Big data and Internet technologies are expected to play an important role for the implementation of a human and affordable health system. However, humans have to be in focus when designing the technology and its use. A decentralized approach and digital sovereignty of users are decisive qualities of the planned technology investments. We are all responsible for requesting and promoting these design criteria for the use of Internet technologies in our health system.

Big Data – Does It Really Help Our Health?

„Dr. Supercomputer“, this is how Süddeutsche Zeitung titled some new technology trends shown at CES 2016 in Las Vegas, USA bringing together big data and health applications (Kathrin Werner: Doctor Supercomputer; in: Süddeutsche Zeitung, #9, 2016-01-13; http://www.sueddeutsche.de/wirtschaft/digitale-gesundheit-doktor-supercomputer-1.2814450?reduced=true, accessed: 2016-01-18, 14:30 UTC). Again it became clear that new functions are evolved based on what is doable and are then presented as wonderful visions for all mankind – whether useful or not. This seems to be in synch with a similar report from the conference Digital Life Design 2016 in Munich, Germany: Andrian Kreye: “Those who still have visions should go see a doctor” (Andrian Kreye: Wer jetzt noch Visionen hat, sollte zum Arzt gehen; in: Süddeutsche Zeitung, #14, 2016-01-19).

Today’s sensor technologies are still at their beginning. Nevertheless, sensors combined with latest computer software and the Internet (high smart phone performance, big data, network coverage, and growing transmission speeds) provide vast opportunities to help people in many prevalent cases of sickness or need for help. Here are a few proof cases:

• Diabetes is still spreading – according to „Bild der Wissenschaft („Diabetes still spreading” [„Diabetes weiter auf dem Vormarsch“]; in http://www.wissenschaft.de/home/-/journal_content/56/12054/938205/, retrieved 2016-01-19, 17:34 UTC) roughly 350 Mio people globally are diagnosed with diabetes. Continuously measuring the glucose level of people with diabetes, hypoglycaemia can be mostly avoided and the level of insulin in a person’s blood can be leveled out during the day. IBM expects that the Watson “computer” can predict a potential glycaemic shock at least 3 hours ahead of time based on millions of data from diabetes patients reflecting their daily life, eating habits, and physical condition.
• Prostate cancer is the second most prevalent kind of cancer in men – in 2012 the WHO estimated 420 new diagnoses of prostate cancer per year for each 100.000 male inhabitants in Europe (in: http://globocan.iarc.fr/Pages/fact_sheets_cancer.aspx; retrieved: 2016-02-21, 16:00 UTC). Available methods of treatment have a strong influence on the quality of life of men with prostate cancer. The optimal method is dependent upon the specific kind of prostate cancer, the individual’s situation, the partner relationship, and the statistical life expectation amongst others. It can be assumed that by collecting relevant data of a huge sample of men, mainly before their first treatment, the optimal treatment method could be derived with a much better success rate as it is the case today.
• Demand for quality support for senior citizens is growing as the population in many countries is aging (for example European countries, USA, Japan; example case Germany: http://m.bpb.de/nachschlagen/zahlen-und-fakten/soziale-situation-in-deutschland/61541/altersstruktur; retrieved: 2016-02-21, 16:19 UTC). In particular the support in case of illness or reduced physical and/or mental health at acceptable cost levels will stress the health systems of many countries. Robots (in their widest sense) may be the best way forward. However, the use of robots at the same time requires the continuous care (and associated collection of data) of the supported persons. We know from Dave Eggers’ book “The Circle” (Penguin Books, London, 2014; ISBN: 978-0-241-14650-7) how quickly positively perceived monitoring can turn into negatively perceived surveillance and control.

All above introduced solution concepts have in common a helpful functionality, which can be assumed to save lives or improve the quality of life of many people while at the same time create tremendous savings for the public health systems. However, the solutions are based on the central recording and analysis of relevant (mostly personal) data and do not take advantage of their decentralized, local processing – in the energy sector we see exactly the inverse trend: a move away from central giant power plants and towards the decentralized generation and use of energy! The inclusion of the users as owners of the data is neglected completely. Regulatory measures do not take requirements into account, which are implied by the concept of data sovereignty.

What do we learn from this? Reject all technical innovations and be daemonized as iconoclasts? Or take each technological trend at face value and thereby support the business model of big Internet monopolists?

