the funding opportunities and two of the active grants.12 Only one has a person's name attached to it, the Martin Delaney Collaboratory for HIV Cure Research. This is a name I recognised from earlier in my life when I was much more involved in HIV research—the late 1980s and early 1990s. My major work during that period was critiquing the FDA for permitting, even encouraging, wide use of unapproved and untested drugs for people with HIV/AIDS. In one particular instance, the FDA approved a poorly designed trial for a substance known as ‘compound Q’. The trial was developed by Martin Delaney, the head of Project Inform, who recruited physicians in San Francisco and New York to do a ‘community-based clinical trial’. The trial was not a success (several people died, although maybe not from the drug), but Delaney's lobbying did help move the FDA to expand access to investigational drugs under its 1990 ‘parallel track’ policy.13

In a long article in the New York Times about the trial, I was quoted as arguing that compound Q was being presented to potential subjects as a treatment rather than as an experimental intervention, an example of what I am referring to in this article as the therapeutic misconception: “We don't have a treatment. What we have is an experimental drug, and most experimental drugs will not work … [nonetheless] the message is that this one is probably going to work. An experimental trial is being looked on as beneficial for an individual, and that's not what they do.”14 Martin Delaney's biographer, Jonathan Kwitny, later wrote that Delaney was very upset at the article—mostly because he thought major parts of it were inaccurate and made compound Q seem more dangerous—and more experimental—than he thought it was, but also because the captions of our two photos were switched. As Kwitny describes it:Turning to the inside pages where the story continued, Delaney confronted two pictures. One was of him, and it was labeled ‘George Annas, professor of health law, Boston University School of Medicine, questions private trials.’ The other, which had the name Martin Delaney under it, depicted a scraggly, bushy-bearded man who looked like a turn-of-the-century Russian revolutionary, or maybe one of the Smith Brothers—at any rate, 180 degrees from the perpetually cherubic-looking Delaney. How much worse could this get?15

There were probably not two people in the world who knew both of us, and probably only a small minority of readers that had ever heard of either one of us. That means that for the vast majority of readers, the appearances were reality: Delaney looked like me and I looked like him. It overstates the case, but not by much, to speculate that research subjects are equally ignorant of science and medicine,4 and are therefore likely to believe as a fact whatever researchers tell them about the drug they are volunteering to test—especially about the drug's potential benefits.

In the 25 years since the Times article was published our expectations have, I think, shifted. We now expect physicians to tell us the truth about our diagnosis, prognosis and the risks and benefits of what they propose to do to us—we do not, however, expect the same from politicians. In fact, few, if any, Americans believe that the federal government can or will ‘cure cancer’ with its new moon shot project. Nor do we believe that the much heralded ‘21st Century Cures Act’ has anything to do with curing diseases but is rather an attempt to limit the authority of the FDA over drug safety.16 The claim is that reducing the FDA's authority will make drug approvals faster, and speed will enable new cures; but that claim misses the fact that the FDA does not make or even test any drugs. FDA's job is to make sure that the drugs others make and sell are ‘safe and effective’. Making drugs that have not crossed this threshold available to sick people, whether through ‘compassionate use’ or some other more expansive programme, combines the worst aspects of the political and therapeutic misconceptions, and is unlikely to cure any diseases, or even any person with disease.17 This was true 25 years ago and remains true today.

Faith in medical research

This is not to say that there is evidence that curing HIV/AIDS is impossible, and it is likely that at least some people with HIV/AIDS, and some researchers in the field, believe that a cure is possible. This is both because individuals with HIV want to believe they will someday be able to stop ART, and because at least some Americans, perhaps including our president and vice president, continue to believe that science can conquer (ie, cure) virtually any disease, including HIV and cancer. As Don DeLillo put it in his White Noise more than 30 years ago:You can put your faith in technology. It got you here, it can get you out. This is the whole point of technology. It creates an appetite for immortality on the one hand. It threatens universal extinction on the other … new technologies every day. Lasers, masers, ultrasound. Give yourself up to it … They'll insert you in a gleaming tube, irradiate your body with the basic stuff of the universe. Light, energy, dreams. God's own goodness.18

The way we put our ‘faith’ in medical research is to privilege belief in what we want to happen over evidence-based expectations of what is likely.17 This is a form of the ‘therapeutic misconception’, which seems to be inherent in agreeing to take part in ‘Cure’ research: we seem to have nothing to lose and everything to gain. The argument is not that the subjects who might volunteer for ‘cure research’ are out of touch with reality, only that many more people are likely to volunteer for a study they believe is designed to treat their disease than for the same study in which their ‘treatment’ will be determined by a flip of a coin. In the SUPPORT study of high or low oxygen concentrations delivered to extremely premature infants, for example, the therapeutic misconception of the researchers was that randomising to determine how an infant would be treated did not have to be disclosed to the infant's parent if both arms of the protocol were consistent with the ‘standard of care’ somewhere in the USA, because the entire experiment could be considered therapy.19 This argument, I think, is untenable since what the patient loses in randomisation is the opportunity to decide for themselves, with the assistance of a physician whose only job is to foster the best interests of the infant, what treatment to employ.4 Only if the two arms of the study had the same likely outcomes (they did not, one was an increased risk of death and the other an increased risk of blindness) would the choice between them be unimportant.19

