Augmented Intelligence and Human Augmentation: Five Critical Foundations to Implementation

Article co-authored by Eleni N. Kalivas, PharmD, MBA , and Echo Schneider

In the first of our three-part series, “Augmented Ethics: The True Frontier of Augmented Intelligence in Healthcare,” we described some of the sensationalism and misconceptions surrounding the idea of augmented intelligence (AI).

Fortunately, the reality is far more promising and optimistic. Healthcare stands at the precipice of a transformative era, with AI revolutionizing how we deliver care. ?This stems from the potential of predictive diagnostics, care expansion at home, and operational analytics to drive targeted, sustainable growth. These shifts can create capacity, empower upskilling and staff growth, and redefine the roles of humans in care delivery.?

However, this transformation can only be fully realized when humans and technology come together in perfect synergy. This synergy is human augmentation; in healthcare, the goal is not to replace humans but to empower them. Staffing is often a point of fear in the AI conversation and whether AI will replace humans in transitioning to an AI hospital.? It’s exciting to see some of the most innovative health systems in the world embracing automation and AI in a way that creates the capacity to reskill and retain their staff, not replace them.? This allows healthcare workers across disciplines to evolve into high-level roles without workforce reduction. This empowerment is a key aspect of human augmentation's synergy, highlighting the value and integral role of all workers in the healthcare system.

Hurdles to AI Implementation

Health systems nationwide have been grappling with the many concerns related to AI integration. Apart from the high-level financial investment, governance approach, and mass employee and patient education, issues that have prevented implementation include vendor selection and assessment, ethical and security considerations, workforce anxieties surrounding job security, and an overall skepticism for the reliability and quality of integration. Given these varied concerns, how can health systems ensure AI enhances, rather than replaces, the workforce? ??

Foundations to Successful Implementation

Thoughtful solutions to successful AI implementation must include these five foundations.

1. Educational Programs: These programs can promote widespread AI literacy across the organization in preparation for integration. Educational modules should include how AI can assist in operational tasks, treatment planning, and diagnostics. ?Empower process owners and those performing repetitive tasks to be part of the automation solution. Asking them to embark on the journey with you allows them to have a voice and become less anxious and more excited about shaping the future. Beyond providing technical growth opportunities, people tend to like things more if they are involved in creating them (this is known as the IKEA effect).

2. Transparency: As future solutions are implemented, education should be on the specific solution, its goal, algorithm, and clinical vulnerabilities. The algorithms can be designed with varying sensitivity for true positives/negatives versus false positives/negatives, making education and AI literacy imperative for healthcare workers as they navigate the sensitivity and specificity of these models.

Transparency in AI solutions in healthcare is crucial to ensuring trust, accountability, and the ethical use of technology. It enables patients and healthcare providers to understand and validate AI-driven decisions and outcomes. It fosters informed consent, mitigates biases, and enhances patient care's overall quality and safety.        

3. Ethical Frameworks: The ethical use of AI in healthcare hinges on protecting patient privacy, avoiding unauthorized access, and complying with legal standards. Staff need to be educated on regulatory compliance so they know the ethical implications of AI and how to use it responsibly. Established data governance frameworks foster trust among stakeholders, including patients, healthcare providers, and regulators.?

4. Incremental Changes on the Implementation Roadmap: Communication is critical as you change. Your organization needs clear roadmaps for these changes to ensure ideas are appropriately vetted, a plan to track financial ROI is established, regulatory considerations are addressed, and feedback loops are established for continuous improvement (e.g., performance dashboards, daily monitoring, etc.).

Tech-forward organizations can efficiently scale the changes and AI solutions with robust safety and governance measures. They can handle larger datasets and more complex algorithms, thus maximizing the benefits of AI across various healthcare applications such as diagnostics, treatment planning, and patient monitoring.        

5. Enhanced Patient Care: Review potential automation or other AI enhancements based on their ability to improve patient care and experience. What tools or surveys are in place to measure a change in this area to ensure those intended outcomes are achieved?

Managing Stakeholder Pressure Post-Pilot

The development of AI tools involves multiple iterations, from initial conception to final production. Each process stage must incorporate a diverse, multidisciplinary workforce to ensure comprehensive testing and validation. This requires interdisciplinary buy-in in healthcare to ensure that all angles of the clinical picture and potential data limitations are considered. A homogenous or hyper-specialized team is likelier to overlook potential issues, and a diverse team would identify outlier situations and circumstances.

It's also crucial to exercise patience during the piloting process, understanding that thorough testing and iteration are necessary to identify and mitigate risks. Different considerations include integrating existing systems, ongoing user training, and continuous monitoring for unexpected outcomes.

Once a pilot is successful, stakeholders often face pressure to scale up and implement the AI system swiftly. However, a pilot's success criteria may only partially encompass real-world implementation complexities. Don’t rush this process, risking potential failure and undermining AI's benefits.

In the third and final article in this augmented ethics series, we will talk about the considerations healthcare leaders face on this journey.