Healthcare organizations use AI-driven tools as part of their clinical workflows to support decision-making and automation of administrative tasks, Shegewi says. In fact, a little more than half (50.8%) of U.S. healthcare providers plan to increase generative AI spending and also demand robust data integration and analytics capabilities to aid their next-generation use cases, according to IDC’s “MarketScape: U.S. Healthcare Data Platform for Providers 2024–2025 Vendor Assessment.”
“By offering real-time data processing, secure integration and scalable infrastructure, modern data platforms lay the groundwork for advanced use cases, allowing organizations to train AI models on comprehensive and high-quality data sets,” Shegewi says.
Tina Esposito, senior vice president and chief data officer for Advocate Health, which is headquartered in Charlotte, N.C., and spans six states, says the cloud allows organizations to modernize their data stack and democratize information so they can make better decisions.
“The amount that the data is growing is enormous, and so as you think about how you want to manage that as an asset, you need to look at cloud, because that is where it will be much more scalable, available and manageable,” Esposito says.
READ MORE: Follow these AI data governance strategies for success.
Why Modern Data Platforms Are Important for AI in Healthcare
Modern data platforms process massive amounts of data, including data from patient records, lab results and wearable devices, explains Sha Edathumparampil, chief digital and information officer for Baptist Health South Florida.
The health system uses cloud-based data platforms and integrates a centralized data lakehouse with AI platforms and models from Microsoft, Amazon Web Services and Google.
“AI and analytics make these systems powerful, supporting doctors and nurses in making clinical decisions, predicting patient risks and preparing hospital operations for busy periods,” Edathumparampil says. “Beyond predictions, they also automate scheduling and paperwork, freeing up time for patient care. They can even analyze human-supervised medical scans and track health trends across populations.”
Integrating both internal and external data is also a central function of modern data platforms for Esposito at Advocate Health.
“It’s very rare that a strategic question can be answered simply from an EHR, for example, or a supply chain system,” Esposito says. “It really is the integration that enables that, so pulling it together is necessary.”
Esposito says Advocate Health has worked with machine learning and predictive modeling and is now moving into generative AI. The health system uses AI to gain operational efficiencies, such as ensuring that they have the right staff at the right time as well as predicting the length of a patient’s stay, she says.
With patients’ lives at risk, data must be cleaned up in modern data platforms before being added to healthcare databases such as EHRs. Diverse data sources must be consolidated, and duplication and inconsistencies eliminated, says Shegewi.
“This is critical for conducting the most elective form of precision medicine and patient-centered care, where duplicative or inconsistent data can jeopardize patient safety and trust,” Shegewi says.
