Déjà-vu at IMFAR 2017

I’m back in the office after my usual annual pilgimage to the International Meeting for Autism Research (aka IMFAR) which took place this year in San Francisco. As usual, it was inspiring to be among so many researchers and, increasingly, autistic people (including autistic researchers, of course) gathered together to share their work. The quality of debate in person and on twitter (search #IMFAR2017 and #AutIMFAR for examples) was superb.

In fact, the quality of the conversation quite often far exceeded the quality of the presentations.  Which is to say, speaker delivery was excellent, and the science included plenty of robust methodologies and impressive statistics, but projects that truly inspired were few and far between. I wasn’t the only one to notice – over and over again I heard people agree that, taking the IMFAR programme as the raw data, you’d be fogiven for thinking that no new ideas had infiltrated autism research over the last 5-10 years. In contrast, over countless conversations in coffee breaks and over meals I was blown away by the commitment, passion and briliance of my colleagues in the research community. If people are having great ideas, why isn’t that reflected in what we see being presented at IMFAR, this apogee in the autism research calendar.

But maybe this is just an unfair impression?

To check, I’ve been leafing through my programme books from the last five years of IMFAR meetings, covering San Sebastián, 2013 to this year. Here are some examples of continuity from year to year – these are all oral session titles, verbatim, though I have sometimes truncated them where there was a sub-title, to save space.

As you can see, there’s a very striking level of duplication over a five year period in topics being presented at IMFAR.  There are many other examples too, such as repeated sessions on brain connectivity, genetic markers, translation from animal models and drug development. I’ve not included these because I don’t know the methods so well and can’t be sure whether what looks superficially like the same topic, really is the same. But I have my suspicions…

Now, to clarify, I’m not drawing attention to this repetition as a critique of the research quality, nor of the relevance of these topics.  The themes identified above clearly have importance – not just for scientific discovery but for people in the autistic and autism communities. Of course there is great value in learning about lifespan trajectories of autism, or refining ways of measuring autism-relevant outcomes. And, as the bottom row shows, new themes are capable of rising up and gaining status. The question is, though, can research make rapid enough progress to justify re-examination of these topics on an annual basis? And the answer is indubtiably, no. Academic research moves slowly. For example, running a randomised controlled trial with a decent sample will take at least 3 years to set up, recruit, administer and analyse.  Longer if you include the period applying for funding before you start, and longer-still to get the results published in a journal.

Of course, different projects in different labs come to an end at different times – in any one year there will be new data in each of these topic areas and more. Individual groups of scientists will continue to submit their data as soon as they are available. But if everyone is submitting their data, on these topics and others, that can’t explain why we should get such consistency in what actually gets presented. Why do we end up with such a high level of repetition from year to year?  I have two hypotheses:

1. Winning on traditional markers of scientific significance

Each IMFAR submitted panel or abstract is reviewed by independent experts in the field.  They will be looking for traditional markers of excellence – things like large sample sizes, longitudinal data, and randomised controlled trials.  All of these things are evident in the themes identified above.  Some research topics, almost by definition, will not have these sort of features. Qualitative studies will normally not make the grade. Brand new topics without a lot of established resource behind them will suffer against these metrics. Small studies with hard-to-find populations, such as pregnant autistic women, will likewise struggle to impress. Instead, the big methods, and the big numbers will normally win the coveted highest ratings. And this means a restricted range of topics make it onto the oral presentation line-up.

Worse still, this pehnomenon could be self-sustaining.  To the extent that an IMFAR presentation helps a researcher win further funding for their field, or achieve tenure, the topics being presented at IMFAR will determine the work that gets done next, and that is available for presentation at future IMFARs.

2. Wherewithal to co-ordinate a submission

IMFAR release their call for submissions months before the conference itself – I uploaded my abstracts in October 2016 for a meeting which took place last week. Priority is given to abstracts with data analysed at the point of submission, which means small-scale studies and brand new topics might not have something to submit in time. IMFAR also requires that data not be published elsewhere, so someone doing a small study of (say) experiences of autistic undergraudate students, might not have data for one IMFAR but might have that study submitted to a journal the following year. As a result, it is big programmatic research streams which are going to be most well placed to handle the IMFAR submisson timeline and rules.

The situation is even more extremem when it comes to Panel submissions. Individual researchers / groups of authors can submit individual papers, but about one third of oral presentation sessions go to ‘Panels’.  These are co-ordinated submissions from a group of academics on a similar theme. Panels ought to include different research institutions, ideally from different countries too.  They are encouraged to incorporate early-career researchers as well.  These are both great criteria, but in effect, I suspect this means that Panels tend to be led by very established research groups who already have active international collaborations and strong leadership in the field. An early-career researcher with the capacity to develop a Panel, or even get invited to join one, if she isn’t already part of an established and high-profile lab, is a rare individual indeed. Cutting edge topics are unlikely to have the kind of wide-spread interest, let alone cross-institution links, to be good Panel applicants. Again, the end result is that the big players are going to have a significant advantage in pulling together a Panel submission.

This isn’t to say that nothing really exciting ever makes it onto the IMFAR programme.  This year we had a fascinating session on ASD and Sexuality, last year there was Perspectives on Pain in ASD and in 2015 Roy Richard Grinker led a keynote-panel on Autism & Society: taking stock of the history and meaning of autism research. These are all truly innovative topics showcasing quality research evidence, using methods suited to their topics, and with relevance for autistic people and their allies. But they are not the IMFAR norm.

So, how do we fix it, Sue?

What is needed is more high-level editorial control over the IMFAR programme to ensure that new themes are given space to be explored. This is hard at the level of the individual reviewer, but topic chairs could be encouraged more explicitly to make decisions with reference to previous years’ programmes, prioritising work which reflects more than just incremental progress. I’d like to see a separate submission category for brand new topics, so researchers can submit abstracts, with a section where they specifically flag how this is new in relation to (say) the preceding three years of IMFAR content. I’d also like to see some sessions more explicitly providing an overview of the field – experts presenting not just their latest data but providing a review of the state-of-the-art of a specific topic in autism research.  These could be selected with attention paid to achieving some discontinuity from year to year.

IMFAR remains one of the highlights of my year, but it’s time to see the kind of excitement and diversity in the programme that I have the fortune to experience in conversations during those few days.