At HLTH USA 2024, artificial intelligence (AI) was a dominant focus, with a dedicated pavilion showcasing its wide-ranging applications, from AI-driven medical assistants to automated medical scribing. Meanwhile, experts took to multiple stages to discuss AI’s amazing potential in healthcare while candidly addressing the real-world hurdles faced during implementation. 

Throughout HLTH USA, several key themes emerged during discussions on AI. These included data optimization and interoperability, as well as the opportunities emerging in combining AI with remote wearables. Meanwhile, the conference highlighted the importance of responsible use of AI and tackling equitable access to AI-enhanced healthcare. 

• Optimization of Data: One of the most significant challenges in applying AI to healthcare is managing and optimizing the vast volume and variety of data generated in the industry. Although healthcare systems produce enormous amounts of data, much of this information remains underutilized due to: its multimodal nature, the need for customization of data in healthcare, and the risk of hallucinations by AI models. To address these challenges, companies are developing technologies, such as foundation models, to streamline data integration, reduce the need for extensive customization, and enhance accuracy. 

  

• Remote Monitoring and Wearables: Remote patient monitoring through wearables combined with AI has huge potential to spot patterns and alert patients, physicians, and care teams to any potential health issues sooner. These tools can track a range of health metrics, from heart rate and blood pressure to blood oxygen levels, all while offering real-time data to patients and healthcare providers. Recent innovations in wearable technology showcased at HLTH 2024 emphasize health monitoring and expanding accessibility. For instance, wearables tailored to assist individuals with visual impairments represent a step forward in equity, offering benefits beyond traditional health metrics.

• Interoperability: Interoperability emerged as a critical theme for advancing healthcare’s digital infrastructure. As AI adoption grows, it is becoming clear that simply accessing data isn’t enough — data must be complete, standardized, and usable. Currently, healthcare systems are largely designed for billing rather than comprehensive data collection in the service of patient care, posing a challenge for AI applications that need clean, research-quality data. However, the push for adopting AI appears to be pulling interoperability into the spotlight, and advances are already being seen in healthcare. 

Responsible and Equitable Use of AI

Discussions throughout HLTH highlighted the importance of responsible AI usage, emphasizing initiatives to establish transparency and safety standards for healthcare technologies. A working group from the Coalition for Health AI (CHAI) introduced its draft Model Cards for health AI, which include essential information such as the developer’s identity, intended applications, and target patient demographics. This structured approach is designed to enhance the evaluation and understanding of AI technologies in the health sector.

Moreover, AI has the potential to greatly expand access to quality healthcare for the 4.5 billion people currently lacking essential health services worldwide. Discussions at HLTH USA underscored the critical need to address biases in AI algorithms to ensure that advancements in technology equitably benefit all populations. By focusing on these issues, stakeholders aim to create a more inclusive healthcare landscape that leverages AI responsibly.

AI's Road Ahead in Healthcare

As described by one delegate, AI isn’t a magic powder that can be sprinkled over healthcare’s problems. It will take a coordinated and intentional approach with commitment from top leaders to get AI to a point where it is integrated and making real and large-scale changes in healthcare, but judging by the quality and commitment of AI companies at HLTH USA 2024, it has and will affect real change for providers and patients.