Improved Patient Care Enabled by Real-Time Interoperable Clinical Decision Support
Intermountain Healthcare, a HIMSS Triple Stage 7 Organization, has developed and implemented a real-time interoperability platform to deliver Clinical Decision Support applications. It is expected to change the model of care, allow for better clinical decisions, and reduce costs while saving lives. An OMG BPM+ environment is built into the FHIR-based interoperability platform allowing clinicians and developers to work together to define and modify care models; we anticipate an enhanced ability to deploy and maintain healthcare applications. Realizing that data latency issues in traditional FHIR servers rendered them unsuitable for the routine delivery of clinical decisions, the platform uses a Publication/Subscription messaging service to trigger data collection and decision support behaviors; reads from and writes to the Cerner EHR through FHIR wrappers; and accesses machine learning-derived tools to compute disease likelihoods. It consumes a variety of clinical data including radiographic interpretations generated by a convolutional neural network developed at Stanford University, reducing the wait for image interpretation to 3 seconds. The use of FHIR for data access makes the model and clinician experience portable regardless of the EHR; the use of BPM+ accelerates the delivery of novel decision support systems. The presenters designed and built the interoperability platform.
Learning Objectives
- Identify key technologies needed to create applications based on the message driven microservices paradigm
- Discuss how Business Process Modeling for Healthcare (BPM+) based software management strategies can provide benefit in Clinical Decision Support environments
- Explain how to integrate different machine learning modalities into evidence-based applications that standardize and rationalize care processes for a specific illness
- Describe strategies to design and build an EHR integrated FHIR application based on the FHIR specifications regardless of whether your EHR supports all the FHIR resources
Speakers


