Richard Stevens: “Why would you base medical decisions on anything but evidence?”

September 5, 2016

Prof Richard Stevens

Professor Richard Stevens is Course Director for the MSc in EBHC Medical Statistics

When I was at primary school I wanted to be a scientist, or a teacher.  As it turns out, I’m a bit of both.

Who has been your biggest inspiration?

I admire statisticians like Martin Bland, Doug Altman and Stephen Senn, whose motivation and preoccupation is to make sure that medical science is done right.  I do also respect the mathematical ability of the statisticians who are determined to create new methods, new formulae and new maths; but actually, usually what medical science needs most is a careful use of the methods and formulae we already have.

A great example is one of Martin and Doug’s classic papers, advocating the approach we now know as the “Bland-Altman difference plot” (Lancet, 8th February 1986, page 307-311).  They weren’t motivated by creating new methods (Martin Bland wrote later that he didn’t even know if it was new.)  They were driven to stamp out an all-too-common erroneous analysis.  That paper has been cited tens of thousands of times and I’m sure it will continue to inspire future medical researchers to use graphical methods in a context where a correlation coefficient misleads.  For me, that’s inspiring in a way that no theorem, no matter how clever the algebra, could ever be.

Why did you get into EBM?

That’s an odd question to ask a statistician.  Why would you base medical decisions on anything but evidence?

What do you feel has made the most difference in EBM?

Again, that’s an odd question to ask a statistician!  You might expect me to answer that great statistics are the key, but you can’t work alongside the team from healthtalk.org, as we do at CEBM, without realising that qualitative research has a role to play too.

Describe your approach to research in three words.

What do you like most about teaching?

That moment when “the penny drops”, as the saying goes – that moment when students faces respond as something makes sense for the first time.  Sometimes it’s something they thought they understood already but didn’t; sometimes it’s something they thought they’d never be able to grasp.  Some of our students are daunted when we hand them a calculator in a teaching session, but by the end of the session they’ll be the ones most delighted to find that actually, they can do this.  No-one with the IQ to become a healthcare professional lacks the ability to understand statistics; they just need the right moment, with the right teaching for them.

Do you have any regrets about becoming a doctor/DPhil?

No.

What has been your most innovative piece of teaching?

When we hosted the Burwalls annual meeting for medical statistics teachers (yes, there really is such a national meeting), I organised a number of events to encourage innovation including a teaching competition.  The competitors were challenged to make the best five-minute introduction to statistics for medical students.  (I entered my own competition and came second – but there’s no shame in coming second to CEBM’s excellent Prof Daniel Lasserson!)  I was especially pleased that we recruited current and new medical students to be the judging panel at a statistics teaching event.  Great teaching has to be driven by the needs of the audience.

When are you having the most fun at work?

When I’m teaching one of my favourite sessions, and the class are coming alive with ideas and contributions.  The exciting thing about the Graduate School in EBHC and Research Methods is that the students in the room at any one time are so diverse, with backgrounds in anything from midwifery to neurosurgery, and they all have professional expertise – often international expertise – in their own fields.  Making the most of that diversity, drawing out the commonalities as well as the differences in their research fields, showing how what we teach is relevant to all of them, adding value to everything they know already, and often learning as I go new perspectives from them and their diverse clinical fields – I have to be on the top of my game, as a teacher, with these postgraduate/CPD classes, but it’s a real buzz.

If you weren’t a doctor/teacher what would you be doing instead?

During the final year of my undergraduate studies, I got to the last round of interviews for the post of Clerk of the House of Lords – but no further.  Meanwhile back at University, my college (Oriel, Oxford) sent me to Dr Raymond Flood, a founding member of Kellogg College for statistics tutorials, which inspired me to do an M.Sc., which led to a PhD…  Twenty-five years later I’m a Fellow at Kellogg College, like Dr Flood … and very glad I’m not devoting my life to procedures and paperwork in Westminster.

What do you find hardest when teaching?

Perhaps keeping to time.  When students are engaged and enjoying the class, they interact and contribute, which encourages digression and all adds to the learning – but I hate to over-run by more than a couple of minutes: by the end of a ninety minute stats class everyone needs to reach the coffee room before all the chocolate biscuits are gone.

If you were given £1 million for research, what would you do?

I’d set up a one-time funding competition for orphaned research ideas.  Existing funders all have their own predefined agendas, for obvious reasons, but sometimes important ideas can fall through the gaps.  To get approved by my funding panel, you’d have to show that your question mattered to patients, that your research methods were strong – and that previous funding bodies had already turned you down!

What one resource should every EBM enthusiast read?

If you haven’t found Stephen Senn’s satirical articles about statistics in pharmaceutical research, written under the pseudonym “Guernsey McPearson”, you’re missing a treat.