Harvard expert supports #WDDTY claims. Really

So last week I wrote a moderately widely read blog, mocking “what doctors don’t tell you” for claiming that UK children suffer harm from environmental toxins, and advising parents to avoid tap water and toothpaste. When these claims got reprinted in Families magazine, and ended up in my children’s school bags, I was mortified. Well, this paper https://docs.google.com/file/d/0Bw0QsePOF36RWGRJcmZST0tkX28/edit?usp=docslist_api came out on Saturday in the Lancet neurology. It’s by Grandjean and
Landigran, of the Harvard school of public health, and broadly supports the claims that I’ve been rubbishing when made by “what doctors don’t tell you”. The language is a bit more moderated, and there aren’t any specific lifestyle implications drawn, but the causal link between the presence of these toxins (including fluoride, manganese, and solvents) in the environment, and an epidemic of neurodevelopmental conditions, is repeated claimed. So am I, as our 1980s playground would have it, “sussed”?
Now, I’m not qualified to do a thorough debunking of this paper, but I wanted to make a couple of observations. Firstly, I don’t think we’re in a neurodevelopmental epidemic. The US has an epidemic of ADHD diagnosis, but that’s a different matter, which I’ve explored before.
Secondly, the paper reports papers chiefly from developing nations, with far less stringent restrictions on environmental exposure that exist in Europe and the US. Of course it still matters if Brazilian children are exposed to poisonous concentrations of manganese, it matters immensely, and the authors should be shouting from the rooftops about the plight of the global poor.
Instead they extrapolate their shocking findings to the US, where exposure is far lower, and the testing in the papers they cite is by that notorious source of junk science, hair toxin analysis. I can’t see any UK studies at all. They show no evidence of increased toxin exposure in clinically diagnosed populations.
So to be clear, this paper is claiming that an epidemic that may or may not be happening in one part of the world, is due to a set of toxins which are in vastly higher levels in a completely different part of the world, where this epidemic definitely isn’t happening. Either Brazilian manganese is causing ADHD in Milwaukee, or this is a load of speculative waffle.
So the paper is pretty transparent speculation dressed up as paradigm shifting insight, and WDDTY isn’t off the hook, but there’s a question and an implication here.
Firstly, why did the Lancet publish this? Would they have published the same words if submitted from Streatham technical college, instead of the mighty Harvard? If not, then they are guilty of precisely the “big beast” bias that their editor, Richard Horton, has been so critical of recently.
Whatever the reason, the implication of this for me is this: as with Aric Sigman’s recent pwning of the UK medical establishment over screen time shows, the rationalist cannot stand within the cool battlements of science, hurling studies and evidence dispassionately at her unenlightened opponents: the barbarians, if I can stretch this metaphor a tad further, are thorough the gates and sharing the high table. This makes debunking bad science riskier, tougher and less of a spectator sport. It also makes it more important.


Is ADHD a helpful diagnosis? How much should we be making it?

Note: This is mainly put together from blog posts from 2012, but since no-one read those, I thought it was worth putting the stuff out again.

There has been a peak of interest in over-diagnosis recently, for instance http://theconversation.edu.au/over-diagnosis-the-view-from-inside-primary-care-8889/  It’s all been interesting stuff, but what I think is not always made clear is where the specific harms of over-diagnosis are to be found. ADHD has been at the forefront of this debate, which dovetailes with a general critique of psychiatric diagnosis, so this seems a good place to start.

I should declare my interest- I do a lot of ADHD work clinically. So I’ve been thinking about this question:.In the UK, we diagnose about 1-2% of children with ADHD. If tomorrow, we increased that to the US figure of nearly 10%, what net harms would result?
Let’s start where we are now in the UK: what are we doing when we make an ADHD diagnosis?
We’re not saying that there is a distinct biological process going on here, like there is in Parkinson’s, for instance. Most people in the field think that ADHD people lie on a spectrum from the ‘normal’, with no ‘clear blue water’ between the two. That’s not to say that ADHD doesn’t have a biological basis, we’re finding stuff all the time, but usually by comparing a “definitely not ADHD” population with a “definitely ADHD” one. Things are much less clear in the grey, messy clinical world.
When the child’s problems are significant and pervasive, and due to their ADHD symptoms, we feel that ADHD is a helpful diagnosis to make. So when I say someone has ADHD, I’m saying that I think they are far enough along this spectrum for this to be the probable cause of a lot of the difficulties that they came to me with. Nothing more. Now, in order to be able to day this you need a pretty comprehensive assessment, of course, as there are many other potential factors at play.
Does making a diagnosis of ADHD benefit children?
I believe that for the children whose ADHD-ishness (technical term) is sufficiently pervasive and impairing to cause them problems socially and in their day to day function, then it is of benefit. There are effective interventions available, both pharmacological and non-pharmacological.
But am I not just exerting social control over undesirable behaviour? After all, one of the most ADHD-ish people I know of is quite happy, and brilliantly functional, but then he is a 1200 year-old Gallifreyan with a time machine. Dr Who doesn’t have to go to school, sit with his siblings at dinner without scuffling, or wait in a queue at Tesco. So yes, in a sense, we are treating behaviour and the distress resulting from the mismatch between this behaviour and social expectations. But that’s not an argument not to diagnose: social expectations can only be adjusted for ADHD if we know it’s there. Once we do know, we do our best.
Many of my patients do OK on non-pharmacological management, at least for a while, as schools adapt to them and parents gain confidence with support from my CAMHS colleagues, but medication is sometimes necessary to allow them to function.
What about stigma?

