Questioning Medical Authority by Accident

Last week I gave a talk about recovery at an eating disorders clinic. It was the third one I’ve done, and it’s always a strange experience. I never had inpatient treatment myself, though my mother and I once visited a friend on an eating disorders ward when I was still ill myself. Being in such places feels like being given a glimpse of a path that might have been mine. Going to hospital was a possibility my parents occasionally held over me as an ultimatum: if I didn’t find a way to make the danger I was in less critical, by independently eating more, that would be my only option.

And now I find myself here, but brought in as the shining example of the person who got to where everyone there is supposed to want to get to. I’m introduced just as Emily (not as the potentially misleading Dr…), I say a bit about my work, and then I launch in, with a brief semi-disclaimer about how I never had inpatient treatment myself, but how I hope what I have to say will still have some relevance.

Looking back, probably I should also begin with a more solid clarification of the kinds of evidence on which I’m basing what I say. Maybe I should have even have offered a proper disclaimer, as I do when I respond to private messages from you, my blog’s readers: I am not a healthcare professional, and what I say is not intended as a ‘substitute for professional advice, diagnosis, or treatment’, nor should you ‘disregard professional medical advice or delay in seeking it’ because of something I say. But then, that would have disrupted the intimacy of the conversation from the outset.

It’s difficult to know how best to pitch these talks. There’s not much point in going on about how brilliant life is after recovery, because I remember how hollow that always used to sound to me. And as I’ve concluded for this blog more generally, it also doesn’t seem very useful to simply ‘tell my story’ with no attempt to connect it with what is known about the human mind and body more generally. But drawing lessons from a single experience is tricky, for reasons I’ll come back to.

What I settled on was the attempt to convey two general points, using my own experience as illustration and evidence.

1) Full recovery is possible.
2) There’s no point in aiming for anything less.

I talked about the simplicity, if not the ease, of the essence of the early and middle stages of recovery: simply eating, according to plan, and keeping on eating until the keeping on eating is no longer a problem. I said that full recovery was incompatible with what is so commonly done: deciding on an ‘acceptable’ BMI like 19 or 20, eating just enough to get there, and starting to restrict one’s intake to stay there. I gave two pieces of evidence that this is counterproductive: the different rates of restoration for fat and fat-free mass, which are likely to result in a temporary bodyweight overshoot, and the gradual normalisation of metabolic rate which is completed only in the final stages of weight restoration.

At this point one young woman put up her hand and asked: her treatment plan involved reducing her calorie intake at a particular BMI level; did I mean she shouldn’t do this?

Source: DWRose, flickr (CC 2.0)

What I quite possibly should have said is: “Well, that’s a question for your medical team, and I can’t really comment on the specifics of your individual case. You could raise this question with them, but given I’m not a medical professional, I shouldn’t advise.”

I didn’t say that.

What I said was (more or less): “No, you probably shouldn’t. I don’t want to interfere with your treatment, but there are strong reasons against imposing dietary restriction on yourself at a predetermined point. I’m speaking from my own experience, and from all the research I’ve done on the science of physiology and diet, as well as from a lot of anecdotal evidence from readers of my blog, and from all these perspectives deciding in advance on a point where you're going to reduce your intake seems likely to be counterproductive. And who knows, by the time you get to the point where your current plan is to reduce your intake, maybe you won’t even want to any more.”

Afterwards I asked the staff member who’d introduced me whether what I’d said had been problematic. She said it was OK, and that if any issue came up, they would simply say that everyone is entitled to their opinion, but I’m not a professional, and her treatment is being directed by those who are.

I found my mind kept turning back to that initial exchange as the day wore on and I made my way home. On the one hand I felt I’d been a bit irresponsible in expressing an opinion on an inpatient’s treatment plan. On the other hand, I still felt that doing anything else would have been a cop-out, and would have betrayed the whole spirit of the talk. As a mini dilemma, it raises some interesting questions about eating and, more broadly, about responsibility and authority.

If I’d had time to talk with her properly, I’d have wanted to talk about how she was finding the amount she was eating now, about her hunger and the things that pull her in the other direction from hunger, about the differences between intensively supervised treatment and doing the re-feeding process independently. I’d have reiterated that I never had inpatient or even day-patient treatment, so I couldn’t completely imagine how it must be to have responsibility for one’s diet wholly or partially in professionals’ hands. I imagine there is an important difference between being given a meal plan that perhaps increases periodically, and at some point may also change in the opposite direction, and being in full control of one’s diet and making the (guided or personal) decision to reduce it.

