Why legacy tech still matters in the AI era

When healthcare policymakers and technology companies talk about the future, the conversation usually revolves around interoperability, AI, and real-time data exchange. But Megan Zakrewsky, Vice President of Product for clinical data exchange at Veradigm, argues that focusing too heavily on the cutting edge risks leaving behind the millions of patients who rely on independent and rural providers still running legacy systems.

“These providers can’t afford costly and disruptive rip-and-replace transitions,” Zakrewsky told Digital Health Insights. “Their systems may lack FHIR-forward exchange strategies, but they’re reliable, deeply embedded in workflows, and rich with longitudinal patient data.”

Her message: innovation must extend beyond hospitals with deep IT budgets. If federal policy and vendor strategy don’t account for smaller players, the digital divide in healthcare will only widen.

Batch data as a population tool

Much of today’s excitement in health IT centers on real-time data — the ability to surface alerts or insights while the patient is still in the exam room. Zakrewsky acknowledges that value, but emphasizes that batch data remains essential for scaling population health.

“Real-time data is powerful for patients already in the office,” she said. “But batch data gives providers and payers the full picture of their population, not just the patients in front of them.” That means surfacing which patients are overdue for preventive screenings or scheduling annual visits before gaps become larger problems.

This distinction resonates with ongoing national efforts. CMS has tied much of its value-based reimbursement policy to preventive care and population health management. Programs like the Medicare Shared Savings Program (MSSP) have shown that systematic outreach can reduce costs while improving quality. In 2024, MSSP saved Medicare $2.4 billion while meeting quality benchmarks, according to the agency. Those gains depend less on real-time alerts and more on systematic, data-driven outreach.

A shared roadmap for payers and providers

Batch data, Zakrewsky argues, also fosters collaboration between payers and providers. By aggregating insights across a panel of patients, payers can highlight high-priority care gaps or quality opportunities, and providers can use that intelligence to plan visits.

This kind of two-way exchange is critical. Historically, the payer-provider relationship has been adversarial, often reduced to disputes over claims or authorizations. But federal initiatives, from the 21st Century Cures Act to the Trusted Exchange Framework and Common Agreement (TEFCA), have emphasized the need for shared information.

Zakrewsky’s perspective aligns with this policy direction: “Batch-based exchanges can capture documentation at the point of care and return it to payers in structured form,” she said. That helps eliminate redundant data entry and administrative burden — a perennial concern for physicians.

Closing the digital divide

Yet the challenge remains: how to ensure that independent practices with aging EHRs can participate in this exchange. Federal health IT policy has often assumed a baseline of modern, interoperable systems. For example, the ONC’s recent information blocking rules and CMS’s interoperability mandates emphasize FHIR APIs. But small practices without the resources to upgrade may find themselves excluded.

“Policy must help close the digital divide, not widen it,” Zakrewsky warned. She points to misalignments across federal entities on data requirements and advocates for standards that bring payer intelligence into legacy systems with minimal disruption.

This echoes broader criticism from rural health advocates. A 2023 Government Accountability Office report noted that rural hospitals face disproportionate financial and workforce pressures, making large-scale IT overhauls even less feasible.

The role of AI: not replacement, but augmentation

Despite the hype, Zakrewsky does not expect AI to sweep away legacy technology anytime soon. Instead, she sees its promise in unlocking the value of existing systems.

“The most successful AI solutions will be the ones that wrap around legacy technology,” she said. That means applying natural language processing to extract meaning from unstructured notes, or embedding predictive insights into existing EHR workflows.

Her stance mirrors recent federal guidance. The White House’s AI in Healthcare report emphasized that adoption will require AI to integrate into existing systems, rather than demand wholesale replacement. Industry analysts have echoed the point: a 2024 KLAS Research survey found that CIOs prioritize AI tools that augment clinician workflow over those that require new platforms.

People at the center

For Zakrewsky, the ultimate measure of progress is not the sophistication of the technology but the inclusiveness of its impact.

“As an industry, we talk a lot about technology’s impact, but transforming healthcare is bigger than technology alone. It’s about people,” she said. “Independent and rural providers care for millions of patients, and they deserve to be part of this journey from the start, not as an afterthought”.