Ubie: Using GKE Autopilot to lower costs and focus resources on delivering an AI-powered patient questionnaire and clinical record system for hospitals
About Ubie
Ubie is a Japan-based healthcare technology startup that automatically generates medical records using an AI-powered patient questionnaire app, saving doctors time and enabling them to provide better patient care. In 2020, the business expanded its offering to healthcare providers in Singapore.
Tell us your challenge. We're here to help.
Contact usWith GKE Autopilot, Ubie is reducing infrastructure costs and maintenance requirements, enabling development teams to focus on achieving the business’s vision of making healthcare accessible to everyone.
Google Cloud results
- Reduces infrastructure costs by 20% to enable the business to invest in areas such as product development and delivery
- Scales application seamlessly to deploy to hundreds of medical institutions across Japan, supporting growth and executing business vision
- Delivers improved availability, meeting internal requirements and providing a high-quality experience to customers and users
Eliminates infrastructure and maintenance tasks requiring several days to complete, freeing resources to develop products
The coronavirus pandemic reinforces the importance of efficient delivery of medical and healthcare services to the community. This is where businesses like Tokyo-headquartered medical technology startup Ubie come in. Founded in 2017 by software engineer Kota Kubo and medical doctor Yoshinori Abe, Ubie aims to direct people to appropriate medical care as and when they need it, and does this with products designed for hospitals and individuals. Flagship product Ubie for Hospital is AI-powered questionnaire software that enables patients to provide details of their symptoms, condition, or other relevant information to specialists, who review the responses prior to an appointment or session.
These responses are captured in a clinical record for review, while the system integrates with and complements existing practice electronic medical record and clinical management systems with capabilities such as pre-clerking, auto-generated case notes, and clinical support. Developed by engineers with input from doctors and launched in 2017, the system helps hundreds of hospitals and medical institutions across Japan streamline patient care.
“Doctors and healthcare workers can be bogged down with administration rather than patient-facing work,” explains Jun Sakata, Software Engineer, Site Reliability at Ubie. “Rather than laboriously entering into a computer patient responses to a paper-based questionnaire, doctors and healthcare workers can now directly access data entered via the AI-powered questionnaire system.” Ubie claims automated record-keeping can save a practice or institution 1,000 hours a year.
The business also offers a smartphone app for consumers that allows users to identify and manage their symptoms by answering AI-generated questions.
“We had to undertake node upgrade maintenance once a month ,and it took our team several hours to complete each time. Furthermore, sometimes we needed to change node instance types to optimize resources for application updates, and this could take a few days. With Google Kubernetes Engine in Autopilot mode, we no longer need to devote time, resources, and cost to these activities.”
—Jun Sakata, Software Engineer, Site Reliability, UbieInitial reliance on different cloud
Ubie relied initially on an alternative cloud to make Ubie for Hospital available to medical practices and hospitals in Japan. However, as the business added new customers and increased the frequency of updates to increase the accuracy and functionality of the Ubie for Hospital product—for example, giving doctors the names of diseases associated with the symptoms displayed—it needed an infrastructure that could support daily deployments complemented by a secure gateway that connected Ubie to a wide range of customer networks and settings.
The business’s evaluation found Kubernetes commands and resources would enable it to deploy, administer, and monitor its system, while Google Kubernetes Engine (GKE) would provide load balancing for its virtual machine instances (nodes), node pools to improve flexibility, automatic scaling of node instance counts, automatic upgrading of cluster node software, node auto-repair, and logging and monitoring.
The business also turned to Cloud Interconnect to provide secure connections between the networks used by many hospitals and private medical practices and the Ubie Google Cloud environment.
Ubie moved to GKE initially running using the Standard mode of operation in April 2019 and began deploying Ubie for Hospital to customers about one month after general availability.
“With GKE Autopilot, we can do more with our business. We can continue developing and upgrading our products, rather than focusing on fine-tuning infrastructure.”
—Jun Sakata, Software Engineer, Site Reliability, UbieGKE Autopilot a compelling option
However, the availability of Google Kubernetes Engine Autopilot—a mode that enables full management of an entire cluster’s infrastructure and provides per-pod billing—presented a compelling option for the business to run Ubie for Hospital more efficiently and cost-effectively. With GKE Autopilot, Ubie could eliminate the need to configure and monitor clusters, while paying only for running pods (with billing per second for vCPU, memory, and disk resource requests)—not system components, operating system overheads, or unallocated capacity.
Ubie elected to move the GKE node pools running Ubie for Hospital to GKE Autopilot mode, retaining all default settings bar those needed to meet its specific business requirements.
The shift has reduced Ubie’s infrastructure costs by 20%, while GKE Autopilot has enabled the business to eliminate Ubie for Hospital infrastructure maintenance and upgrade tasks that could take hours and days respectively to complete. “We had to undertake node upgrade maintenance once a month, and it took our team several hours to complete each time,” says Jun. “Furthermore, sometimes we needed to change node instance types to optimize resources for application updates, and this could take a few days.
“With GKE in Autopilot mode, we no longer need to devote time, resources, and cost to these activities.”
“With our AI-based symptom checker and clinical support technology, which we have been developing since 2013, we will make healthcare accessible to everyone by creating a ‘search engine’ for healthcare. We are going to go full throttle to realize our company mission as soon as possible.”
—Yoshinori Abe, MD., CEO, Ubie Inc.Building on the rich benefits of GKE Standard
GKE Autopilot also improves on the already high availability levels enabled by GKE Standard, which enables the business to create multi-zone and regional clusters that distribute Kubernetes resources across multiple zones within a region.
With Ubie for Hospital running effectively for hundreds of customers on GKE Autopilot and GKE Standard powering its smartphone app and common medical engine, the business can now focus its resources on performance, features, and customer acquisition. “With GKE Autopilot, we can do more with our business,” says Jun. “We can continue developing and upgrading our products, rather than focusing on fine-tuning infrastructure.”
Yoshinori elaborated in a recent media interview, saying “I strongly believe that our company is going to play a crucial role in combating the current crisis. During these uncertain times, our mission, ‘to develop a healthcare guide for everyone,’ has never been more important.
“With our AI-based symptom checker and clinical support technology, which we have been developing since 2013, we will make healthcare accessible to everyone by creating a ‘search engine’ for healthcare. We are going to go full throttle to realize our company mission as soon as possible.”
Tell us your challenge. We're here to help.
Contact us
About Ubie
Ubie is a Japan-based healthcare technology startup that automatically generates medical records using an AI-powered patient questionnaire app, saving doctors time and enabling them to provide better patient care. In 2020, the business expanded its offering to healthcare providers in Singapore.