Way to go HealthRI!

On Thursday 17th of January was the 4th HealthRI meeting, marking already four years of bringing together a number of Dutch health data research projects and infrastructure (such as CTMM TraIT, BBMRI, ELIXIR-NL, DTL, NFU etc.) under one umbrella organization. It was a great conference and this is a good time to reflect on what has been accomplished so far, and more importantly what's the way forward for HealthRI.

Lucien Engelen (well-known from RadboudUMC and REshape, his book "Augmented Health", and now working for Deloitte) kicked off with a presentation about innovation which I found quite useful. If only to introduce a number of words and metaphors to describing innovation. Here are some of my favourites:

  • SOY (the Shit Of Yesterday) is often what really holds back (implementation of) innovation, because it sucks up too much energy and budget. SOY is a term explained by Peter Hinssen in his book The Day After Tomorrow. The above figure also explains why it makes sense to focus on radical innovation.
  • The apple and the orchard: innovation is like an apple. It can be beautiful and delicious, but it's part of a larger ecosystem (the processes, people etc. where the innovation needs to be embedded). If you want to grow apples, you need to tend to the orchard!

Why is it so hard to build infrastructures?

In the Personalization track, we had a discussion about this question, also triggered by a presentation from a medical doctor showing all the systems he has to still manually enter data into today. I think that in general the amount of energy, money and persistence required to successfully build an infrastructure is underestimated. How often are (public and private) infrastructures delivered on time and on budget? Building and maintaining infrastructures that actually work in daily operation costs blood, sweat and tears.

It probably also doesn't help in that regard that the HealthRI community is quite researcher heavy, because researchers by their very nature are focused on new and innovative ideas, and once the concepts are understood already think ahead to the next challenge. However, to build infrastructure, it's not enough to create and demonstrate concepts. The concepts actually need to be implemented, and the resulting implementation needs to be maintained. Activities which will not yield any new scientific publications or insights in the short term. In the meantime, if the IT departments and CIO is focused on risk management on the SOY, it's not obvious where this investment is coming from.

Add to this that to come full circle with re-using routine healthcare data for research, we are also talking about a chain-computerisation problem (Jan Grijpink) which depends on collaboration and commitment of all main stakeholders to work, and it's clear why it is not easy to turn the vision of HealthRI into a day-to-day reality for doctors and nurses, patients and citizens, and researchers. (BTW: Grijpink's paper on chain communications (in e.g. the criminal law system) is actually a really interesting read and has quite some parallels with the FAIR principles and the Personal Health Train.)

In my webinar about health data networks from last year, I explained this by loosely reversing Conway's law: only by building the infrastructure with the input, involvement and commitment of all stakeholders you can ensure that it will actually be used, maintained and populated.

A solution built in isolation, even if it's the most well-funded, innovative and just generally splendid in the world, will not work in this context. In that regard it is very different from the winner-takes-all consumer market. And indeed, if the healthcare sector does not build and adopt such an infrastructure in time, it will likely be disrupted by a company that drives innovation by directly engaging the patients, probably bypassing much of today's healthcare system by providing better direct value to the end user. And if that happens, my bet is that the researchers, being the avid consumers themselves that they often are, will actually be the first to follow suit and leverage the opportunities and data that will come with this!

Where should HealthRI now focus its efforts?

In the general assembly, there was a poll about what the main focus for HealthRI should be, and a majority voted for 'Building out the infrastructure'. Of course, being from an infrastructure building company I was quite happy to see that, however, I actually voted for promoting open science and FAIR. Because I think HealthRI will have the most lasting impact in the long run, if it succeeds in building and constantly growing a community of people from all possible stakeholders that care about sharing and (re)using health data for the benefit of the patient and citizen, and share the same vision and principles.

A good KPI for HealthRI would be the number of (new) people that get trained every year in these core principles. Even most life science researchers still haven't heard about FAIR, data stewardship, data re-usability etc. Let alone the doctors, nurses, patients etc. Growing our community and building tools should go hand in hand, but tools come and go, strong communities last. Which is by the way why The Hyve's mission focuses on open source communities and not open source tools.


Let me close by saying that in general, I'm still very optimistic about the opportunities for HealthRI to achieve its goals in the coming 10 years. Via my job I'm involved in all kinds of FAIR initiatives in various countries and organizations, and yes it's true that some other countries have much better funding situations for national health data networks. But what HealthRI has going for it, and also demonstrated, is that it can leverage the Dutch spirit of networking to achieve consensus results ('polderen'). That includes leveraging European and global infrastructures such as ELIXIR, GO-FAIR, GA4GH etc. If HealthRI can build a powerful and inclusive community of people that share the same vision, and then add sufficient funding as well as a laser-focused execution plan and the right operational focus, it can be done! It's much more effective to do it in this order than the other way around.
And, as Lucien suggested here, maybe we should add the P of Pro-active to highlight that!