Still, though, there’s the inalienable essence of that reduction: once you've done it, you'll feel less full than you used to on the meals you were accustomed to eating, feel hungrier sooner after them, realise you’re now expected to feel full on less.

The obvious objection is: well, this only seems a problem because you’re assuming the reduction will happen before the point where it should. You’re assuming that the person being treated will still be hungry for as much as she is being given, and that less will therefore be too little. But if the time has been judged right, that won’t be the case: the current amount will be too much, and it will feel right and good to reduce it slightly to ‘maintenance’ levels.

Eating Disorders Essential Reads

The Quiet Radicalism of Looking After Your Body

Hope and Hesitation in the Journey of Anorexia Recovery

That’s true. But the crucial point is that the moment where just right becomes too much cannot be predicted with complete confidence. Probably not even a ballpark figure can. When I got slightly anxious about my weight creeping up towards and then beyond the ‘healthy’ BMI range, my therapist, wonderful as she was in all other respects, didn’t say ‘stick with it, you’re still hungry all the time, carry on doing what you’re doing until you’re not’. She advised me to replace whole milk with semi-skimmed, and to stop eating so many puddings. I did as she suggested, and it felt wrong, and I went back to what I’d been doing before, until eventually I realised the hunger had gone.

Not that my experience should be taken as typical. Just that there is a very great danger in telling someone in recovery from anorexia to eat less. And when the moment of telling them to do this has been decided long in advance, the danger is far greater still. What you want to do is get to the point where you no longer need to be on a treatment plan with a daily energy allowance, where inpatient treatment has given way to daypatient and outpatient because your eating has become something unproblematic and natural, and where less may be eaten, but not a counted amount less, and where your bodyweight is what it is when you are living well, not protected like the sacred number it long was. This is an ideal, of course; but should we ever aim for anything less?

There was no chance to talk any of these things through: by the time the session was over, it was their lunchtime. When I first replied to the question, though, what was my primary responsibility to the group? What principles should I have let guide my reply?

The most important thing of all in such a situation is presumably not to compromise the recovery of the person in treatment who asked the question, or of the other people in recovery who were listening to her question and my answer. The second most important thing must be to aim to further her and their recovery. But as soon as we ask how best to do either of these things, we run into difficult questions. Should we worry more about compromising some aspect of recovery right now, today, this minute, or should we take the longer view on a process that lasts months or years? And does a greater danger lie in complicating a person’s relationship with the clinicians supervising her treatment, or in missing a chance to impress on her how vastly much more likely it is that she will remain underweight than become overweight, and how crucial ditching the idea of predetermined dietary restriction is in getting away from underweight?

The answers to these questions are as much about individual and group psychology – about motivation and stage of recovery, about relationships with those guiding it, about responsiveness to authority and to conflicting perspectives – as they are questions with generalisable answers. But if we are seeking some general principles, what do we find?

There are two obvious questions to ask. One, how likely is it, on average, that someone recovering from a restrictive eating disorder goes on to become and remain overweight? Two, how do the physiological dangers of being underweight and overweight compare?

On the first question, I haven’t managed to find any targeted research. Science of EDs offers an interesting overview of issues relating to eating disorders that arise out of a prehistory of being overweight, including the difficulties clinicians may have in dealing with eating disorders where objective thinness is not present. But in all the clinical trials I’ve read, I don’t remember ever coming across a mention of participants overshooting their weight goals. The question is always whether they will achieve their targets at all, and more often than not those targets are on the distinctly low side (more on this later). Whether the absence of mentions of overshoot means it never happens, or it’s not reported because it’s viewed as a failure, or it’s not reported because it’s viewed as no more important a success (or failure) than a participant simply reaching their target, I don’t know.

On the second question, there have been a few studies highlighting the fact that overweight (as distinct from obesity) is associated with lower ‘excess mortality’ rates than underweight, and in some cases even lower than ‘normal/healthy/optimal’ weight (Flegal et al., 2005; Visscher et al., 2000; see also Keith et al., 2013; Roh et al., 2014; and Cao et al., 2014; and here’s a quick overview in The Independent by Kendrick, 2015).

Whether or not ‘likelihood of death’ is the most meaningful measure of health can of course be debated. But it’s pretty clear that there are at least equivalent, quite possibly greater, risks to underweight as overweight. And let’s make our working hypothesis that remaining underweight is far more likely than becoming lastingly overweight for those who have experienced restrictive eating disorders. Why, then, would inpatient treatment for anorexia involve a pre-planned reduction in daily energy intake? Why is the patient not expected to be free to make her own decisions about this because her re-feeding program will bring her to the point where she is physically and psychologically well enough to do so for herself?

