Using Machine Learning To Solve Critical Challenges In Healthcare It
Artificial Intelligence and Machine Learning are rapidly gaining momentum in Healthcare and - as some say - on the verge of becoming the most important breakthrough for healthcare since penicillin.
Most of these technological advances rightfully target improvements of patient care - for example in the areas of better diagnostic services, precision medicine, personalized drug development and decision support.
Healthcare IT is undergoing similar revolutionary transformations with AI and Machine Learning techniques that are helping to address critical challenges. Two of the following our recent use cases highlight these challenges and reflect on successful applications of AI and ML in our practices.
COMPLEX PROCESSES FOR PATIENT IDENTIFICATION
Reliable patient matching is one of the most complex and most important elements in communications between Health systems. Matching records to the correct individual is more complicated when patients receive care in multiple settings and when organizations and providers use different systems to share records electronically.
In the process of creating rare diseases registry one of the organizations we have been working with has a need to identify and capture medical records matching target disease criteria. On behalf of this client we have integrated with multiple healthcare institutions and are receiving all visit summaries from these institutions. Only small number of the visiting patients match the eligibility criteria and expected to be captured in the registry. Additionally, records coming from these external systems have to be correctly matched to patients which are already in the registry.
We have utilized and integrated existing open-source machine learning components with proprietary algorithms developed and optimized in-house.
We subsequently trained these components on large sets of synthetic and real-life patient records to achieve desired matching accuracy and included these matching processes in the existing data ingestion pipeline.
The above infrastructure is built as part of the integration pipeline based on Intersystems HealthConnect integration engine. The machine learning components are implemented using Python libraries and a special adaptor was built into HealthConnect to support patient matching workflow. The infrastructure is deployed on Microsoft Azure under Docker environment.