Recommended: Decentralized Structures Avoid Data Misuse

Using available technologies we can implement solutions fulfilling functional requirements and cost efficiency of future health systems as well as the individuals’ requirements with respect to data sovereignty. How could a decentralized solution look like in case of the above examples?

• A small device close to the body can predict dangerous hypo/hypoglycaemic blood levels if correctly parameterized. It is not necessary to uniquely identify an individual as the device is owned by its user. If the device recognizes a potentially hazardous blood level, it can automatically call an emergency number and at the same time alert its user. The diabetes patients decide whether to make their body data anonymously available to a central analysis computer for further refinement of the prediction parameters. This could open new service fields for doctors, health insurances, nursing services, which help the patients AND are trusted by the patients. Examples are providers of “trusted” apps for the devices, anonymization of the data or confirmation of transferred anonymous data. Regulating the use of the accumulated data is a must.
• The treatment of prostate cancer can be adjusted to the specific needs of each individual patient assuming a sufficiently high number of case data. But in order to do this, it is sufficient if a statistically relevant subset of patients make their anonymized data available for further analysis. The identification of an individual is not necessary. Also the required data do not need to be continuously collected using body trackers. A trusted entity (doctor, health insurance, nursing service) can monitor the anonymization as well as assure the consistence of the transferred data. By analyzing the anonymized data the best treatment method can be developed in cooperation between big data services, doctor and patient taking the current patient case into account. The access to the analyzed data (for treatment development) could be paid for by making a patient’s anonymous data available for scientific analysis. Again, this scenario could open a wealth of new business opportunities for service providers.
• You may have guessed already: also helping senior citizens can be implemented in a trusted and cost efficient way. The individuals take advantage of local devices under their control with centralized analysis and management technologies. The exchanged data remains under the control of the person. The device is “personalized”, i.e. it is parameterized (in artificial intelligence speak it has “learned”) to fulfill the needs of its user. If an emergency situation is coming up, the device can summon the best service on site. This does not require the existence of a central, omniscient “Circle” to monitor the individuals in need of help – the individual­ remains the sole point of control.

What are the characteristics of the recommend solutions? They all use local, personalized devices under the control of the individual. The devices are parameterized for optimal operation with data from big data services. Such big data systems develop decision strategies using immensely rich sets of (anonymised) user data and make those strategies available to users in the form of parameters for personalized devices. Beyond this, there are many other advantages compared to today’s centralistic approaches as enforced by Google, Facebook or other Internet players:

• Control over data and devices remains with the individual user, i.e. data sovereignty is maintained.
• The local device can be easily switched on and off.
• The localized approach opens new business opportunities for entities which operate close to the users and their needs.
• The dangers of hacker attacks on central infrastructures are minimized (no interest for anonymous data), impacts on users are close to zero.
• The European economy with its many SME players and their decentralized operation is strengthened and may lead to similar approaches in other parts of the world.
• SME companies can increase their competitiveness compared to big Internet players with their customer mindedness and being close to their customers.
• Big data technology and the use of large volumes of anonymized data are an important part of the solution concept and its associated ecosystem.

We could show that data sovereignty and intelligent business opportunities in the Internet can complement each other, when latest technical evolutions around sensors, data analysis, big data, personalized information and communications technology are intelligently combined. Advantages for the individual are obvious, but also society in general can profit. New business models become possible taking advantage of the decentralized use of modern technologies. At the same time, the disadvantages of centralized structures – hacking, misuse of data etc. – can be avoided. So what are we waiting for?

And now?

As with the energy change and its transition to the distributed generation and management of energy, the implementation of decentralized structures in forward looking, profitable, and creative Internet based services and solutions is extremely complex and difficult. The focused and concentrated move of many actors is required to be successful. This includes the discussion with political players to stand up against data misuse by international conglomerates. We need to be clear about user interests: Improvements to the health system without the monopolistic accumulation of most personal data. Thirdly, we as users should actively go for alternatives to the monopolies’ offers and use them - for example an app offered by our trusted health insurance. And last but not least, we should eventually become aware of the value of the data which we voluntarily and/or unawarely give away to Google, Facebook and others - we should get more than a few gaming apps for our smartphones!