Government Accountability Office (GAO’s) latest report has identified several challenges affecting the development and adoption of machine learning (ML) technologies in medical diagnostics. The report also offers some policy options for addressing these challenges and enhancing benefits.
The 103-page report “Technology Assessment Artificial Intelligence in Health Care: Benefits and Challenges of Machine Learning Technologies for Medical Diagnostics” to congressional requesters has been jointly published by the GAO and the National Academy of Medicine (NAM).
Cost of diagnostic errors
Emphasizing the importance of effective, efficient, and accurate clinical diagnostic process, the GAO’s report quotes a report by the Society to Improve Diagnosis in Medicine according to which each year diagnostic errors affect more than 12 million Americans with aggregate costs likely in excess of $100 billion.
The report says that challenges to the development and use of ML medical diagnostic technologies raise technological, economic, and regulatory questions. “… AI tools developed using historical data could unintentionally perpetuate biases, reduce safety and effectiveness for different groups of patients, and produce disparities in treatment,” notes the GAO report.
Some of the several challenges affecting the development and adoption of ML in medical diagnostics found by the GAO are:
- Demonstrating real-world performance across diverse clinical settings and in rigorous studies as, according to experts and stakeholders, medical providers may be reluctant to adopt ML technologies until its real-world performance has been adequately demonstrated in relevant and diverse clinical settings.
- Meeting clinical needs, such as developing technologies that integrate into clinical workflows as medical providers are less likely to adopt ML technologies that do not address a clear clinical need, and many ML diagnostic technologies do not progress from development to adoption for this reason
- Addressing regulatory gaps, such as providing clear guidance for the development of adaptive algorithms as gaps in the regulatory framework may also pose a challenge to the development and adoption of ML technologies.
The policy options
As per the report these and many other challenges affect various stakeholders including technology developers, medical providers, and patients, and may slow the development and adoption of these technologies.
To address these implications, the GAO report has identified three policy options to help address challenges or enhance benefits of ML diagnostic technologies. These options include encouraging evaluation of these technologies, improving high-quality data access, and promoting collaboration across stakeholders.
- Evaluation: Policymakers could create incentives, guidance, or policies to encourage or require the evaluation of ML diagnostic technologies across a range of deployment conditions and demographics representative of the intended use.
- Data Access: Policymakers could develop or expand access to high-quality medical data to develop and test ML medical diagnostic technologies.
- Collaboration: Policymakers could promote collaboration between developers, providers, and regulators in the development and adoption of ML diagnostic technologies.