Big Data in Medizin und Gesundheitswesen

Abstract

Healthcare is one of the business fields with the highest Big Data potential. According to the prevailing definition, Big Data refers to the fact that data today is often too large and heterogeneous and changes too quickly to be stored, processed, and transformed into value by previous technologies. The technological trends drive Big Data: business processes are more and more executed electronically, consumers produce more and more data themselves - e.g. in social networks - and finally ever increasing digitalization. Currently, several new trends towards new data sources and innovative data analysis appear in medicine and healthcare. From the research perspective, omics-research is one clear Big Data topic. In practice, the electronic health records, free open data and the "quantified self" offer new perspectives for data analytics. Regarding analytics, significant advances have been made in the information extraction from text data, which unlocks a lot of data from clinical documentation for analytics purposes. At the same time, medicine and healthcare is lagging behind in the adoption of Big Data approaches. This can be traced to particular problems regarding data complexity and organizational, legal, and ethical challenges. The growing uptake of Big Data in general and first best-practice examples in medicine and healthcare in particular, indicate that innovative solutions will be coming. This paper gives an overview of the potentials of Big Data in medicine and healthcare.

Ethical Implications of e-Health Applications in Early Preventive Healthcare

“…Summa, 2020 : 98; Braun and Dabrock 2016 : 326). These merges could arise, for instance, by linking health data to lifestyle choices or social environment data from social media, forums, blogs, or specialized communities ( Rüping 2015 : 794; Krüger-Brand 2015 : A1026f. ; Müller and Samerski 2016 : A1749).…”

“…Furthermore, it has been argued that with the use of apps and digital infrastructure, risks for children could be better captured and lead to more research data and better access to existing knowledge (see e.g. Rüping 2015 ). Increased initiatives could even promote “deep medicine,” as Eric Topol (2019) suggested.…”

Digitale Gesundheitsanwendungen – Rahmenbedingungen zur Nutzung in Versorgung, Strukturentwicklung und Wissenschaft – Positionspapier der AG Digital Health des DNVF

“…Häufig werden die 3 "V's" Volume (Menge), Velocity (Schnelligkeit), Variety (Vielfalt) zitiert [9,10]. Diese Merkmale sind jedoch eher selbsterklärend [11], daher scheinen folgende Merkmale von größerer analytischer Schärfe [12,13]: ▪ die Daten sind umfassend, unstrukturiert und offen, ▪ sie werden automatisiert zu anderen Zwecken als zur Analyse erhoben, ▪ stammen oft von Konsumenten/Patienten selbst und ▪ greifen oft nicht auf Hypothesen, Modelle oder Kausalitäten zurück. Diese Eigenschaften werden gelegentlich durch weitere "V's" wie "veracity" (Datenqualität) und "value" (Wertschöpfung) beschrieben [14].…”

“…Die Nutzer hatten nicht nur ihr Verhalten geändert, sondern sie hatten (im Wissen um die Nutzung ihrer Daten durch Google) ihr Nutzerverhalten als Reaktion auf die Intervention, nämlich die Analyse durch Google, verändert -der klassische Fall eines Interventions-sensiblen, sich über die Zeit verändernden, komplexen Kontextes. Die Ansicht, man bräuchte, um Wissenschaft zu betreiben, nur in einem quasi unendlichen Meer von Korrelationen "schwimmen", die sich automatisch der Wirklichkeit anpassen, ist also kaum zu halten, zu groß ist die Gefahr von Scheinkorrelationen[12]. Allerdings können auf diese Art entdeckte Korrelationen ihrerseits den Ausgangspunkt für eine Hypothesen-gestützte Big Data-Analyse darstellen.…”

Interpretable Clinical Decision Support System for Audiology Based on Predicted Common Audiological Functional Parameters (CAFPAs)

“…By using large amounts of data, precise classifications become possible that rely on better statistics than a human expert could achieve with a limited number of seen patients. In addition, new relationships in the data could be discovered which the human expert did not know before [ 10 ]. A disadvantage is that the CDSS can only represent what it has learned in training, and therefore, it could miss some special cases.…”