The healthcare industry generates about 30% of the world’s data. However, historically this data has been locked in silos and healthcare organizations have failed to generate real-time operational and clinical insights at scale.
A new generation of unified data ecosystems, intelligent automation and modern medical technology stacks are changing that, delivering real-time insights that improve decision-making.
To learn more, Becker’s Healthcare recently spoke with Kushboo Goel, head of healthcare strategy and consulting and Pragadeesh J, director of data engineering at Neurealm.
‘Not an acceptable model’
The typical health system uses 10+ different system in addition to the gravitational center the EHR leading to siloed data that can include patients struggling to access their own data, clinicians ordering duplicate tests and limiting interoperability with federal and state systems.
To solve these challenges, organizations tried to pull data from disparate systems into a central data lake for reporting and analytics. But the data wasn’t standardized, definitions varied and access was limited to a small team of experts so end users had to request reports. By the time reports were delivered, the information was often already outdated.
Ms. Goel calls this “not acceptable,” as these existing structures mean organizations lack a real-time unified view of each patient and don’t have real-time clinical and operational insights to make decisions.
Having isolated data from one system or data from even a few hours ago is no longer good enough. Data must be unified and in real time. “The concept of near-real time is gone,” added Mr. Pragadeesh J. “Real-time datasets are needed to make decisions much faster.”
What health systems need
The goal of optimizing data availability is for end users to make better decisions through real-time insights, derived from a unified view of all data sources.
Mr. Pragadeesh J and Ms. Kushboo Goel refer to this as a unified data ecosystem, where clinicians have access to a unified view of their patients, including data from their EHR, lab results, monitoring systems and even a patient’s social determinants of health. AI algorithms would then use this data to identify each patient’s health status and risks, determine a recently discharged patient’s risk of readmission and help clinicians and patients with decision-making.
Operationally, the concept of a “hospital command center” has taken hold. Command center leaders want enterprise-wide visibility for capacity planning and resource-allocation decisions. This includes predictions about ED volume to guide staffing decisions and other factors.
In addition, Ms. Kushboo Goel noted that organizations want their data to be AI-ready, so they can build AI use cases that seamlessly integrate into clinical and operational workflows. Being AI-ready also means having the right engineering foundations, such as continuously processing data and reusing key clinical features, ensuring models always operate on accurate, up-to-date information at the point of care.
Automation and the modern data stack
Automation has advanced rapidly over the past decade, from basic rule-based tools to high-volume robotic process automation and now intelligent automation that turns ambient clinical conversations into accurate documentation for the revenue cycle, improving efficiency and potentially boosting revenue.
Utilizing intelligent automation can also have an immediate impact on patient care with some organizations using it to review patient data and identify hospital-acquired infection (HAI) early to positively shape outcomes early.
A modern data stack enabled by automation creates the foundation for predictive analytics and scalable AI. It streamlines how data is collected and cleaned, giving teams a single, reliable view of information much faster.
The components of a traditional data stack were data warehouses, datasets, ETL tools and business intelligence. In the era of intelligent automation, the ecosystem is changing to a data intelligence platform that does everything. It handles data ingestion, has open storage, is cloud-based and scalable, works with both structured and unstructured data and uses generative AI to quickly answer questions and create custom reports.
Organizations optimizing to be more “intelligent” enable strategic agility to pivot and quickly adapt to an ever-changing modern healthcare landscape.
