Back to Resources

A fast, resource efficient and reliable rule-based system for COVID-19 symptom identification

John Sartori, PhD

John Sartori, PhD

Co-founder, CTO

Dr. Sartori is an Associate Professor in Electrical and Computer Engineering whose research expertise spans multiple fields.

This blog article summarizes a study co-authored by Q-rounds cofounder John Sartori, Ph.D., and a team of researchers.

Existing annotation systems faced significant challenges in scalability and resource utilization when it came to promptly identifying COVID-19 symptoms in clinical decision support systems (CDS). 

The runtime of a rule-based gazetteer designed at the University of Minnesota was compared with five annotation systems—BioMedICUS, cTAKES, MetaMap, CLAMP and MedTagger—to determine document processing times, resource needs and performance in terms of weighted microaverage and macroaverage measures for precision, recall and f1-score.

Identifying limits in CDS scalability

With COVID-19 came an unprecedented need to identify symptoms of COVID-19 patients under investigation (PUIs) in a time-sensitive, resource-efficient and accurate manner, but traditional annotation systems struggled to keep up. For example, the processing time for 12,000 clinical notes took MetaMap taking 105 hours, CLAMP 28 hours and cTAKES nine hours, severely limiting the scalability of CDS.

Integrating with CDS

A rule-based gazetteer—a dictionary of terms derived from a given lexicon—specifically designed for COVID-19 symptom identification was designed to mitigate CDS integration issues and bolster CDS capabilities in real-time.

Rounding it up

The rule-based gazetteer facilitated real-time symptamology identification for COVID-19 and integration of unstructured data elements into the CDS, with lower resource utilization and faster processing times than traditional annotation systems. It also reduced processor and memory utilization, making it an ideal fit for medical sites lacking robust healthcare infrastructure. Additionally it matched traditional annotation systems in terms of weighted micoaverage and macroaverage measures for precision, recall and f1-score. 

Furthermore, it’s ideal for large-scale deployment across a wide variety of healthcare settings for surveillance of acute COVID-19 symptoms for integration into prognostic modeling. Such a system is currently being leveraged for monitoring of postacute sequelae of COVID-19 progression in COVID-19 survivors.

Q-rounds uses AI to learn provider preferences and suggest an optimal rounding order, in addition to sending automated, real-time notifications letting care teams, patients and their families know when to be present for rounds.

Interested in learning more? Read the full study or see how Q-rounds can help.