Posted by on Friday 31 March 2017

Aaand relax! Thoughts on the REPOSE trial

I've been watching Twitter with some interest today, following a conversation about a recent piece of research published by top-notch diabetes Doc, Sheffield's Dr Simon Heller.

I first heard of the REPOSE study (Relative Effectiveness of Pumps Over MDI and Structured Education) in 2016 when I was able to sneak into the Diabetes UK Professional Conference, where Dr Heller published some of the early findings. It has now been formally reviewed and published in full and, as much as I am able to understand it, I find it fascinating reading.

If you are interested, you can read the study here:

The intention of the study was to evaluate whether insulin pumps or multiple daily injections (MDI) gave the better outcomes for people living with type 1 diabetes who have received comparable training and support. Essentially they took over 300 people, spread across multiple centres almost all of whom had HbA1c results over 7.5% and offered them structured education/training in intensive insulin therapy. Of those that undertook the training (267) the study then followed 260 people over 2 years to see what changed. Of the study group, the pump vs MDI split was approx 50:50.

What did it show?
Here are some of the results that caught my eye...
  • Supporting people with effective, detailed structured-education/learning/training* in the complexities of managing type 1 diabetes improves outcomes whatever method people use to deliver their insulin (* delete the phrase/s which annoy you the most)
  • On average people do a little better on insulin pumps (-0.85% with pump treatment and -0.42% for MDI)
  • When results are adjusted for differences which could have skewed the data, the pump 'win' was pretty modest at around an extra -0.25% (-2.7 mmol/mol) reduction in A1c for pump users vs the MDI group - the averages may, of course, hide significant individual variation
  • Pump users additionally scored better for 'quality of life' and 'treatment satisfaction'
What I find interesting, reading between the lines, is that this study almost exactly replicates my own meandering journey towards insulin pump therapy. This whole blogging endeavour sprang out of a non-Wittertainment related 'unfortunate event' which acted as a catalyst for me to re-assess my own relationship with type 1. Chiefly that reassessment was that I was nothing like as good at dealing with it as I was allowing myself to think. I then went through a period of significant re-education, not by attending a formal course, but by my own experiments and learning from others living with T1D, many of whom were using updated intensive-insulin-therapy-type approaches. After much work, and many successes (including a reduction in A1c and elimination of severe hypoglycaemia) I realised that I had reached the limit of my MDI skills, and if I wanted to improve further - which in my case meant more or less maintaining HbA1c, but further reducing mild hypoglycaemia and glucose variability, then the next step for me was an insulin pump.

The small irony is that I had been offered an insulin pump repeatedly for going on 5 years at that time, but had never liked the idea. In fact, looking back, I think if I'd accepted a pump in the years before we started writing this blog, and before I had discovered the #doc - I am really not sure whether it would have done anything much for me. Or whether I would have just have tried it for a bit and then stopped using it out of frustration.

No magic bullet
One of my take-aways from the REPOSE trial is that insulin pumps (or any other diabetes technology for that matter) are never going to entirely 'fix' things. This was certainly one of the things that put me off pump therapy for years. If many of my errant results were down to 'user-errors' of judgement, I reasoned, what difference could it make whether that 2u, 3u, 4u dose was delivered by syringe, insulin pen or pump/cannula. It's pretty obvious really - or at least it should be - but a lot of the potential power and effectiveness of insulin pump therapy comes down to the way an individual thinks about their diabetes management. The techniques they use, the strategies they have been given to review and adjust on an ongoing basis. I attended a pump users event when I had only been using Artoo for a few months and was slightly shocked to meet people who had been using insulin pumps for years but had never used (or wanted to use) the combo/dual/square/extended bolus options. It may be that they didn't need to, or that they had never been shown how to. But the lack of curiosity was a genuine surprise.

Not for everyone, and not second best
It is all too easy inside the hothouse of the 'Twitter bubble' to let yourself believe that absolutely everyone is using an insulin pump, full time CGM, Nightscout, and open-source cloud-based Artificial Pancreas algorithm. If you understand even half of that sentence, you will know exactly what I mean. But the simple truth of REPOSE is that insulin pumps are just one option, and they will suit a particular type of person, with a particular set of approaches at a particular time in their lives. That doesn't inherently make them better or worse, and it certainly doesn't make them right for everyone all the time. I loved this post by everyone's second favourite ex-teacher-Libre-wearing-cat-loving-warm-ball-of-GBDOC-loveliness Adrian Long about his continued love of multiple daily injections. People can do brilliantly on MDI, and people can do brilliantly on insulin pumps. What matters most is the support and information/training they have been given. Diabetes is going to be infuriating, uncooperative and apparently willfully disobedient whichever insulin delivery method people use. What I need, in order to face those daily frustrations is a comprehensive set of strategies, and individually tailored personalised goals to make small incremental improvements towards better outcomes. In addition a good deal of understanding and moral/psychological support goes a long way to help.

The shiny gadget effect, sticks and carrots
I do find myself wondering about the possible catalytical nature of a new piece of diabetes technology. For some people, and I think I might be one, having a shiny new diabetes toy (or the promise of one) can re-energise them into a new, more active relationship with their diabetes management. It may also be that people who have never really fancied attending Structured Education, for a whole host of reasons, may decide to take the plunge in order to get access to the technology they are hoping will help.