When I commented on Twitter that my patients seem to be less stigmatised after they receive an ADHD diagnosis than before, other people were skeptical. This interested me, so I thought I would record my thoughts on this.

Here’s my narrative:

When kids arrive at the point of being assessed, they have typically been labelled as ‘naughty’; they are frequently described in school reports as ‘choosing to behave’ in a particular way. Their behaviour is ascribed by schools to their parents’ lax discipline or other shortcomings, about which schools can do very little.

When I explain ADHD, I talk about the distribution across the population of ‘ADHD-ishness’, at the top end of which the child starts to run into problems, an

Once the diagnosis is made, I find that teachers see the chiIld in a more positive light, as finding it hard to conform rather than ‘choosing’ not to, and as someone who can be helped to engage better with education. Parents feel relieved of the burden of having caused the child’s difficulties, although they do now take on a significant therapeutic responsibility (which not all are delighted by….). Children also seem to benefit from the positivity.

It’s important to note that none of this requires medication- i prefer to discuss this after other interventions have been put in place.

There is another narrative, of course:

Child makes trouble at school, and is sent to a paediatrician for ‘sorting out’. Checklists seem to show ADHD, so the child is told they have an incurable brain disorder and needs to be on medication for life. Teachers give up on him as a lost cause. Parents understanding is that there is something ‘wrong’ with the child, nothing to do with them, and that they have no responsibility for helping the child, as the medication will take care of it.

I’m not sure that this second narrative is entirely wrong, and unfortunately it may be highly accurate in some cases. I try hard to avoid the pitfalls of the second narrative, but I’d be interested to know anyone else’s thoughts/ experiences on this.

What if you increased the diagnosis rate?
I’m quite strict about the diagnosis, aiming for a diagnosis rate of about 2% by insisting on pervasive and impairing symptoms which score highly on rating scales. But if my patients, the top 2%, benefit from treatment why not the top 10%?
Well, we would be starting thousands of children on medication which has a significant side-effect profile (although better than most psycho active agents). On the other hand, the evidence seems to be that quite a few of these children would find it easier to control their behaviour and concentrate. For other ‘over-diagnosed’ conditions, there is a strong argument that the drugs used are made less effective when you broaden the diagnosis; for ADHD I’m not convinced that we have that evidence- it may well be that someone only in the top 10% for ADHDishness would benefit from both environmental modification and drug treatment.
This idea scares people- it certainly scares me. But if we move away from ‘yuck’ reactions, what is wrong with the idea?
Firstly, some less persuasive arguments for harm:
There are arguments made about ADHD diagnosis stigmatising children. I’ve dealt with this above
Would we be giving children the message that they don’t need to deal with their difficulties, because a pill will take them all away? Would that not discourage self-determination and self-reliance in children? I don’t know. But recent research (http://jme.bmj.com/content/early/2012/08/27/medethics-2011-100224.full?g=widget_default) suggests that children on ADHD meds can recognise these threats and can, with help, counter them.
How about the argument that making a diagnosis of ADHD absolves families of responsibility for the child’s behaviour, so that they don’t need to discipline them as it’s not their fault? Some families do interpret the diagnosis in this way and it is very frustrating. Yes, the family will not have caused the problem, but they need to continue to impose boundaries and consequences on their children, with the help from professionals as necessary, and if they simply give up because he has ‘a condition’ they are letting the child down, just as a family whose child has diabetes would be if they refused to follow dietary advice. But the point here is that this response is a mistake and no reason to reject the idea of ADHD as a diagnosis.
It’s argued that we shouldn’t expose thousands of children to side-effects. But if the benefits outweigh the risks for one child, why not for ten? You may ask who makes this judgment of benefit vs risk; I would answer parents, doctors, and young people, together.
So I don’t buy the standard critique. However, there are some better arguments:
Firstly, you dilute the effect of the diagnosis: if one in 10 children have ADHD, 3 per class, it becomes difficult to distinguish those who need a little re-focusing from those who need near-constant attention.
We might, by ascribing behaviour to ADHD, miss an important factor such as domestic violence, language delay or anxiety. This can, in my experience, be one of the most serious consequences of being too quick to diagnose ADHD; stories from the US of patients diagnosed after a 20 minute consultation chill my blood.
It is often argued that we ought not to give families an ‘easy way out’ with drugs, when they can do so much themselves in terms of parenting, diet and lifestyle. For my 2% cohort, this argument is hogwash- they often need all the help they can get, and their lives are never easy.
But I can see the argument having more traction in a 10% ADHD world, as at the milder end they could mostly do OK without medication (although they may do even better with meds). This raises the interesting question of whether we should aim for ‘best’ function or ‘good enough’; one for another time.
Finally, increasing the diagnosis rate would swamp our service, and the mental health service, leading to long waits and sub-optimal care. The drugs bill would be collosal, and this money is probably better spent on making schools more ADHD friendly, and opening up access to parenting classes (in fairness, the latter is already happening).

