So I have just attended my first datathon aimed at “bringing inter-professional critical care trainees and clinicians (data novices to gurus), statisticians, data scientists and administrators together in the same room to tackle clinical problems with large datasets. The datathon aims were to generate patient-centred or service-centred projects and research which can be published in scientific journals and ultimately make a real difference to patients.”
I went with the advised ‘open attitude’ and ‘to collaborate’ mindset and it did not disappoint. From initial ideas brainstorm from those attending, and then linking in to a project of interest or specialty, this was all spontaneous. Next came the collaborative group work formed from a team of medical, nursing, librarian, government, dietitian, researchers, statisticians and data scientists discussed and brainstormed to generate a meaningful research question. This was all supported by research, medical, scientist ‘gurus’ to challenge and advise on topics and frame the research process.
Having minimal research experience, a dataset of 1.5 million admissions is just scary and needs to be managed with the respect like you were directly caring for each of these patient admissions. Time for the data scientists to come in and within hours had provided a initial dataset with required variables (cut down to 100 after much discussion) from 1 million admission episodes. Many cups of coffee were required to keep up with these specialists. No matter the level of experience, all aspects were fully explained so the group were all up to speed.
Big data analysis and time for ‘The Statistician’ to come and make some meaning from this huge amount of data. This was all done before the morning coffee break, that is time critical working.
Time to review the results and look for outcomes and demographics with important findings. Comparing these results to past research and knowledge provided a where were we at picture, to look at where are we now.
Ready to Rumble
How To Pitch Your Presentation by @MishManners:
- Frame the problem
- Current solution / knowledge
- Your solution!
- The ask
The datathon showed you need a collaborative approach to complete such a research project. Multi-disciplinary working, we must do more. Can the team continue to collaborate and finalise the work to publish and share these findings with a wider audience? As for another datathon, I would thoroughly recommend, even for a lost weekend of days off.