Artificial intelligence (AI) and machine learning (ML) are opening up new areas in healthcare with great potential to improve clinical outcomes, manage resources, and support therapeutic development. It also raises ethical, legal and operational issues, which can result in increased risk.
Where are AI and ML today? Go, stop, go.
2023 has provided a roller coaster of activity considering the phenomenal progress and its implications, resulting in an effort to corral unchecked expansion. Many industry leaders, seeing the rapid growth of AI technology, called for a pause in continuous progress for at least six months, but saw other companies continue to take advantage of the target-rich opportunities. I just had to. This push-and-pull reflects the need to be cautious in investing in and using AI/ML.
Activities at the government level are also evolving rapidly. In late 2022, the White House will prioritize civil rights and democratic values and unveil the “AI Bill of Rights Blueprint” to guide the deployment, design and use of automated systems. On April 3, 2023, the FDA issued a draft guidance for developing FDA’s regulatory framework for AI/ML-enabled device software capabilities. This guidance proposes an approach to ensure the safety and effectiveness of AI/ML that uses adaptive mechanisms to incorporate new data and improve in real time. Given the lack of comprehensive federal legislation on AI, states have been active in enacting privacy laws. Additionally, to comply with patient-centric health-related AI standards, the Coalition for Health AI released his “Blueprint for Trusted AI Implementation Guidance and Assurance in Healthcare” in early April.
This accelerated development calls for action internationally. Italy temporarily banned ChatGPT in April and launched an investigation into alleged GDPR violations of the application. Spain, Canada and France also expressed similar concerns and launched investigations. EU lawmakers are calling for an international summit and new AI rules, including a proposed AI law. As a result, oversight and accountability practices for AI/ML technologies are increasingly becoming a regulatory priority.
Key Areas of AI Growth
- Personalized care: AI has the potential to detect disease and guide treatment by integrating current medical research and treatment resources in real time. Predictive elements of AI technology can predict treatment outcomes, improving quality of care and minimizing costs. Examples of patient-specific applications include predictive analytics to determine patient outcomes with high accuracy, personalized provider matching based on modeled variation in provider outcomes and patient-specific diagnoses , including timely clinical intervention through wearable monitoring with AI decision-making tools. AI’s pattern detection capabilities are particularly useful in medical imaging, as pattern recognition supports disease diagnosis and prognosis. Preclinical AI helps streamline workflows, monitor hospital bed availability and readmission rates, and identify health equity gaps.
- Early detection and diagnosis: AI algorithms will be able to accurately detect and diagnose serious diseases such as ALS, kidney failure, and Alzheimer’s, years before conventional diagnosis is possible. AI detection capabilities are also being implemented in general wellness areas such as sleep, diet and mental health monitoring, leading to early detection of related ailments and can improve the effectiveness of treatments. AI algorithms have been shown to predict diabetes disease with a high accuracy of over 90% and achieve clinical accuracy comparable to that of the average physician in diagnosing written test cases.
- Development and Discovery of Therapies: AI can scrutinize and analyze vast amounts of digitized drug information to address complex clinical problems. As a result, there has been a notable increase in partnerships between traditional pharmaceutical companies and AI-driven companies. AI is particularly relevant to drug discovery, screening, and molecular design. Clinical trial design. and manufacture of pharmaceuticals.
Legal and industry considerations
The goal of AI/ML technologies is to deliver “smarter” care, but so far, the patient-provider relationship remains critical to ensuring patients receive the right care. is. The growth of AI in healthcare and life sciences is also giving rise to new legal and regulatory considerations, particularly in the following areas:
- FDA and SaMD: The use or assistance of AI algorithms in clinical decision making may fall within the scope of FDA regulatory agencies if the technology meets the definition of a “medical device.” The FDA has developed a framework to regulate AI/ML-enabled medical devices and AI/ML-based technologies that are “software as medical devices.” As technology evolves and public interest grows, the FDA is actively working to issue guidance on these topics.
- Ethics and research: As the application of AI expands to a range of services traditionally performed by licensed physicians, it could raise questions about unlicensed medical practices. The use of patient data in the development and testing of AI technology requires informed consent and may begin oversight of an institutional review board. As AI becomes more prevalent, the need, or lack thereof, of human oversight may continue to be a concern, especially to monitor AI’s ability to produce false results and cause unnecessary or erroneous care. . Moreover, malicious and unintended applications of AI, such as biohacking, bioweapons, and weaponization of health information, require careful protection and proactive vigilance by all to ensure proper oversight.
- Intellectual property and data assets: Healthcare innovators in the AI/ML space face a different IP environment, as AI/ML systems may not receive the same protections as traditional outputs. For example, copyrights and patents may not accompany output that is not the work of human authors or developers. Rights to data assets, such as raw and derived data underlying AI algorithms, should also be monitored.
- Privacy and data rights: Medical privacy laws and regulations can be relevant at both the federal and state levels. Patient information may be subject to protections under HIPAA and other state laws and may be required to be anonymized before such data is shared and used to develop AI/ML products. Additionally, private lawsuits related to consumer privacy laws and data rights provide grounds for an individual to monitor and, in some cases, oppose the use of her personal data in AI development.
- Refund and Compensation: The use and adoption of AI by healthcare providers and institutions is highly dependent on financial incentives, such as reimbursement rates based on new AI iterations of innovation and whether AI services are covered by payers. As the industry moves toward value-based care, AI may provide additional tools and opportunities.
- Potential bias and inaccuracy: Despite the groundbreaking and revolutionary potential of AI/ML technology, algorithms in AI technology may detect patterns using human-annotated data. The patterns may (1) be based on old, homogeneous, or incomplete datasets, and (2) be susceptible to replication and persistence. Prejudices based on race, gender and even age. As a result, increasing attention is being paid to diversifying and expanding medical datasets to identify and mitigate these potential biases.
pivotal moment
The growth of AI/ML is at a pivotal moment, driven by the dichotomy of pushing AI technology development or calling for a pause. As industry and government consider AI’s enormous potential and risks, they will closely track developments to ensure that innovations are implemented in ways that promote societal benefits while mitigating unintended harm. is most important.
Despite the uncertainties and risks, the implementation of AI with the right compliance framework and infrastructure presents an exciting opportunity to improve patient outcomes, increase efficiency and transform healthcare into new frontiers. Offers.
Photo: ipopba, Getty Images
