Data: the key to personalised care
This article has been translated by Commit2Data staff.
Personalised care is becoming increasingly important in the Dutch healthcare landscape. With personalised care, we can better offer the right care at the right time. Various ICT techniques are already being used for this purpose and we can expect a lot more in the near future. In particular, the use of data and data analytics leads to promising applications for personalising care. In this article, we list a number of them and take a look at the future of the healthcare landscape.
author: Pieter van Kuilenburg, project advisor at ECP | Platform for the Information Society.
Personalisation of care is already being worked on in many places. Personalisation makes diagnosis and monitoring more person-specific and contributes to the most suitable treatment and medication for the individual patient. In the long term, this will also contribute to controlling the rising costs of care, since treatments will be more suitable for individual patients and therefore more effective.
Fully personalised care is a great goal, but there is still a lot of work to be done before that happens. Within the research and innovation programme Commit2Data, various projects are working on personalising care by means of data-based applications. This can be done in various ways: from decision support for the doctor to personalised lifestyle advice for the patient, and from support with self-management to the personalisation of prostheses.
Deciding together
A correct assessment of risks helps doctors and patients to make decisions together: data plays a valuable role here. In the SNOWDROP project led by Ameen Abu-Hanna (University of Amsterdam), techniques are therefore being developed to extract relevant information (including free text) from the patient's file in order to develop a prediction model for fall risk in the elderly.
This project combines expertise from geriatrics, general practice, medical informatics and communication science. The aim is to optimally integrate the developed software into the existing workflow of the care providers and to find effective ways to communicate with patients.
Serious gaming
Data research is also being used for a new approach to persistent problems, such as lifestyle change. Lifestyle adjustment is an integral part of the management of diabetes, for example, but for many patients this is a complex task. Moreover, for some of them, a 'dry' explanation by healthcare providers is not effective.
In the DiaGame project, the team of Natal van Riel (TU Eindhoven) is developing a serious game for self-education of people with diabetes. By living with diabetes in the game, they learn to cope better with their illness. The game is aimed at changing habits whose health effects have been scientifically and clinically substantiated. The engine of the game is an existing simulator. By applying data techniques, it is fully personalised for each player and made as realistic as possible. This makes the game very appealing to the patient's imagination and makes internalising what has been learned easier.
Learning from brains
Developing new approaches is truly pioneering. In recent years, for example, large databases of brain scans have been created. These could then be used in psychiatry for objective diagnosis and for predicting whether a patient would benefit from a particular treatment. Before that happens, however, the right analysis techniques have to be developed.
This is happening in the project Deep Learning to extract biomarkers for the diagnosis and personalized treatment of neuropsychiatric disorders. Researchers from the Amsterdam UMC (led by Guido van Wingen) are developing advanced data techniques to analyse brain scans. The results of these analyses are used to predict how patients with severe depression respond to a specific therapy. This can then be used to select the most appropriate treatment for the patient. The new analysis techniques can possibly also be used to personalise the treatment of other psychiatric disorders.
Implants from a 3D printer
Data may be abstract, but some applications manage to literally convert them into tangible results. In the project Accelerating Mass Personalization in Orthopedics facilitated by Machine Learning and Bone MRI-based Digital Fabrication, Peter Seevinck and his team at the UMC Utrecht are working on the personalization of implants that are used for bone fractures, among other things.
Through a combination of various AI-based data techniques, information from MRI scans is used to make a synthetic CT scan. This synthetic scan is then processed and used as input for a 3D printer. In this way, personalised bone segments, titanium implants and instruments can be created, without the use of harmful radiation required for a CT scan. This aspect is particularly important for children. Ultimately, this approach contributes to smaller and simpler operations that promote faster healing and shorter hospital stays.
Safe and responsible
Personalisation using data automatically raises questions about privacy. That is why various projects are developing techniques to handle data in a safe and responsible manner. For example, in the CARRIER project led by André Dekker (Maastricht UMC), work is being done on algorithms that personalise lifestyle interventions for the prevention of cardiovascular diseases. To this end, data from general practitioners, hospitals and CBS are combined.
The research team makes use of so-called federated learning via the Personal Health Train approach, in which algorithms are brought to the data instead of data to the algorithms. In this way, the data itself does not have to be shared, but only the results of the data analyses are shared. Moreover, the researchers apply methods that allow them to perform calculations on encrypted data without decrypting it. During the development of the algorithms, all data thus remains encrypted and in the possession of the original controller, thus guaranteeing privacy.
How does it work?
Although fully personalised care is not yet a reality, the examples given in this article show that major steps are being taken in many areas. The Commit2Data research programme aims to valorise the results of the various projects in the direction of applications that have an impact on everyday healthcare practice.
A number of trends from these research projects will continue in the future. Data, for example, contributes significantly to the increasing focus on preventive care. The possibilities of making this data available in a safe and responsible manner are being further developed. This is important for the production of individualised prostheses, devices and medication. In addition, new data-sharing techniques can provide connections between different sources, which means that personalisation will be introduced in more and more areas, integrally.
