How to Successfully Implement Artificial Intelligence in the Healthcare Sector

2019-12-25

Implementing artificial intelligence in the healthcare field can bring about transformation, providing opportunities for improving patient care, reducing costs, and impacting population health. However, a recent report from the National Academy of Medicine in the United States emphasizes the need for caution and thoughtful balance when it comes to implementing AI in healthcare. The report highlights several drawbacks of AI in the healthcare sector, such as unrealistic expectations, excessive hype, unfulfilled promises, and potential risks. Representatives from institutions such as Boston Harvard Medical School, Rochester, Mayo Clinic, OptumLabs, and Epic describe the challenges of implementing AI in healthcare and outline the measures that vendors must take to successfully implement AI in their healthcare systems. The report emphasizes the need for wider adoption of data models to support the development, deployment, and maintenance of AI tools. To address these issues, all stakeholders and the broader healthcare community must advocate for policies, regulations, and legislative mechanisms to improve data collection and enhance transparency for better utilization of patient health data while balancing financial incentives. Ethical healthcare, fairness, and inclusivity should be explicit goals in the development of medical AI. This will ensure representative population datasets and prioritize inclusivity and equality in healthcare. As the scale of AI implementation may expand, existing inequalities may be exacerbated, making the impact of individual persons or organizations far less significant compared to the global or national impact of AI technology. Additionally, the sustainability of AI tools should be assessed to understand how various deployment environments may affect fairness and inclusivity. Transparency is key to building trust between users and stakeholders. The report suggests that in the field of AI, the composition, semantics, sources, and quality of data used to develop AI tools should be fully transparent. Furthermore, it is crucial to clearly separate data, algorithms, and performance reports in AI discussions. Human-centered tools should be prioritized at the forefront of AI implementation, as machine errors are unacceptable to most people. The focus should be on promoting, developing, and evaluating tools that support humans rather than completely replacing them with full automation. Moreover, the AI community must develop an implementation and maintenance framework by integrating existing best practices in ethical inclusivity, software development, implementation science, and human-computer interaction. The framework should be aligned with goals and objectives, and the cost of AI tools should be balanced with use case requirements.

RECENT AI NEWS

RECENT AI TOOLS