In my opinion, one of the largest (and foundational) obstacles to both interoperability and actionable analytics lies in the path that data follows– provenance. The healthcare industry generally has a poor grasp on where data goes, how it moves, and what it changes into in traversing care-facing applications. Even if we were to resolve the technical and privacy-related aspects of interoperability, other questions remain.
For example, does a LOS (length of stay) mean the same thing at Health System A as it does in Health System B? What about the definition of an inpatient visit? Does it count if a patient stayed overnight in a bed but was under observation and thus billed as an outpatient encounter? Since many organizations struggle in getting consensus on such concepts internally, I would doubt that translation between organizations would be successful.
The ONC (Office of the National Coordinator) recognizes this in promoting the Oh, the Places Data Goes Health Data Provenance Challenge. $180K in prizes are within reach of those who can come up with the best ‘demonstration’ of health data provenance via both a white paper and a prototype.
This is a fascinating contest in many ways. Sadly, however, I feel the need to put on my Data Debbie Downer Hat:
- How does a technical prototype work when the main component is data governance, which is a framework and not a product one can point at?
- Technology alone isn’t going to be enough to solve the problem. Agreed upon business rules (e.g. a Robert’s Rules of Order for data) need to be defined. This is a socio-cultural activity.
- The maximum team award is $80,000. Putting together a prototype will cost more than $80,000.
- If a winning prototype truly does the job, it is worth a ton more than $80K. Add some zeroes.
- Data is plural. Should it be ‘the places data go‘? (Ok, Debbie Downer is done.)
I suppose successful proposals will be focusing on a clear niche business area or clinical scenario as opposed to solving all data provenance challenges. This makes sense given the stated goal of the Challenge is “...to inform ONC about provenance problems the industry incurs, why it is important and possible ways to solve them.”
So it seems that they don’t intend on finding a solution that ‘cures world hunger.’ To keep with the metaphor, it sounds like the ONC wants to know more about who is hungry, why they are hungry, and what they may want to eat…. of data.
On a positive note, this challenge highlights the difficulties with data provenance for those who may not realize the gravity of the issue. At times, data scientists can be a touch naive with their aspirations in mining healthcare data. But maybe one of them will surprise us and come up with a great solution?
I’m definitely open to that.