Examples of data analytics in healthcare are sufficient. Nor is it lacking in high-end visions. But actual applications in clinical practice are not yet or hardly at all. This is because hospitals find it difficult to bring the pilots to production. What is involved? And how do you set a hospital-wide platform that can be used by anyone who want to gain more insights from data? 

Personalized medicine requires a vision of data

Every doctor wants to deliver the right care to the right patient at the right time and at the right price. We also call this personalized medicine. It is a word that every hospital has in its strategic plan. These plans are often vague about the elaboration of this. It is quite a lot when the translation is made to the resources that are needed for this, such as dashboards that clearly present predictive data to the medical specialist. And self-learning algorithms that predict which treatment will work best with which patient.

From pilots to production

As vague as the strategic plans are, the pilots are often so concrete. From apps that advise people with hay fever when they are allowed to increase their medication to image recognition algorithms that indicate whether suspicious things can be seen on a scan or in a biopsy. Just look at what is being developed during events such as Dutch Hacking Health. There are also more and more medical specialists who are curious about the application possibilities of big data analytics. In short, it buzzes and in the pilot-like atmosphere a lot is happening. But as already mentioned, that hardly translates to widely supported applications in clinical practice.

Vision is missing

This is mainly because a vision supported by the management is often lacking. If that vision and accompanying strategy do exist, then translating it to a functional design is relatively simple. Our customer UMC Utrecht shows this, for example, in the widely supported program Applied Data Analytics in Medicine, which is working on clinical applications of big data analytics. Ethics

Dat komt vooral doordat een door het management gedragen visie vaak ontbreekt. Als die visie en bijbehorende strategie er wel zijn, dan is de vertaling ervan naar een functioneel ontwerp relatief eenvoudig. Onze klant UMC Utrecht laat dit bijvoorbeeld zien in het breed gedragen programma Applied Data Analytics in Medicine, dat werkt aan klinische toepassingen van big data analytics.

Ethics

The most complex aspect of the step from pilot to clinical practice is that the validation of the solutions must be scientifically tested in one way or another. Running a randomized controlled trial takes time, then you are only a few years further. UMC Utrecht is therefore now working with experts in ethics and methodology to look at how they can shorten the lead time. They turn the issue around: is it still ethically justified to not use an algorithm that was successful in a pilot in practice? To remember the patient’s insights? They also do not want to rely entirely on the algorithm, it will always remain a teamwork with the medical specialist. He eventually makes the decision, together with the patient.

Develop a strategy

Does your hospital also want to get value from data? Then start by developing a vision and base a strategy on it. In addition to internal, also use external expertise to translate this strategy into a functional design and associated technology platform. Because if you do not have to do one thing as a hospital, the focus in the project is on technology. Focus on possibilities to improve care, using data and analytics. Then we arrange the rest.

Do not forget to let enthusiastic medical specialists, nurses, people with a sense of data and information and patients actively participate in the development of the vision and strategy. Involve them above all in the implementation. Because care remains people’s work.

Erik Fransen (Connected Data Group)

Laurens Faassen (Health Consultant Open Line)