Reflection - Part 2: EHR

From Xiruo:

I like the article of Death by a Thousand Clicks on Fortune. It unveils to us how fragile, unstable, counter-intuitive, and uncommunicable the EHR systems were and probably still are. And it is quite interesting that Epic’s President is still promoting those big fancy terminologies as part of the future of EHR, and I would say some of them are even headaches in academia. The AMIA EHR-2020 Task Force was taking a somewhat different but still similar way of envisioning EHR’s future, and some of their suggestions are to the point brought up by the Fortune article. It is just a little interesting when we look at their fifth goal as “The EHR in 2020 must support person-centered care delivery” five years later and it seems still in an unsatisfactory stage, quoted from MedStar’s Raj Ratwani and Seema Verma, the administrator of the Centers for Medicare and Medicaid Services.

From Yue:

Novel approaches that complement and go beyond evidence-based medicine are required to guide clinical decisions. Although Electronic Health Records (EHRs) are common and lots of standards have been developed to improve the quality of structured data, a wealth of clinical information remains locked behind clinical narratives in free-form text. A promising avenue to unlock the full potential of clinical notes is applying NLP methods to understand patient clinical trajectories and improve the decision supported by the analysis of structured data. Besides predicting the condition of the patient in ICU stay, the adoption of NLP for clinical notes provides a unique opportunity for chronic disease management, clinical decision support, and potentially delay or prevent disease.

From Jake:

From an NLP perspective the EHR provides a rich source of data. With traditional paper notes there would be no ability to do clinical NLP on data. Our three papers show how NLP and information extraction techniques leverage EHR data to provide key insights. This goes towards one of our main benefits of EHR adoption is the promotion of research. While we learned that EHR adoption isn't without it's drawbacks (cost, timeliness), we also see the NLP that comes from it to improve cost and quality of care.