Phil Ballentine, MSc's Linkedin Analytics

I'm a clinical data engineer and have been working with EHR and medical/pharmacy claims data since 2014. After working for 3 years testing and performing build and implementation of Epic at major hospitals, I've worked in Agile data teams ingesting, validating, enriching, and integrating claims and clinical data from many EHRs and different clinical systems and engineering and utilizing huge datasets to drive research and operational improvements, with a focus on NLP extraction of clinical entities from semi- and unstructured clinical text and notes. I am especially interested in using FHIR alongside HL-7 2.x to speed up EHR integration pipelines for predictive modeling, quality improvement, and generating Real World Evidence for scientific research. I graduated with a MSc in Health Informatics and Analytics from the Tufts University School of Medicine in August 2022. My MSc capstone project, Implementing a Flexible, Modern, and Repeatable FHIR-based EHR Integration Framework, included design and build of a Spark pipeline in Databricks to interface with a cloud FHIR server that I created, used clinical data from 1 million de-identified hospitalizations to train an inpatient mortality prediction ML model, and returned results as a FHIR Observation resource. The model achieved a >.87 AUROC in predicting inpatient mortality from an extremely minimal set of data that would be easily available from a USCDI-compliant bulk FHIR interface.