Overcoming Data Fragmentation to Unlock AI’s Potential in Healthcare

In an era where technology promises to revolutionize healthcare, the potential of artificial intelligence (AI) is undeniable. From streamlining administrative tasks to enhancing patient care, AI could transform the industry. However, a significant hurdle stands in the way: data fragmentation. This issue affects both payers and provider health plans, making it crucial for organizations to address if they hope to fully leverage AI.

On a recent webinar sponsored by Verato, experts from SCAN Health Group and the Alliance of Community Health Plans discussed how their organizations are tackling this challenge. The conversation highlighted the importance of shared data definitions, robust data governance, and strategic planning in achieving AI readiness.

Thomasina Anane, Associate Vice President of Enterprise Analytics for the Alliance of Community Health Plans, emphasized the need for consistent data governance and shared data definitions. "Fragmentation shows up as documentation gaps, misdiagnoses, undercoded acuity," she explained. These issues can lead to inaccuracies in payment accuracy and risk adjustment, which are critical for the financial health of health plans.

Vinay Kulkarni, Chief Information Officer of SCAN Health Plan, stressed the importance of high-quality, interoperable data and strong data privacy practices. "AI readiness is an operational and structural reality," he said. "True AI readiness means your data workflows and compliance guardrails are built so that your machine learning models can rely on them."

Addressing Data Fragmentation: Practical Steps and Strategies

To address data fragmentation, organizations must take practical steps to ensure their data is reliable and usable. Anane identified payment accuracy and risk adjustment as key areas where fragmentation has a significant impact. "If you’re not capturing this information accurately, you’re not being paid accurately," she noted. This can affect a health plan's competitiveness and ability to thrive in the industry.

Kulkarni outlined what AI readiness means for a health plan. "You’ve got to have embedded human-in-the-loop checkpoints and mandatory manual reviews," he said. "These are structures that you have to put in place." He also emphasized the importance of deterministic data lineage, which allows organizations to trace AI-generated output back to its raw data inputs and applied business rules.

The webinar also highlighted the need for internal operational alignment within health plans. Anane discussed the challenges of fragmentation across different functions and the necessity of consistent data governance. "Shared data definitions and consistent governance are crucial," she said. "Without them, it's difficult to ensure that everyone is on the same page."

Why It Matters

Addressing data fragmentation is not just a technical challenge; it has real-world implications for patients and healthcare providers. Accurate and interoperable data can lead to better patient outcomes, more efficient care delivery, and reduced administrative burdens. As universities like Penn and UC San Diego explore how AI can improve healthcare outcomes, the importance of robust data governance becomes even more evident.

The University of Pennsylvania, funded by a $215 million five-year National Science Foundation award, is leading research into how AI can enhance healthcare. Similarly, UC San Diego is building an AI-enabled health system that spans basic discovery, clinical research, and patient care. These initiatives underscore the potential for AI to transform the industry, but they also highlight the critical need for high-quality, well-managed data.

As healthcare organizations continue to navigate the complexities of AI readiness, addressing data fragmentation will be a key factor in their success. By implementing robust data governance practices and ensuring interoperability, health plans can unlock the full potential of AI, leading to better care and more efficient operations.