Machine Learning for Precision Medicine

Markus Wenzel, researcher at Fraunhofer MEVIS in the field of machine learning and instructor of the module Machine Learning for Precision Medicine of the Fraunhofer Transformative Digital Skills for Healthcare program shares some insights on the digital transformation of healthcare:

How will digital technologies transform the future of work in hospitals?

Overall, digital technologies can serve as a common exchange platform that can remove barriers between departments to the betterment of care delivery. Digital technologies are one important factor fostering trans-disciplinary collaboration. We have experienced in our projects how, for example, radiologists and pathologists started communicating with each other simply because their data was visible in the same computer application once the pathology department switched to digital slide scanners. The same is true for other disciplines, for example surgeons. Based on digital representations of the anatomy, perhaps even displayed on a mobile device or augmented reality headset, they can assess their priorly planned access paths and make sure that risk structures are at safe distance. Prior to surgery, they can even plan biophysical properties of devices they intend to implant, like tendon replacements or stents for vascular surgery. Computer simulations based on models of the physical behavior of tissues and devices are the basis for this.

How could the digital transformation reduce the strain for hospital personnel? Can it improve their work?

Strenuous, repetitive, and error-prone tasks can be supported by technologies like trained software assistants – often dubbed medical AI. They are today often based on machine learning techniques like deep learning, deriving powerful models from large data collections. Such machine learning models are suited for personalized risk prediction, support in differential diagnosis based on patient data and images, or accurate, repeatable measurements using image analysis methods on radiological exams. This development is currently well visible in commercial and research image analysis solutions for radiological tasks but also enters many health services, including some in apps on personal devices. For example, in underdeveloped countries with poor medical service availability, the power of small devices and wireless data exchange networks enables huge advances in care delivery to remote areas, for example, used in automated and remote reading scenarios for early tuberculosis detection.

Which benefits will digital technologies offer to patients?

The healthcare landscape can become much more transparent for citizens and patients.

With the introduction of the digital electronic patient record in many countries, the expectation is that many more options for digital health services will become available — besides the obvious benefit that my own health data is stored in one place, instead of distributed across hospitals, private practices, and other services I use. This allows health data mobility on small and large scales. So, when I have a health checkup in inpatient care in a hospital and aftercare with my private physician, the data no longer needs to be transferred in printed letters or a DVD but is „instantly available“. If I move to another country, the same would be true if all providers adhere to standard formats for data representation. Such standards have another benefit: Startup companies and small health care providers can develop solutions with broad applicability. For me as a patient, being able to give access to my health data to such a specific service provider for a specific purpose will allow me to choose among offerings – which can range from AI-based online services for everyday personal health tracking, a second opinion in a health issue I have, or easy comparison of cost estimates for elective treatments I consider.

You want to know more about the module? Visit M6 | Machine Learning for Precision Medicine

 

Kommentar schreiben

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert