AiCure’s AWS-Powered Platform Supports Clinical Trials by Learning Patient Behavior

Clinical trials can struggle to remotely capture patient engagement and assess symptoms. Accessing meaningful patient data and obtaining insights into real-time engagement can provide a holistic view of the patient’s symptoms and experience to treatment.

AiCure reduces costs, improves the predictability of study timelines, and accelerates those timelines through remote patient engagement and assessments, such as measuring digital biomarkers and real-time monitoring of patient dosing. AiCure uses a microservices-driven architecture, leveraging a variety of solutions from Amazon Web Services (AWS) to collect, catalog, and process video, audio, image, and text data from patients participating in clinical trials. Services including Amazon API Gateway help connect patients’ smartphones and the backend architecture, while AWS Step Functions and AWS Batch are leveraged for continuously improving AI models and inferencing at scale to make data actionable. Other services, such as AWS Lambda, are used to generate metadata for each execution of the AI pipeline to create an audit trail and full traceability across the system. This data can then be used by the clinical trial sponsor to enable patient engagement and retention, and to support medication adherence.

Building on AWS enables AiCure to “leverage cloud services for scalability, flexibility, reliability, auditability, and security. There’s no need to reinvent the wheel,” says Lei Guan, chief technology officer at AiCure. Authentication services like Amazon Cognito, AWS Key Management Service (AWS KMS), and AWS Lambda are key to AiCure maintaining HIPAA compliance, while storage solutions like Amazon Simple Storage Service (Amazon S3) ensure that it doesn't have to worry about scalability as its user base and data collection volume increase. A microservices-driven architecture means that AiCure can run its platform at scale, focusing on constantly refining its core AI technology. And since moving to its AWS-based architecture, AiCure has collected more than 600,000 videos, 180 million image frames, 600,000 dosing events, and 17,000 intervention events from 30 countries, all at near-real-time performance.