1

New horizons for health and life sciences data

If there’s one positive to come from the COVID-19 pandemic, it is the acceleration of digital technologies and working practices throughout every industry sector across the globe, and none more so than the healthcare sector. Digital transformation plans have had to be put into place within days or weeks instead of years.



But having the hardware and software in place is only part of the story. The most important component of any digital system is the data that resides within it, and the purposes to which that data is put.



“Meeting the challenges of a global pandemic means that every healthcare stakeholder — from patients and providers to insurers and pharmaceuticals — must share data from disparate geographies, scenarios, and populations in the effort to help understand, treat, and eventually eradicate COVID-19,” said Dr Mark Lambrecht, Director of the Global Health and Life Sciences Practice at SAS.



“No doubt, the processes built today — as well as the key lessons learned and preparations made — will define the future of healthcare.”



Some of those lessons and preparations will emerge from a new strategic partnership formed between SAS and Microsoft, which aims to accelerate healthcare innovation through artificial intelligence and computing.



Transforming our understanding



Although massive efforts are underway to connect healthcare data that comes in every form, standard, and quality imaginable, “Gaining insights from the intersection of patient observations and clinical trials, for instance, can feel Sisyphean — yet that ability will likely define the future of healthcare,” said Lambrecht.



One standout example of success is the Healthy Nevada Project, which is already demonstrating the value and possibilities of connecting patient data. This community-based, genetics study uses SAS machine learning and artificial intelligence to improve population health in Nevada.



“By gathering information from citizens who enrol in the program, geneticists can identify predispositions for certain diseases, or alert asthma patients when they travel to a part of the state with poor air quality,” said Lambrecht.



“To amplify these applications of data analytics and AI on a global scale requires massive compute power and a secure infrastructure to share and control patient data,” he added.



“That’s what excites me about our new partnership with Microsoft and the combined power of SAS analytics running on Azure.”



This idea is supported by Heather Cartwright, who leads a team on new cloud and AI technologies for health data at Microsoft. She says that unsustainable workloads in the healthcare space have acted as an impetus for adoption of the cloud.



“There is an overwhelming increase in the types of data care teams need to manage. As the number of inputs clinicians use to treat patients grows, we need to leverage different tools for health data,” said Cartwright.



“Cloud technology provides the scale which is urgently needed to manage health data workloads, but just as important, it enables machine learning with that data,” she added.



“Health leaders understand how that will transform our understanding of human health and how we deliver care in the future. So healthcare is finally saying, ‘Okay we need to go to the cloud, and we need to know how.’”



But sometimes it’s not easy, says Lambrecht. “As the leader of SAS’ scientific response to COVID-19, I can testify to the difficulty of bringing observational patient data derived from healthcare claims, healthcare registries, clinics, and all types of patient interactions together for analysis,” he said.



“To lead the way forward, healthcare organisations need a comprehensive enterprise cloud strategy and an analytics strategy that drives insight from real-world data.”



Lambrecht says that Microsoft and SAS are “committed to meeting healthcare organisations where they are,” with cloud-based solutions that are ready to run on day one, but can also scale as organisations grow.



A good example of this is Mercy, a leader in both technology and clinical care. Among the first organisations running SAS analytics natively on Azure, Mercy boasts a virtual health division and an analytics culture that helps it bring information together about COVID-19 patients and rapidly package that data to make it available to other health organisations working on innovative therapies.



The safe and secure cloud



Although the healthcare sector traditionally takes a conservative approach to innovation, it needs to be able to scale and drive insights from different types of data sources. SAS AI and analytics provides that scalability.



As Cartwright puts it, “When you’re innovating, trust is essential. We want to make sure that health systems maintain control over their data when they move it to the cloud, that they can define database access and bring their own identity.”



“We make sure these security measures are in place so our customers can trust that their data is in the right foundation, because that frees them to really focus on innovation.”



Flexibility is important too, which is why Azure Synapse provides the ability to work in whichever environment healthcare professionals are already comfortable. “Scientists shouldn’t have to learn a new language in order to work with a different data set,” said Cartwright.



Critical to that flexibility are the feedback loops and machine learning that enables dynamic decision-making at every level of healthcare.



“It is so important to bring the front lines of healthcare into that machine learning process,” said Cartwright. “Feedback loops are essential to make models better… refining, expanding even, or identifying new algorithms we need to develop.”



“SAS and Microsoft are building solutions that physicians can trust,” added Lambrecht. “We’re rapidly creating simpler interfaces that do not hide the analytical complexity or the data complexity, but still allow decision makers to make the right decision, to extract insights that correctly steer how they need to run their organisation.”



For Cartwright, transparency in AI development is key.



“People using data models should be able to go deeper and understand what is happening in those models, what the inputs are for those models and the parameters, so that they can have trust in it,” she said. “And then we can continue to validate and make sure that they are working at the right levels.”



In other words, acceleration shouldn’t come at the cost of proven, hierarchical data validation processes.



Oncology is a good example. Without substituting the expertise of the physicians, SAS’ deep learning algorithms and models helped Amsterdam UMC automate the read-out of metastatic liver lesions due to chemotherapy treatment by rapid calculation of various metrics like volume or surface of the lesions.



The algorithms didn’t hide the complexity of the analytics, but they did provide enormous support for oncologists who would otherwise spend a lot of time on error-prone tasks.



Which just goes to show that if healthcare organisations have their data in the cloud, with analytics engines running and data science teams working closely together, they will have a ‘readiness machine’ to make decisions in a crisis, says Lambrecht.



“I’m thrilled to work with Heather as SAS and Microsoft build that secure and powerful readiness machine together,” he said.



For more information, click here.



Image credit: ©stock.adobe.com/au/Shutter2U

Подробнее
[fixed][/fixed]