15x increase in urgent care bookings with a Virtual Care Platform on AWS

01

Business needs

A Fortune 500 US-based healthcare services innovation company wanted to improve the performance of its Virtual Care Platform (VCP) to meet the evolving demands of patients and enable healthcare providers to deliver round-the-clock care.

The following were their primary business requirements:

• Enhance the VCP with an on-demand scheduling feature that allows users/members/patients to view wait times and book appointments promptly
• Implement a highly reliable and optimized system to efficiently retrieve the earliest available time slots from a diverse pool of healthcare providers
• Establish a sophisticated monitoring and alert mechanism that promptly notifies users when wait times exceed the three-hour threshold
• Address existing issues, such as slow response time, suboptimal user experience, limited customizations, and insufficient sorting and grouping capabilities, which are negatively impacting appointment bookings

03

Solution

With deep domain expertise, the Impetus team architected a comprehensive solution to help the client address the technical needs of the Virtual Care Platform (VCP), keeping up with unprecedented demand. Here are a few solution highlights that contributed to the project’s success:

• Implemented a Pub/Sub-based event system to refresh the availability of whitelist providers and visit reasons for multiple days or selected day/provider/visit
• Synchronized baseline schedules and appointment data into the ‘slots availability’ storage system for seamless scheduling, rescheduling, and cancellation
• Synchronized appointment slots data for multiple providers, locations, and visits through parallel processing using Python Consumers and AWS Simple Queue Service (SQS)
• Utilized schedulers to verify and retry in case of stale or missing information about provider availability in the ‘slots availability’ storage
• Enabled filtering and sorting based on speciality, visit reason, distance, insurance, availability, and the provider’s profile status by implementing our proprietary algorithms in Java and Elasticsearch (ES) query builder
• Automatically created indexes when new organizations onboard the system using the Change Data Capture (CDC) technology stack
• Implemented auto index updates when changes occur in provider details, insurance, reasons, etc., with CDC
• Rearchitected and redesigned the legacy search implementation using Elasticsearch capabilities.
• Designed a highly optimized Elasticsearch (ES) query builder for efficient provider search based on input criteria.
• Built logic with Java, data structure, and Elasticsearch query builder to find alternate providers for selected visit reasons, improving the appointment booking process
• Implemented separate microservice to capture real-time updates for provider profile details and availability
• Developed a multi-threaded system to manage updates in the provider index efficiently
• Leveraged Kafka for publishing and consuming events to keep the updated data in the provider index
• Implemented robust alerting and monitoring using New Relic and AWS CloudWatch to track high response/error rates, throughput, and CPU usage
• Established a wait time alert mechanism to notify stakeholders if there is insufficient coverage in a particular location by implementing custom scheduler jobs in Java (Spring Boot, MySQL)
• Ensured zero downtime during the refreshing of availability slots data at the organization and provider levels
• Ensured availability of Horizontal Pod Autoscaler (HPA) based on CPU usage thresholds to ensure system scalability and effectively manage traffic spikes in peak utilization hours