Big data analysis creates enormous improvements in medicine and helps patients.
Big Data

How big data can revolutionize healthcare

Experts are confident: Data analysis can benefit all areas of the healthcare industry enormously.
In the search for a new active ingredient, costs can currently run to several billion EUR. Experts believe this figure can be reduced to a few million in a few years. Cancer treatment – in the future it will be more effective with more individual methods, say the doctors at the National Center for Tumor Diseases (NCT) in Heidelberg. Or with mental disorders as well, where behavioral patterns can be accurately identified by comparing thousands of patients. One thing is certain – the successes to be expected with the massive use of big data analytics in the healthcare industry will be gigantic.

Europe-wide database

Experts have already taken a hopeful step forward for many seriously ill patients. Leading European scientists are currently working on a database that enables a more reliable prognosis with hematological diseases, and allows decisions on the best possible treatment for the individual patients. To do this, the experts collate anonymized patient data on diseases such as leukemia or lymphomas (lymph node enlargements). This project known as HARMONY (Healthcare Alliance for Resourceful Medicines Offensive against Neoplasms in Hematology) was co-developed by Professor Lars Bullinger, Oncologist at Ulm University Hospital. The "Innovative Medicines Initiative", a public-private initiative for the rapid development of safer drugs, funds the project with a total of 40 million euros. During the course of this big data project, 51 partners from eleven European countries collate anonymized patient data.

Find order – discover correlations

To do this, the scientists line up a series of promising projects as data analysis can demonstrate its real strengths in medicine and the healthcare industry in particular. The big data experts at consultancy firm PwC point out: "More and more data-producing technical equipment is being used during treatments. Examinations such as X-rays, CTs, MRTs, blood tests or dialysis generate a massive amount of heterogeneous data. Together with doctors’ reports, the personal patient data, their historical disease progression and the costs for these treatments, this produces an actual 'sea' of data." Furthermore, medicine is virtually made for this with the heterogeneity of the different data sources: "Big data technologies can be optimally used here to find order in this unstructured data, to discover correlations and consequently reduce costs for healthcare. During a treatment, for example, the data of similar patients can be used to determine patterns. These then contribute to making the right decision on which treatment methods and drugs a patient needs. A doctor can consequently introduce an individualized treatment with similarity analyses," says PwC expert Barbara Lix.
Dr. Carsten Bange
Dr. Carsten Bange

Performance capacities far from exhausted

The performance capacities of big data in medicine are far from exhausted with the thematic blocks referred to by Barbara Lix. Big data expert Carsten Bange identifies numerous other methods, which, from a technological point of view, must now only overcome some smaller hurdles (see interview). More complex here are the non-technical areas, such as data privacy and data law. A very pragmatic approach to resolving these problems was recently developed by the Swiss Academy of Engineering Sciences. Prof. Dr. Christian Hauser of the Swiss Institute for Entrepreneurship (SIFE) at HTW Chur, who headed up the project: "The legal framework conditions are from a time in which big data applications were still unimaginable. This enables the enterprises to position themselves ethically." The scientist is therefore asking on one hand for the laws to be adjusted to enable a more beneficial handling of patient data. On the other hand, he also advises companies to not only consider the 'business case' when using the technology, but rather to also identify conflicts early on and include an 'ethics case' in their considerations. To do this, the Academy developed eight ethical standards and values, such as "Control of own (digital) identity", "Informational self-determination" or "Contextual integrity", with which the healthcare industry can leverage the full potential of big data, without neglecting patient protection. Big data can therefore lead to success in the healthcare industry with simple principles.

Prevention of child abuse

How this could be done here in this country is illustrated in particular by the two data specialists, SAS and Mindshare Technology. The jointly developed technology will support social workers in "recognizing, practically in real-time, how high a risk or actual danger is for minors", say the two companies. The solution will "automatically alert the frequently overworked social workers" to imminent dangers for children, so they can immediately become active in urgent cases. Social services and child protection organizations actually receive notices of possible child abuse or neglect all the time, however how acute the danger actually is to a child mostly only becomes clear when further data sources are included. Police reports or health authority databases, for example, say SAS and Mindshare Technology.

Three questions to Carsten Bange

Dr. Carsten Bange is Managing Director at BARC GmbH. The Business Application Research Center (BARC) is a leading research and consultancy institute for corporate software.
Mr. Bange, why is big data incredibly important for the healthcare industry?
Because the possibilities for using data analysis are so diverse. Medical equipment and devices will be equipped with sensors in the Internet of Things, which will improve their quality. The equipment will become even more reliable and can be maintained with predictive analytics. With deep learning methods, image-providing procedures can be optimized, as the image information can be compared with millions of other images. And crimes such as insurance fraud are already detected quicker today thanks to data analysis.
The extensive spread of the topic is nonetheless still rather difficult?
Yes, and on one hand that is due to the vexing issue of patient data. Of course, we must also act here with plenty of sensitivity and data privacy. On the other hand, the fragmentation of the entire healthcare system causes problems here. It is better to have data, of a patient for example, together and to organize processes such as billing via data exchange platforms. Furthermore, the entire health market is actually really encrusted. There is a still a lot to do here, but ultimately, I am confident that this will happen, because it is a basic requirement for the success of big data projects in this environment.
In what way?
Well, a good deal of the value of analytical methods is in the ability to link data with one another. Only, when I can extract and link data from many sources I gain a high level of value creation from it. Only then can I develop the data models that have an especially high level of information value. I am very sure, that this can be guaranteed with strict data privacy and data security.

Further Articles