The most charitable explanation is that this is a strategy designed to help the patient feel less terrified about recovery, and so commit to it with less ambivalence. In a kind of protective second-guessing, the doctor predicts that if the patient knows there is no predeterminable end to the weight-gain process, and/or if weight gain continues beyond a point that is by some obvious metric ‘necessary’, the patient may fail to agree to start recovery in the first place, or ditch it somewhere along the road out of fear of the unknown. This strategy thus serves as a calculated compromise: better to get the patient to a moderately safe BMI than to fail in the advocacy of something greater.

If this is a reality, it's a sobering one. Unless we have established with confidence that aiming for full recovery makes it significantly less likely that someone will achieve even partial recovery, this is a pretty important gamble. And let's be clear: this has absolutely not been established.

So what else might be going on here? The pull of medical targets doesn’t explain it: getting a patient to more than her target weight can’t be a bad thing. Or maybe it can: maybe someone ending her treatment at a BMI of 26 would constitute as much of a failure in a clinical write-up as the patient relapsing or never progressing beyond 18. Maybe clinicians are even worried about the legal repercussions of re-feeding someone to ‘past where they ought to be’.

Targets certainly come into the picture when we look at treatment programs developed as part of clinical trials. In my first post on metabolic rate, I mentioned the strikingly low BMI values often used as the threshold of the ‘recovered’ category in published eating-disorder research, as well as the questionable practice of significantly reducing energy intake after the re-feeding phase (in Krahn et al., 1993). To get a rough idea of how BMI figures in the current clinical practice of anorexia treatment while minimising search bias, I’ve just done a brief analysis of the top 20 Pubmed hits for ‘anorexia nervosa treatment’. Seven of them describe clinical trials or other structured interventions, and one of these (McIntosh et al., 2005), mentioned no fixed recovery/remission criteria, with end-of-trial BMIs between 18.1 and 18.8. Of the remaining six, all define recovery, weight restoration, or full remission with reference to a BMI of 20 or less.

Schebendach and colleagues (2017) use a BMI of 20 plus a cutoff point on the Food Preference Questionnaire; for Berends and colleagues (2016), 20 marked the end of treatment and the start of the ‘aftercare programme’, while relapse was defined as falling below 18.5. Tubić and colleagues’ (2016) study ended at a BMI of 19; Steward and colleagues (2016) used the DSM-V criteria plus an Eating Disorder Inventory score to define full remission, and the group average was a BMI of 19.1. For Moody and colleagues (2016) weight-restored was defined as being above 18.5, and for Egger and colleagues (2016) a recovered BMI could be as low as 17.5 or above (even though AN could here be diagnosed between 15 and 18.5), plus an appropriate Psychiatric Status Rating, and the final observed BMIs were 18.2 and 17.9 for the two forms of treatment under investigation. It saddens and angers me to observe the uniform inadequacy of these BMI markers of recovery.

Statistically, few healthy adults have a BMI that falls naturally between 17.5 and 20, and after serious malnutrition there are important physiological reasons to expect and encourage a temporary overshoot in bodyweight restoration (see above). The BMI targets in this sample of studies – which to me appear decently representative of the wider field of research – thus seem entirely inappropriate. (The broader question is whether BMI deserves the central status it always has in this field: the general consensus tends to be that it’s flawed but the alternatives are more flawed. The US Centers for Disease Control and Prevention conclude that it ‘should not be used as a diagnostic tool’, which it generally is in the eating-disorders realm.) This small snapshot of clinical research practice in anorexia treatment suggests a worrying state of affairs, where the threshold for physical recovery is placed implausibly low, and no mention is ever made of encouraging patients to continue weight restoration beyond that minimal level. Where do these clinical habits come from?

