Ethical Data Sharing? As I see it!
I love the phrase Ethical Data Sharing because it is so provocative. The most common reply to this tagline is, “but that’s an oxymoron.” So it begs the question, is it?
When you search for Hospitals sharing data, Project Nightingale is in the majority of top hits. Every hospital CIOs worst nightmare. This is an over-told story of what transpires when fractional ownership is not fully considered. While this has shell-shocked many hospital leaders from posting about data sharing, it has not stopped hospital systems from looking for solutions. Considering hospital systems have gross revenues in the Billions but operating margins of around 6%, they are scrambling to capture value. And they see value sitting right under their nose.
Now let’s flip over to the data user side. Who wants access to health data, and what do they want to do with it? Pharma companies need large volumes of data to understand rare diseases. smart radiological technology (AI diagnosis) needs tens of thousands of scans to get detection to a clinically accurate level. Notice, I didn’t say they need to hold patient data or look at patient files to advance medicine. So what do they need?
This question needs to match what risks are data owners are concerned with.” This is not a one size fits all problem.
It takes a collaborative agreement between data owners and data users to define the essential points in data sharing. One of the fundamental aspects of maintaining trust and balance of power is keeping the data with the data owner. This is the cornerstone of the new data-sharing paragram in which you don’t share the data. We are moving away from a world where you make a lengthy agreement and then hope nothing terrible happens.
If the data stays with the owner and they are granting algorithm-only rental access, the data user can still draw insights without violating stakeholders or increasing the risk of breaches. This shared understanding of Ethical Data Sharing allows for a new way forward with a trust-but-verify model.
Patients and hospitals are not against furthering medicine; they are against an imbalance of power that creates unanticipated risks. Those in medical discovery are not in the business of owning your clinical data. They are in the business of insight discovery. For truly good actors, this creates a happy path. Creating exclusionary criteria for bad actors and leaning on a trust-but-verify system gives data owners end-to-end control.
Promoting medical discovery without the typical risks in data sharing is possible. The second-order effects increased insight velocity. In changing nouns from “data velocity” to “insight velocity,” we create the happy path to what is next in Modern Medicine. Create the effect of data sharing without sharing your data.
* Krister leads operations at Sympatic. With Piers Nash Ph.D. and Christian Dompierre, they have been building the tools for Ethical Data Sharing.