The problem with joining ‘cure’ with ‘research’ to produce ‘cure research’ is that it creates an oxymoron, and it is an oxymoron that has a pedigree in HIV/AIDS. Perhaps the most effective disease lobbying group ever, ACT-UP (AIDS Coalition to Unleash Power), for example, had as its slogan: ‘A drug trial is health care too’. This is simply not true. A drug trial is research designed to test a hypothesis, not to help individual patients.17 Telling research subjects that they are patients, and even worse, that they have ‘nothing to lose’ by trying the ‘new’ drug, as was common in the early days of AIDS when there was no effective treatment, hopelessly confuses research with treatment and makes ‘informed’ consent virtually impossible.

So my conclusion is that ‘cure’ is simply too strong a word—promising so much more than it is likely to deliver, that it is misleading per se and so should not be used either in the name of the study, or in the informed consent form or discussion. Even if a researcher did not accept this argument, a powerful pragmatic argument cuts the same way. The Mississippi Baby can be seen as an example where researchers bet that the Baby was ‘cured’ and doubled down on the bet by opening up a new, research programme: HIV Cure Research in Infants. Because the story of the Mississippi Baby was at the very core of this infant research endeavour, finding HIV in the Mississippi Baby threatened to halt the infant HIV set of research projects. This is not unusual—there are a number of very famous cases in which ‘first-of-its-kind’ research was stunted, or even halted, for decades or more because premature research was widely reported in the popular press that ultimately turned out to be useless or even disastrous to the subjects.

Subjects who halted research projects

The most famous example is Barney Clark, the world's first recipient of a totally implanted artificial heart. Although a few additional Jarvik-7 hearts were implanted in subjects after Clark, it was Clark's horrific and widely publicised personal experience as the first recipient that doomed the Jarvik 7 as a viable medical device.20 It was decades before another company was willing to try another artificial heart in a clinical trial.21 More directly pertinent to the Mississippi Baby is the case of Baby Fae—the first (and last to date) infant to be the recipient of a cardiac xenograft, a baboon heart. Baby Fae was front page news around the world.22 But she and the experiment were both short-lived, and when her heart hyper-rejected it spelled the end of xenografts for decades.23 Similarly, the death of Jessie Gelsinger set back gene transfer research at least two decades.24 ,25 These cases, anecdotes for sure, are powerful indications that the use of individual cases to obtain informed consent is not only misleading to potential research subjects, it also puts the entire research project at risk of early failure based on the experience of one research subject—which may not be representative. And, of course, the fact that there is no cure for a fatal disease does not make experimental drugs ‘therapeutic’, any more than a mechanical or baboon heart can be classified as therapeutic.17 Less well-known examples of highly-publicized early failures setting back research include mitochondria transfer experiments in China,26 and uterus transplants in the USA.27

Where does this leave the ‘Berlin Patient’, Timothy Brown? After the recurrence of HIV in the Mississippi Baby, Brown is now described as ‘the first and only person known to have been cured of HIV’.28 amfAR, a private organisation that promotes HIV Cure Research, has said that almost half of its research budget will be devoted to ‘cure-focused research projects’, which it also describes as ‘the holy grail of AIDS research’.28 It has recently branded its effort as the ‘Countdown to a Cure for AIDS Initiative’ with a goal of ‘developing the scientific basis for a cure by 2020’, and has also launched the new amfAR Institute for HIV Cure Research at UC San Francisco. Their ‘2015 Year in Review’ message to potential contributors reads in relevant part:Teams of amfAR grantees are racing to locate, understand, measure, and destroy the HIV reservoirs that have stymied efforts to eliminate the virus. To cite just one promising investigation, researchers at Rockefeller University are conducting a multi-country clinical trial of a very promising ‘shock and kill’ strategy to flush the virus out of these reservoirs and then boost the body's immune system to destroy it. Progress like this is why we can't stop now—there's simply too much hope on the horizon, and too many millions of lives at stake … thank you for standing with us to conquer AIDS.29

Of course, this is a fund-raising letter, and some exaggeration is expected. Nonetheless, the military metaphors employed suggest the Iraq war strategy of ‘shock and awe’, a strategy that seemed to work in Iraq in the short term, but was a disaster in the long term. It also suggests that our goal should be to ‘conquer AIDS’ and that only by doing so can we save ‘millions of lives’.30 The message also equates ‘cure’ with ‘conquer’, the first being a medical goal; the second a purely military one. The December 2015 letter does not mention the Berlin Patient, but their October 2015 newsletter makes it clear that his case remains the primary reason to think a cure may be possible, even though his treatment, ‘with a high-risk stem-cell transplant is not a practical means of delivering a cure …’.28 The ‘shock and kill’ strategy is also very high risk—especially for patients whose HIV is being controlled by standard treatment—which is likely all of them who are ‘eligible’ for the HIV Cure trials.