Join today’s leading executives online at the Data Summit on March 9th. Register here.

In September 2020, Microsoft launched Cloud for Healthcare, a managed service offering that’s designed to help health care organizations manage their operations. In an expansion of the original vision, the company today announced Azure Health Data Services, which aims to bring in data from clinical, imaging, and medical technology APIs so that it can be viewed together in data visualizations.

The launch comes as many health providers struggle to manage — and make use of — data across their organizations. According to a 2019 survey from the Society of Actuaries, only 60% of health care executives are using predictive analytics within their practices, despite the technology’s potential to cut costs and improve patient satisfaction. Another recent poll found that most health organizations don’t trust their own data, with 80% telling health tech company InterSystems that they lack trustworthy and interoperable data — even though it’s fundamental to making clinical decisions.

“We built Azure Health Data Services to address the industry’s distinctive needs to deliver better insights and care,” Cartwright told VentureBeat via email. “Health care organizations generate a tremendous amount of data every second. In addition, it is unique from other industries — in its challenges, in the way its data needs to be handled, and in its objectives. Our industry continually seeks ‘interoperability,’ but in reality, health organizations are using disconnected data points to make ad hoc solutions work while larger data sets continue to sit in silos. The good news is that silos of data are a pain point that cloud technology can solve, and as the health industry begins to expand in the cloud, how you bring data into the cloud makes a big difference.

Expanding visualization

The goal of Azure Health Data Services is to help organizations discover insights by bringing disparate data together and connecting it with tools for machine learning, analytics, and AI, according to Microsoft VP of health and life sciences Heather Cartwright. Using the new product as a part of Cloud for Healthcare, organizations can manage, anonymize, transform, and view health data across different parameters.

“One of the biggest challenges facing the health industry is the sheer volume of data it produces, which is too often unstructured and inaccessible. For most providers, 50% to 90% of their data is siloed, limiting advances in medical treatments and breakthroughs while maintaining regulation compliance,” Cartwright added. “This not only wastes valuable time on data processing, but also means that the data is unusable for analysis and AI and machine learning at scale. … Azure Health Data Services is purpose-built for personal health information — taking into consideration how data needs to be brought together from disparate sources, while still maintaining compliance boundaries and the privacy of the patients.”

Cartwright gave examples of how customers might use Azure Health Data Services at the point of care. A physician could query an MRI and show a patient’s medical history next to it, or analyze data collected from a wearable device, like a smartwatch, to see if a patient is trending toward an unhealth lifestyle. Meanwhile, clinical review teams could leverage Azure Health Data Services to study how many patients over a specific age had a particular symptom while using a given drug.

“The possibilities for AI and machine learning are what make Azure Health Data Services so exciting — the reality of having organized data in open-source formats along with tools that can quickly anonymize that data and have it ready for analytics, AI, and machine learning – that changes the game for data science. We know firsthand from our customers and partners that it is very complicated to capture insights and ready data for AI and machine learning without a foundation of data standardization and secure pipelines built for personal health information in the cloud. Azure Health Data Services creates that strong cloud foundation for big data, which makes true, deep AI and machine learning possible.”

Enduring challenges

Azure Health Data Services — and, by extension, Cloud for Healthcare — is in many ways Microsoft’s answer to the Google Cloud Healthcare API, albeit more holistic. But it might also be perceived as a response to the increasing demand for digital health care technologies in light of the pandemic. In a 2020 Optum survey, more than half of executives at hospitals, life sciences companies, health plans, and employer organizations said that they were planning to accelerate or expand their AI and analytics deployment timelines.

“Previously, health care data was separated in disconnected systems. With Azure Health Data Services, we are able to integrate data from different systems and combine it seamlessly in one place. Its ability to provide information using a centralized, secure location can accelerate the pace of research, helping scientists more rapidly analyze information and thereby translate discoveries to new treatments for patients,” Cleveland Clinic chief information officer Matthew Kull told VentureBeat via email.

Many blockers to successful — and responsible — deployment remain. The Optum survey found that the majority of executives were concerned about a lack of transparency in how analytics technologies use data and how the technologies arrive at their predictions. The concerns aren’t unfounded. Biases that wind up in analytics tools can lead to inaccurate clinical decisions, missed diagnoses and worsened clinical outcomes, particularly for already-disadvantaged groups of people.

Cartwright maintains that Azure Health Data Services has been engineered to minimize the potential for misuse, for example with documentation that explains the limits of its capabilities. Responding to a question about security — in 2021, more than 40 million patient records in the U.S. alone were compromised in various breaches — Cartwright says that Azure Health Data Services takes precautions to ensure that sensitive health data remains protected in and on a range of computing environments and devices.

“Azure Health Data Services is a platform as a service to ingest, manage, and persist data. Data access is controlled by our customers– they define user identities and access, and we make sure data is accessible only through APIs as our customers have defined it,” Cartwright continued. “Particularly as our customers are developing more AI and analytics, controlling access to data through APIs offers a new level of refinement as they use data for research.”

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn More