Introducing the Harlem Harmonization Hackathon!
When we were at the NINDS TBI Classification and Nomenclature Workshop in DC last month, the lack of data harmonization between research datasets and data existing in the medical record came up several times. With the vision of Ari Ercole (Cambridge) and inspiration of David Okonkwo (UPMC) to “just get things done,” several of us got together to figure out how we might make this happen.
First of all, what's the problem?
Though there are local harmonization efforts such as the Common Data Elements for research, we lack a global harmonization schema for bedside data (physiology, therapeutic actions, etc.) and data in the electronic medical record in hospitals. The need for this is becoming more acute as we start to mine data across hospitals and previous observational datasets – such as the TRACK-TBI datasets – and correlate it with other factors in the medical record.
Why hasn't this been done?
The most likely answer is that a small group of very busy researchers need this done urgently, and the need has not been felt by a wider audience. And, at least for those who recently participated in the Curing Coma Campaign Common Data Elements effort, it’s a labor of love and not something you get career points for.
So, how do we get this done?
At our meeting in DC, we acknowledged the tedious nature of the work and decided we had to make the meeting as fun as possible. We decided to have two meetings, one in Europe and one in the U.S. To keep the harmonization theme, we decided to meet in Harlem, NYC and Haarlem, Netherlands. We plan to call it the Harlem Harmonization Hackathon. We will promise attendees t-shirts and other enticements.
While in DC, we got to know Shubhayu Bhattacharyay, who is finishing his PhD in Cambridge with Ari and starting medical school at Harvard this fall. His work is very relevant to this topic, and he was one of the main instigators for this effort. We decided to make the two of us yet another meme for harmonization; a young, bright Indian guy who represents new knowledge and the future, and an old, senile, white guy who represents all the legacy systems that we can’t get rid of and therefore need to interface. So, here’s to Peace, Love, and Harmonization:
Plans so far:
Ari Ercole has agreed to lead the Harlem Harmonization Hackathon effort, and Peter Smielewski will participate. We have the blessings of INCF (via Mathew Abrams), and it’s likely we would submit our output to INCF. We are contacting others to help organize the effort.
We are hoping to do the first meeting perhaps in the summer in NYC and the second one perhaps right after the Cambridge INTS meeting in early September (or later in the fall). All of this is tentative. A goal would be to report on our work at the Neurocritical Care Society meeting in October. Our role (Moberg Analytics) will be to help with logistics, publicity, and funding. We are working on some grants for travel support.
How can you get involved?
If you are working in harmonization on the data science or clinical side and want to have fun (i.e. get involved), please email dick@moberganalytics.com. We would like to have a planning meeting in the next few weeks so we can get some funding applications submitted in March. We will set up an informational website once we have some details figured out.
