Earlier this month, I wrote a blog item laying out some of the problems Epic is facing in getting its clients to share de-identified patient data among themselves. In sum, I concluded that while it might have the right idea, getting providers to pony up their data is no picnic.
Of late, Epic has been selling its customers a vision of building “One Virtual System Worldwide,” a data-sharing network designed to support clinical science breakthroughs by making relevant data available. However, according to Chilmark Research analyst John Moore, the vendor has had trouble getting clients to participate.
Of course, that doesn’t mean Epic’s efforts are out of line. In fact, at least one of its major competitors seems to be on a related track.
Cerner has just announced that it’s launching a new initiative designed to give clinicians easier access to patient health data. Working in partnership with the Duke Clinical Research Institute, the vendor is rolling out its Cerner Learning Health Network, which is intended to automate data collection from EHRs and multiple additional sources.
The two organizations will kick off their efforts with a pilot project and study intended to evaluate the use and potential impact of proven therapies for chronic cardiovascular disease. For the pilot, Duke will use Cerner tech to analyze de-identified patient data from the University of Missouri Health Care and Ascension Seton in partnership with Dell Medical School at The University of Texas at Austin.
After the pilot and initial study are complete, Cerner expects to make its health data sets available to clients. When they access the Health Network, clients will be able to use Cerner’s HealtheDataLab to aggregate Cerner and non-Cerner EHRs. HealtheDataLab, in turn, draws on the capabilities of Cerner’s HealtheIntent, its data analytics tool for population health management.
So, have either Cerner or Epic made a breakthrough here? Probably not. While this all sounds good, the reality is that getting providers to share data freely is a tough problem, particularly given the need to comply with HIPAA privacy and security regulations.
To be fair, there are other approaches to health data sharing which could help close some gaps. For example, one possible model under discussion in the academic research world is using blockchain technology to lock down health data sharing between institutions. For example, earlier this year I reported on a study looking at how one blockchain-based model could not only protect clinical trial data shared between institutions but also make that data traceable and immutable.
If it works as promised, a blockchain-based data-sharing model might offer a basis for more extensive data exchanges. Unfortunately, though, compliance and security challenges are just the beginning. There are also serious ethical and financial questions that still overshadow the potential benefits of such record sharing, including but not limited to the extent to which patients should have the right to approve each transaction. A lot is still in flux.
Ultimately, it seems unlikely that sharing de-identified patient data will become popular until everyone potentially impacted by such sharing gets a vote on how it happens. Getting there is going to involve one heck of a journey.