Precision health holds the promise of providing the right treatment to the right person at the right time. But to do this, it needs the help of augmented intelligence (AI), often called artificial intelligence, to efficiently organize, analyze and interpret data so physicians can deliver targeted diagnoses, prognoses and treatment recommendations.
Physicians already appreciate the role AI can play—an AMA survey of more than 1,000 doctors found that nearly two-thirds can see AI’s potential benefits. But a lot still needs to happen to ensure its application to precision health is safe, effective and equitable.
An AMA Ed Hub™ CME series introduces learners to foundational principles in AI and machine learning, a subdomain of AI that enables computers to learn patterns and relationships from data without being explicitly programmed by people. Developed by the AMA ChangeMedEd initiative and the University of Michigan DATA-MD team and geared toward medical students, it is also suitable for residents, fellows, practicing physicians and other health professionals.
The sixth module in the series, “AI and Precision Health,” examines the potential of AI to fulfill the goals of precision health.
From AI implementation to EHR adoption and usability, the AMA is making technology work for physicians, ensuring that it is an asset to doctors—not a burden.
Why they go together
“Precision health is a health care approach that tailors diagnosis, prognosis, and treatment to individual patients based on their unique genetic, biomarker, phenotypic, or psychosocial characteristics,” the module says. “Al augments clinicians' capacity to analyze and interpret complex data, aiming to provide more personalized, efficient and effective care to ultimately improve patient outcomes.”
At the heart of this relationship is data. Precision health requires huge amounts of it, and it all has to be organized, processed and analyzed to be useful in clinical settings.
“Al has the potential to effectively serve this purpose,” the module says. “With vast datasets encompassing genomics, biomarkers, clinical records, environmental data and social factors, Al is the driving force powering precision health.”
The AMA has developed advocacy principles that build on current AI policy. These principles (PDF) address the development, deployment and use of health care AI. Meanwhile, AMA Ed Hub also features a separate, 16-credit CME course on artificial and augmented intelligence in health care.
Where datasets can still fall short
For all of their promise, however, AI and precision health have to reckon with several significant challenges in data utilization. These include:
- Lack of diverse and representative datasets due to the limited availability of high-quality, essential health data—encompassing factors such as race, ethnicity, gender, sex and geography—to train models.
- Data privacy concerns around encryption, deidentification and access control to protect data from misuse that could lead to patient harm.
- The need for longitudinal data, as many patients’ health histories are incomplete in their electronic health records.
In addition, precision health faces challenges beyond data. A case in point: worsening inequities driven by varying access to resources within health systems.
“Al-powered precision health has the potential to magnify these gaps,” the module notes. “This highlights the need for equitable access to Al-driven tools and underscores the importance of bridging the digital divide to ensure that health care advancements are beneficial to all patients and health care systems.”
Learn more with the AMA about the emerging landscape of augmented intelligence in health care (PDF).
What needs to happen
Drawing lessons from the success of the University of Michigan Precision Health Initiative, the module suggests several strategies to help ensure that precision health, aided by Al, transforms health care for the better.
Assemble large, longitudinal cohorts. “Long-term, integrated cohorts like the UK Biobank, Million Veteran Program, FinnGen, and All of Us have accumulated vast amounts of genomic, lab, and lifestyle data, presenting opportunities for medical breakthroughs,” the module says. "The next step is to merge data from multiple cohorts to create a comprehensive dataset for researchers.”
Commit to improved diversity. One major challenge in precision health is insufficient diversity in the pool of research participants, the module notes. “This gap can worsen health inequities and limit biological discoveries. We must commit to improved diversity. Enhanced data depth allows for refined measures beyond race and ethnicity. Simultaneously, diversifying both study populations and the biomedical research workforce is essential.”
Pivot to routine genomic analysis. So far, clinical genomic analysis has been performed only when evaluating specific cancers or rare genetic diseases. “Moving forward, whole genome approaches will become a standard step in understanding, preventing, detecting and treating diseases,” the module says.
Capture expanded phenomics and environmental data. Precision health will see the expanded use of phenotype, exposure and lifestyle data, and this will require integration of data from health claims, national vital statistics and geospatial resources. Wearables will play a part too—by providing data about activity levels, physical metrics and exposures.
Protect data privacy. The success of precision health depends on the involvement of large populations, but participation won’t always be easy to come by. “Historically, science has not always been trustworthy or treated participants equitably,” the module notes, adding that this has been especially true for patients from historically marginalized racial and ethnic groups. “Transparency, active community engagement and involvement of participants in research governance will enhance trust and drive precision health in a more culturally sensitive direction.”
Embrace stakeholder collaboration. “To realize the vision of precision health for all communities, a strategy involving international collaboration, engagement of diverse participants and researchers, comprehensive measurement of populations, wide dissemination of clinical and research data, and integration of this knowledge into clinical practice within a learning health care system is imperative,” the module says. “Collaboration among stakeholders will enable us to embrace precision health holistically across populations.”
The module also explores how precision health is tailored to an individual’s characteristics, as well as the role of precision health in the response to the COVID-19 public health emergency and other key accomplishments in precision health over the last two decades.
Periodic knowledge checks test the user’s understanding of how concepts are applied.
The CME module “AI and Precision Health” is enduring material and designated by the AMA for a maximum of 0.75 AMA PRA Category 1 Credit™.
It is part of the AMA Ed Hub, an online platform with high-quality CME and education that supports the professional development needs of physicians and other health professionals. With topics relevant to you, it also offers an easy, streamlined way to find, take, track and report educational activities.
Learn more about AMA CME accreditation.