Dimensions of Artificial Intelligence in the Healthcare Industry – Healthcare Blog
Steven Zecora
On December 19, the U.S. Department of Health and Human Services (“HHS”) issued a request for information seeking to leverage artificial intelligence (“AI”) to reduce health care costs and make America healthy again.
As discussed in this article, artificial intelligence can be used in many ways to help reduce medical costs and improve care. However, to achieve major breakthroughs in artificial intelligence, the U.S. Department of Health and Human Services will need to overhaul its regulatory approach to drug discovery and development.
Dimension #1. Incorporating artificial intelligence into drug discovery
The greatest benefits of AI to healthcare industry performance can be realized through drug discovery. Taking into account the cost of failure, FDA drug approval costs society nearly $3 billion on average and takes decades from birth in the laboratory to market.
In contrast, artificial intelligence can identify potential treatments faster than traditional methods by processing large amounts of biological data, revealing hidden causal relationships and generating new actionable insights.
Artificial intelligence holds particular promise for complex, multifactorial diseases, such as neurodegenerative diseases, autism spectrum disorders, and multiple chronic conditions, where traditional reductionist approaches have failed.
In the short term, the U.S. Department of Health and Human Services should direct its funding toward basic research generated by AI, especially for hard-to-treat diseases. At the same time, the FDA should establish a new approval system for AI-initiated projects to achieve breakthrough treatments on compressed timelines.
Dimension #2. Incorporating artificial intelligence into the drug development process
Relying solely on AI for drug discovery while incorporating its advances into the current approval process will undermine the use of this technology.
Instead, improvements in artificial intelligence have made it possible to meet exhaustive regulatory documentation requirements, which now account for up to 30% of compliance costs.
In the short term, AI can improve drug development by:
- Automate and validate regulatory documents
- Strengthening trial design and participant stratification
- Monitor safety and effectiveness in near real-time
- Reduce administrative and compliance costs
In the UK, for example, the Medicines and Healthcare products Regulatory Agency reported that clinical trial approval times for artificial intelligence and related reforms have doubled.
To realize greater long-term benefits, HHS should break down all clinical efforts utilizing AI into one long-term trial rather than discrete Phase I, II, and III trials, given that AI can be used to continually update and validate documentation. This change does not require a statutory change or agency rulemaking because clinical trial design is not codified in FDA rules.
Safety results can be reviewed and reported in real time as participants are added to the trial. Once a trial exceeds a certain number, such as 1,000 participants, and proves its effectiveness and meets specified safety protocols, it will be approved for rollout. In this approach, the government's role is to act as an auditor to verify the results of the trial. This function will include experimental validation, mechanistic understanding, and ethical oversight.
With these changes, FDA personnel will transition from occasional gatekeepers to ongoing auditors, which will require fundamental changes in organizational culture. While safety concerns will still be important, responsibilities and liabilities will be more equally shared between applicants and trial participants. Additionally, the long-term suffering of existing patients will be factored into the public benefit analysis when preliminary safety results are reviewed.
Dimension #3. Strengthen data collection and empower artificial intelligence
Comprehensive and accurate data is critical to the success of artificial intelligence. This is another area where the healthcare industry is failing.
The industry evolves as each provider or family of providers encourages their patients to sign up for a customer portal. Providers generally treat the information on these portals as their own for research purposes. However, the provider does not own this data. Each patient has his or her own data.
To expand the scope and applicability of health care data, HHS should develop national standards for patient-facing data collection:
- Use interoperable formats
- Capture diagnostic results and associated explanatory variables
- Preserve patient ownership and informed consent
- Enable longitudinal tracking while protecting privacy and security
Once this format is established, HHS should set a goal of recruiting 100,000 participants within two years.
DDimension #4. Using artificial intelligence to establish standards of care and price caps
The United States has no national standards of care for disease or other health conditions. Patients often do not understand the nature of their pain, treatment options, or the costs of various treatments.
At the same time, HHS may fund basic research on a specific disease, the FDA may (or may not) approve it, Medicare may (or may not) cover it, some insurance companies may cover the treatment, and some may not.
In addition, the cost of different treatments at different treatment facilities can vary widely without the patient being aware of it.
Underpinned by market dysfunction, healthcare practitioners have a desire (and financial incentive) to provide the best (and potentially most expensive) service to their patients.
In short, market failure is primarily related to a lack of actionable information.
In the short term, AI can help address these failures by aggregating and analyzing how care is delivered across the country and identifying patterns associated with better outcomes and lower costs. These insights can be used to inform evidence-based minimum standards of care and increase transparency on pricing and performance.
In the long term, the output of these systems can be used to establish minimum standards of care for all (or most) diseases. These standards will be mandated to be covered by insurance. At the same time, the output of these care standards can be supplemented by regional price caps for various practices based on comprehensive industry analysis.
As experience is gained from these informative AI systems, future versions could be programmed to automatically calculate prescribed minimum standards of care and price ceilings to mimic the function of demand and supply curves. The algorithm can be constructed using a specified level of subsidy provided by the federal government as the equilibrium. When federal subsidies exceed certain preset limits, AI will address the imbalance by providing lawmakers with options that would lower price caps for certain conditions and/or lower minimum standards of care.
Beyond stated federal subsidies, certain categories of patients will not receive payment for the best available treatment (unless they have supplemental insurance) and/or certain health care providers will suffer reduced profits.
This approach would require congressional approval, but this trade-off is occurring now—and there is no sensible alternative. In this dimension, artificial intelligence can be used to solve the massive information failures in the industry and solve the problem of increasing subsidies.
Dimension #5. Incorporating artificial intelligence into HHS’ internal processes
Artificial intelligence can also improve the efficiency and effectiveness of HHS's internal operations. While the potential percentage gains will be smaller than those in the discovery and development dimensions, given the scale of federal health care spending, even modest improvements could yield meaningful savings.
in conclusion
Artificial intelligence offers the opportunity to significantly improve healthcare outcomes and efficiencies, but only if it is integrated into regulatory and governance frameworks designed for its functionality. Forcing AI into existing structures will dilute its impact and increase implementation risks.
Each of these areas requires a separate, dedicated, multidisciplinary team reporting to the Office of the Deputy Secretary. After determining the strategic direction for each dimension, these teams should be tasked with:
- Develop a detailed implementation plan, including budget requirements
- Identify any legal or regulatory barriers
- Establish timelines, milestones and evaluation criteria
- Address ethics and fairness issues
Drug discovery and drug development are the dimensions where AI implementation will have the greatest impact. HHS should leverage outside expertise to develop the details of an appropriate regulatory framework for these aspects.
Detailed plans to implement artificial intelligence should be approved and finalized by the end of 2026. As this article outlines, HHS should take a proactive, forward-thinking role in leveraging AI to limit health care costs and improve care.
Twenty-three years ago, Steve Zecola sold his web application and hosting business when he was diagnosed with Parkinson's disease. Since then, he has worked in consulting, taught at graduate business schools, and practiced extensively
