The digitalization of medicine and health information is changing healthcare today. Potentially, Big Data analytics allows diagnostics, therapy and the development of personalized medicines to provide unprecedented treatment options (video). Yet, the task is daunting: biomedical information is available in structured and unstructured formats; structured databases contain highly varying types of content (patient records, omics data, literature, device data, to name a few), and on top of that multiple data standards exist.
For various reasons, in healthcare, digitalization has been progressing slower than in other industries and organizations. In handling biomedical Big Data, privacy concerns and data security are high on the radar of all stakeholders, foremost the patient/consumer. This has led to a highly-regulated industry, which puts stringent limitations on the applications and methods used. There is considerable patient skepticism to some of the advances (example: telehealth). This is also not just a challenge about mining of Big Data with powerful algorithms to advance precision medicine, but also about digitalization in everyday practice, such as the question of how physicians and patients can already be supported by digital applications and content today. There is a need to “combine-to-mine”: to effectively bring all relevant information together – perhaps leading to predictive maintenance for patients.
Numerous opportunities are associated with this health-data revolution. Machine Learning and medicine seem to be made for each other. The integration (example: CBMed) and near-instant-response analytics across very large datasets (example CancerLinQ – video) can impact patient outcomes (while containing cost. Example Gustave Roussy), by supporting care-givers and researchers to test and discard hypotheses more quickly. Physicians want to be enabled to compare a patient to other similar patients, to learn from best practices for treatment from their peers, across large sets of parameters. Direct and real-time interaction between physician and patient is likewise becoming more important, allowing the physician to monitor and support a patient practically ‘live,’ instead of an interaction once every few weeks. Similarly, researchers want to interpret large datasets to find trends in drug-candidate behavior, treatment regimens and clinical trials. Methods and systems that facilitate these functionalities must be open, flexible and extensible, and must support and integrate with legacy tools and processes. They must also be adaptable to the digitalization maturity-level of an individual organization.