Medical data science research and development will often never make it into solutions that are implemented in clinical practice. There are multiple different barriers causing this gap, including issues such as lack of involvement of clinicians at the early stages of projects, differences in research culture, difficulties in designing appropriate trials for evidence generation, regulatory obstacles, etc. All parties are losers when this translation does not happen.

The data scientists will never see their research come to fruition and will not be able to plan their research progression in a direction supporting clinical impact. The clinicians will not harvest the otherwise promising benefits from data science research and will not have a path to bring their actual needs forward to data scientists. Finally, overall, the research investment fails to fulfill its potential for actual societal benefit.

In this workshop, we will identify some of the barriers and discuss potential strategies and solutions based on the organizers’ experiences within medical imaging, as well as participants’ own cases.

Program

Introduction to session (5 minutes)

Bring your own case (15 minutes)
– individual exercise followed by group interaction to identify barriers and experiences from participants

Lectures (10 minutes each):

• Retrospective data for understanding patient outcome using AI,
  Sarah Stougaard, data science PhD student, Odense University Hospital
• Prospective clinical trials of AI based methods for evidence generation,
  Stine Korreman, Professor of Medical Physics, Aarhus University
• Need-driven innovation in the clinic,
  Jasper Nijkamp, (Interim) Head of Innovation, Aarhus University Hospital

Group work based on cases brought by participants (20 minutes)

Plenary presentations of group work and discussions (15 minutes)

Target audience

The workshop is meant to be for anybody working with data science research with potential application in the clinic. The organizers are mostly focused on medical image analysis, but the principles of the topic should be more broadly relevant.

Workshop outcome

The intention is to share experiences and build networks between people working with medical data science. Furthermore, we aim to assist participants in getting closer to potential clinical application, by understanding and identifying barriers and discuss possible solutions.

Level

The workshop is aimed at intermediate to advanced levels, at any stage of career. As there will be a “Bring-your-own-case” element of the workshop, it is ideal that at least some of the participants are actively working on projects that are aimed at clinical implementation.

Organizers

• Professor Stine Korreman (lead), Aarhus University, stine.korreman@clin.au.dk
• Associate professor Christian Rønn Hansen, Odense University Hospital, Christian.roenn@rsyd.dk
• PhD student Sarah Stougaard, Odense University Hospital, sarah.stougaard@rsyd.dk

All three organizers are associated with the Data Science Infrastructure in Radiotherapy (DESIRE), supported by a grant from the Novo Nordisk Foundation