I'm not altogether sure how I feel about this if I am honest.

I would hope that REPOSE leads to more individualised, supportive care, rather than people being forced to unnecessarily attend officially endorsed education courses purely to box-tick the process (and add delay into the bargain). My own journey towards pump therapy was excellent in that regard. In the pre-pump assessment I was offered the education, but in conversation it was decided that I was already using exactly the techniques and strategies that the course advocated, so that it was entirely up to me whether I thought it would be beneficial or not. I would not want people who might excel at pump therapy to be put off by a hardline education requirement... but at the same time, my own experience tells me that often you don't know what you don't know. And that many people who attend diabetes education expecting very little from it actually leave the course with their relationship with their own diabetes utterly transformed.

My slight worry in all this, is not what this study actually shows, but more how it might potentially be interpreted and skim-read - short version: pumps expensive and not much better. The current guidance over the use of insulin pumps (and when people may additionally benefit from CGM) is pretty clear cut. And yet, some people in some places find themselves having to jump through interminable hoops, or simply get enquiries brushed off for having 'too good an HbA1c'.

Insulin pump therapy really does work well for some people, but it is expensive and the 'working well' does not come automatically. What I hope comes out of all this is a greater level of support and assistance both for pump users and MDI whizzes to aim for those elusive  and aspirational treatment targets alongside a decent quality of life and an avoidance of diabetes burnout. With finite NHS resources it is only right that these therapies are used effectively, and clearly a vital component of that is the support, education and encouragement that people receive, not just initially, but on an on-going basis.

Simply blindly chucking technology at T1 is never going to work.

I'd be interested in your thoughts on this trial, and how you think it might impact you in your journey with diabetes. Please do leave a comment below.


  1. Great breakdown of a very long piece of study Mike, thanks for that. Unfortunately I had spent a couple of hours reading it myself and have come to very much the same conclusions as yourself on it. I am concerned that this study could be a set back for people who want to access insulin pump therapy, if their clinic is not particularly pro pumping, a skim read of this does come to the conclusion of lots of money for not a huge amount of return. That could have knock on effects possibly

  2. Hi Mike - thanks for a very interesting blog. We have gone for Libre & its transformed my 2 young adults. They are slowly getting more willing to engage in further education but its baby steps. The whole gig can be so knackering that its small wonder folk bury their head in the sand. I especially like the non-Wittertainment related 'unfortunate event' reference - Kermode & Mayo must have a corner of the church for T1's & their parents!

    1. Hehe! Thanks. T1's Transept perhaps? Good to hear about your two. You are doing brilliantly to keep them engaged :)

  3. HbA1c is, in my view at least, a pretty crude method of deciding which treatment is "better". When I was on MDI my HbA1c was very, very good; but it disguised a lot of up and down swings and hypos every day. By contrast, when I went on the pump my HbA1c results were "worse" (slightly) but my quality of life was vastly improved as my BG wasn't swinging around wildly all the time. So I have my reservations about HbA1c as a measure of which is "better".

    1. I absolutely agree Tim, and I think that the focus on HbA1c is certainly shifting towards including more subtle/richer markers including quality of life, exposure to hypoglycaemia, and (importantly) glucose variability alongside A1c. The snag, of course, is that when trying to pool data from difference sources, all the other things are measured and recorded differently, while HbA1c has an international lab standard and can be easily compared.

    2. Tim, this is such a good point! I am about to start with a minimed 640g and cgm next week, so doing a bit of a read of "real" blogs...and I had the exact same thought about the sole use of HbA1C as a sign of tight control or not.

      In the past when my HbA1C has been much lower than it is now (too embarrassed to post it actually) I have experienced many hypos and bgl swings... not a true indicator of tight control!

  4. Great post Mike, and as others have said before me, very balanced. One point to note about REPOSE is that anyone who met NICE TA151 (insulin pump therapy) criteria, or who was deemed to need pump therapy by the investigator, was excluded from this study. That puts a different skew on the findings. This study wasn't intended to undo everything about pumps that is already known, but rather to look at the idea that 'everyone' should have one. NICE TA 151 still stands. If anyone who meets (or believes they meet) the criteria is waiting an unreasonable time or struggling to get one, please contact INPUT via

    1. Thanks Lesley, that is a really important point, and well worth mentioning.

      The fact is that insulin pumps really can make a huge difference, and the subtlety and precision of doses, micro-corrections and carefully tailored basal profile(s) they afford can't easily be matched by the relatIvely crude tool that is MDI. But I guess all that precision comes at a cost of time and effort. And that cost can be difficult for people long-term. Plus there is a potentially increased exposure to major BG highs if a set fails. Don't get me wrong, I love my pump and would really struggle to return to MDI, but as you say - important to remember that pumps aren't for everyone, and that a pump in itself can be used effectively or ineffectively, just as MDI can.

  5. I've read the REPOSE study a couple of times, and the exact same things stand out for me, namely that while it was taken as a "Are pumps a magic pill?" type of study, it's very easy not to interpret it that way, as much of the twitter response demonstrated.

    Once again, we find ourselves in a position where the published data that NICE will take into account when making recommendations shows that those who aren't bothered either way probably don't benefit a huge amount, whereas there is no good evidence relating to those who want to use newer treatment methods and are engaged enough to ask.