See Part 1 of my HIMSS14 Session Review: Overcoming The Current and Future Challenges ACOs Will Face

In this HIMSS14 session, University of Pittsburgh Medical Center presented its challenges in integrating 49 clinical source systems across its health system. The driver behind this integration was based on the industry redesign of how healthcare will be delivered. Included among these drivers are the:

  1. Switch from a facility centered record to an episodic record
  2. Requirement for strong analytics to foster accountability
  3. Necessity for payers and providers to be more tightly integrated
  4. Need to de-noise data so clinicians can focus.

Rasu Shrestha, MD, VP of Medical Information Technology at UPMC, presented the following graphic to demonstrate how interoperability fits into the overall picture of care.

in the pyramid, which leads to patient-centered accountable care, semantic interoperability is the glue between raw data and the applications which need to use the data and present it to clinicians in a meaningful way. the pyramid emphasizes the additional challenges that present itself as data begins to flow from facility to facility. 

first there is the identification layer, which includes not only the patient but also provider and facility identification as well. then comes the semantic layer, which requires applications across facilities to use consistent structure coding to represent the clinical events. dr. shrestha talked in detail about the painstaking efforts that took place to map the coding systems across all facilities. upmc had to contend with 18 different coding systems just for problems. once that semantic mapping takes place, the data can be stored into a central data repository and made available to the applications for data mining, decision support, and an episodic view of clinical events for a single patient.

as described by dr. shrestha, the end goals for achieving this semantic interoperability were to seamlessly exchange and integrate information in a meaningful way that the data can be contextually understood, accurately interpreted, and be actionable.

sounds like perfect job for an interface engine to me!

Tags: ,
 Print Friendly