Research studies in all fields are guided by the biases of needing to get published, and by the fact that it’s far easier to get a positive result published than a negative one (Matosin et al., 2014). In the field of eating-disorder research, though, there may be an additional pressure at play. Researchers and clinicians working with eating disorders live in the same world as the rest of us. Just as we all are, they are bombarded with a perpetual pressure to value the standard variants on slimness: 1) the fashion-model version – near-emaciation; 2) the glamour-model version – the extremes of slim thighs and waist and large breasts and bum; and 3) the fitness-model version – significant muscle mass and very low body fat, often also with the basic proportions of (2). What distinguishes these researchers and clinicians from the rest of us is their proximity to a clinical community, within the field of disordered eating and perhaps further afield. Maybe they are better and more continually informed about the medical consequences of obesity, which is more prevalent a problem than the opposite at a national population level. Maybe they therefore feel inclined – reflectively or not – to defend against obesity as actively as they support recovery from anorexia. Or maybe it’s all just habit and inherited practice, and the following of guidelines because their evidence base is assumed rather than interrogated.

These possible answers areke all speculation, but the questions are important to ask. By now, though, we have shifted from the realm of responsibility to that of authority. Coming back to the woman who asked me the question: what kind of authority does it make most sense for her to be guided by? There are non-arbitrary reasons to place faith in mainstream medicine, and specifically, in the logic of the randomised controlled trial (Haynes et al., 2012). There are good reasons to resist the intuition that translates automatically from experience and anecdote to data, or from correlation to causation: it seemed like this for her, or for me, therefore it must be so.

But no system of knowledge is infallible or total. Smith and Pell (2003) give an excellent, and only semi tongue-in-cheek, demonstration of where the logic of RCTs gives out. And when it comes to diet, RCTs are difficult to carry out on the required scale. This makes dietary science, and its intersections with the study of eating disorders and obesity, an excellent example of the fallibility of the scientific method as currently practised, because it’s so heavily reliant on observational ‘pseudoscience’ (Taubes, 2012). Even when controlled experiments are possible, and are carried out, their logic is a systematic elimination of variables conceived of as separable and reducible. This may not always reflect the structure of reality. It is misguided to reject out of hand the value of controlled experiments, because they can separate out correlation and causation in an epistemically powerful way. But for reasons like these it is also dangerous to reject the value of other forms of learning, including through singular experience. When my experience tells me that at a BMI of 19 or 20 my recovery had barely even begun, that is a meaningful fact to place in dialogue with the Pubmed sample. The vast discrepancy is not the end of the debate, but the start of one.

If we apply these principles to the context of eating-disorder research and treatment, we may conclude that it is just as wrong to reject the general usefulness of objective measures like BMI as it is to reject the relevance of subjective realities that can’t be easily quantified or experimentally isolated or aren’t just about identifying cause and effect. Standardised BMIs or questionnaire scores tell us a lot about recovery; there are levels for both below which recovery from anorexia is not possible, I would say, but recovery also involves lots of things those measures will never capture. That is, they are necessary but not sufficient markers of recovery. The reality of everyone’s recovery has differences on a bedrock of human similarity, and it’s equally easy for clinicians to underestimate the individual differences and for sufferers to overestimate them. In sickness and in health, it can be just as tempting to assume one moment that my own what-it’s-like-to-be-me is the same as everyone else’s, and the next moment that mine is utterly unique.

And here we are at the most profound questions of them all: at the mystery of how and why we have conscious experiences that feel privately our own, and at the conundrums of what realities are relevant when and to whom, and how on earth to arbitrate with confidence the truths we want to be guided by.

Conducting deeply interdisciplinary research brings you up against questions like this all the time. My current research practice is poised somewhere between literary studies, experimental psychology, and clinical psychiatry, all of which make contrasting assumptions about the nature of evidence and argument and about the kind of knowledge that is of value. I used to reject most of what my home discipline, literary studies, said about everything; now I’ve found my way to a point where I feel I see the psychologists’ blindspots and weak points as clearly as the literary scholars’. Not at all to claim that this view from all angles lets me overcome all the failures of any angle alone, but it has at least become clear to me that the easiest mistake one can make in any inquiry about any psychological phenomenon is to declare that one system of knowledge is always right and superior, or that the answer is only ever one thing and never the other.

So, in the case of me going into an eating-disorder clinic and saying that I don’t think a pre-planned structured reduction in energy intake is an appropriate part of a recovery plan, where does all this leave us? We’ve come some way from the easy ‘I’m not a medical doctor therefore I shouldn’t have commented’ anxiety I felt immediately after the talk. It now seems clear to me that the issue clearly can’t be easily reduced to a straightforward humanities/sciences split, or a grand divide between personal experience and generalised medical practice, though it is these things too. Beyond these apparent polarities, it’s about different understandings of recovery, different risk-benefit calculations, different interpretations of evidence, and different relationships to the person who asked the question. In this sense, perhaps our little flash point of a question-and-answer was a valuable window on to reality.