Sharing personal health data
Comparing individual health data with group data allows doctors to give personalised advice and patients to learn from one another's experiences. Wessel Kraaij, Professor of Applied Data Analytics, shows how personal data can have a valuable predictive function. Inaugural lecture 24 February.
Anyone visiting a doctor will generally receive treatment that has been shown to have a beneficial effect on test subjects in clinical studies. Research from the US indicates that outside the lab the most widely prescribed medicines are only effective in at most one in four cases. And in some cases the medication has an adverse effect. There are a number of different reasons for this, including genetic susceptibility, lifestyle differences and environmental factors which mean that patients respond differently to medications.
Wessel Kraaij, Professor of Applied Data Analytics, believes that analysing big data offers opportunities for improving the effectiveness of treatments and interventions aimed at prevention. He will deliver his inaugural lecture on 24 February. 'Collecting health data gives us more scope for a personalised approach. Longitudinal data allow us to make predictive models. The crux of the matter is that these datasets have to be sufficiently specific to relate to an individual, but at the same time they have to be general enough to have some predictive value.'
Collecting data themselves
Citizens and patients will in the future gather more data themselves and share their experiences. These data and experiences can generate important additional insights that will help determine the causes of illnesses and the effects of treatments. Kraaij is working on a secure infrastructure that will allow citizens to manage their data themselves and - with their permission - make their data available for research.
Data are valuable not only for combatting illnesses, but also for preventing them. Kraaij mentions the example of the COMMIT/SWELL project, in which highly educated subjects monitor their mental and physical health themselves on the basis of data science. Many knowledge workers are affected by burn-out at some point in their working life. With a ‘digital alter ego’ - that measures such factors as your activities, tiredness and physical condition - you can build a picture of your habits and their negative consequences and can also determine which habits have more positive effects. This kind of intervention enables people to modify their behaviour and develop a more healthy lifestyle.
‘The main challenge is to convert all kinds of dissimilar data into comprehensible health information,' Kraaij says. 'This is something we are working on in lab studies. We have measured the heart rate and skin conductance of test subjects while at the same time filming their facial expressions and posture with 3D cameras and standard cameras. What we found is that when people are under stress, this is clearly visible in their posture and facial expression.'
Ultimately there has to be a balance between personal ownership of data and the possibility of combining and analysing personal data. Kraaij explains that plans are already under way for a national - and maybe in time even an international - infrastructure for analysing health data. This should make it possible to learn from datasets distributed for research purposes that cannot be stored centrally under current legislation on personal data and security.
In another project Kraaij – together with the Vrije University in Amsterdam, the Hogeschool Amsterdam and the Universidade Estadual de Campinas in Brazil – is looking at ways of giving wheelchair users personal advice on movement and nutrition. These people often have a complex range of medical conditions. By comparing their movement data with experiences of comparable wheelchair users, it will be possible to make personal recommendations that should eventually lead to a healthier lifestyle.
According to Kraaij, combining data will have benefits in other areas than healthcare alone. He is currently also involved in research to give policymakers tools for analysing large-scale data so that they can devise evidence-based policies. His research focuses on the metropolis of Rotterdam-The Hague. Together with public administration experts from Leiden University and the Center for Big Data Statistics at Statistics Netherlands, he is working on developing real-time indicators for policy, so that the effects of interventions in such areas as mobility and sustainability will be made available more rapidly and the learning cycle will be shortened.