So yes, there are potential harms in over-diagnosing ADHD, but they might not be the ones that those arguing against increasing diagnosis rates are thinking of.

Why does this matter? Well, I wanted to build an robust argument for not broadening diagnostic boundaries, as DSM-V seems to intend. And because I think that we will need it in the next few years to stop the UK drifting towards the US model for ADHD.

So how do we keep diagnosis rates at a reasonable level? I’m going to be a bit didactic for the sake of brevity. Apologies if it makes me seem a bit of an arse. Please feel free to disagree.

1. Start with an open mind
I slightly wince when services set up ADHD diagnostic clinics; doctors sometimes forget that we are as prone to cognitive bias as anyone else, and seeing a patient who may have ADHD in an ‘ADHD clinic’ seems a good way of ensuring over-diagnosis. I’m a bit of a purist about this- I won’t allow ADHD-specific questionnaires to be completed until we’ve had a chance to see the child- some schools and some parents are keen enough on the diagnosis already,  without allowing them to present it to us as a fait accompli!

2. Beware incentives to diagnose
The most obvious incentive occurs when the person who is paying for the assessment wants a diagnosis. I’m not making allegations, but I’ve never known a family go private and not get the diagnosis that they want, even if several NHS teams have denied them. It seems possible that this is an engine of the US explosion in diagnosis.
But even within the public sector, we need to be careful. If a child needs you to keep them on in your clinic, managers should not make this dependent on them having a diagnosis. Obviously if, say, the government farmed out ADHD services to private contractors who had a clear interest in diagnosing, medicating and following up large numbers of children, that would be a terrible in so many ways… but it’s never going to happen SO THAT’S OK, and anyone who accuses the government of the break-up and privatisation of the NHS is scare-mongering.

Outside  health, education authorities need to bear in mind the clear central guidance that provision should be based on need, not diagnosis. Finally, decent quality parenting intervention should be universally available, not just to diagnosed cases (as per NICE guidance).

3. Impairment not symptom counts!

I like criteria- they help keep everything consistent. But the DSM and ICD symptom checklists can contribute to over-diagnosis. Let’s look at the DSM on ‘Hyperactivity’ (you need 6 from this and impulsivity for diagnosis)

  • often fidgets with hands or feet or squirms in seat
  • often leaves seat in classroom or in other situations in which remaining seated is expected
  • often runs about or climbs excessively in situations in which it is inappropriate (in adolescents or adults, may be limited to subjective feelings of restlessness)
  • often has difficulty playing or engaging in leisure activities quietly
  • often talks excessively
  • is often ‘on the go’ or often acts as if ‘driven by a motor

Useful questions, every one. But what does ‘often’ mean here? These are behaviours that many children exhibit from time to time, and it seems to me that an exasperated parent may well answer these in the positive even if they are not the main problem causing the exasperation. So a consistently defiant  and slightly lively child can look on this checklist like someone with ADHD.

What I think is important is to try to quantify the degree of ‘ADHDishness’*,  this is not easy, but rating scales combined with careful qualitative interpretation of reported behaviours helps. You then look at impairment (can they go to the supermarket, go on a bus, eat a meal with their family etc) and if:

a) the ADHDishness is real, significant (I prefer to see 2 standard deviations above the mean) and inconsistent with developmental level.

b) the impairment is significant, long-standing and pervades several settings

c) the impairment can be attributed to the ADHDishness

Then I’m happy to diagnose.

4. Triangulate with school
Schools have the child for the most productive hours of the day, 5 days a week. If you haven’t sought their opinion on one of their pupils when diagnosing ADHD, you shouldn’t be doing it. Even if you do have an office on Harley Street.
5. Don’t abide by my rules: Think for yourself, draw your own conclusions. There are enough clashing cymbals in debates on ADHD (like me, for instance). See loads of children and get an idea of which ones benefit from the diagnosis.

*: I know this is a) not a single dimension and b) much more complex than I make out.