Bridging the AI Divide: Building Local Capacity in Healthcare Systems

Bridging the AI Gap in Healthcare: A Push for Local Capacity Building

In recent years, the adoption of AI tools in healthcare delivery has been concentrated in resource-rich settings, such as academic medical centers, leaving lower-resource environments struggling to keep up. This disparity is reminiscent of past technological transitions, including the adoption of electronic health records (EHR) and telehealth, where similar challenges were met with targeted government interventions.

AI in Healthcare: Opportunities and Challenges

Healthcare organizations are increasingly turning to AI tools to improve care delivery, enhance patient experiences, and reduce costs as part of the quintuple aim. Tools like large language models (LLMs) are being implemented to predict inpatient mortality risks and improve clinical documentation. However, organizations with limited resources face significant hurdles in adopting these technologies effectively, leading to a widening digital divide.

Many health delivery organizations (HDOs) lack the technical expertise, funding, and data infrastructure to manage AI tools throughout their lifecycle, resulting in inefficient implementation that may exacerbate existing inequities. This divide is evident with more than 20% of office-based practices not adopting certified EHR technology, which supports many AI applications. Without targeted interventions, the gap in AI adoption continues to grow, impacting quality of care and health equity.

Lessons from Past Technological Shifts

Historically, technological shifts in healthcare, such as the implementation of EHRs and telehealth, were successfully supported by federal initiatives. The HITECH Act and the establishment of Regional Extension Centers (RECs) provided crucial support to smaller, underserved practices, driving significant EHR adoption by offering technical, legal, and financial assistance. Similarly, Telehealth Centers of Excellence have facilitated telehealth adoption, bolstered by infrastructure investments and reimbursement changes.

These initiatives underscore the importance of a comprehensive support system, beyond financial incentives, to drive widespread adoption of new technologies in healthcare.

Implementing a Hub-and-Spoke Model

To address the AI digital divide, the Health AI Partnership (HAIP) proposes a hub-and-spoke network model, where resource-rich hubs connect with lower-resource spoke sites to provide necessary technical, regulatory, and educational support for AI adoption. This model has been piloted by HAIP in several facilities across the United States, focusing on building local AI capacities and fostering a community of learning and support among participants.

This approach involves collaboration among various stakeholders, including technological firms, payers, professional service firms, universities, and coordinating centers, to offer comprehensive support across the AI product lifecycle. Such a model aims to empower all healthcare settings to effectively utilize AI tools, ensuring that benefits are realized across the board.

Moving Forward with Strategic Investment

To achieve meaningful AI adoption across diverse healthcare settings, strategic investment from federal and state governments, as well as private sectors, is imperative. These investments should focus on building local capacities, providing infrastructure, and offering incentives for AI adoption. Learning from the pilot programs, further public investment is necessary to expand these models nationally, addressing the AI digital divide comprehensively.

As the healthcare landscape continues to evolve, drawing on past experiences and innovative models like the hub-and-spoke approach can lead to a more equitable utilization of AI in healthcare, ultimately enhancing patient care and health outcomes.