A report from the Ada Lovelace Institute, in collaboration with the Nuffield Council on Bioethics, has provided UK healthcare leaders with an in-depth analysis of the potential and the risks presented by AI powered genomic health prediction (AIGHP).
AIGHP is the use of AI in genomic analysis to carry out polygenic scoring. The institute described polygenic scoring as the assessment of "the collective impact of multiple (individually small) genetic variations on the likelihood of a given person exhibiting a given trait (such as developing a particular disease), relative to the rest of the population".
Polygenic scoring uses high levels of complex data, which AI is able to analyse. At present, AI is not widely used for polygenic scoring, but the report points out that the use of AI is likely to increase and has attracted major investment. AIGHP could provide healthcare providers with the ability to make predictions about a patient's future health and how they respond to medication using genomic data.
The report emphasises that this can support a move to preventative healthcare.
Potential for insights
"If appropriately integrated into a healthcare system, AIGHP could provide people with insight into their risk of developing particular diseases, inform beneficial lifestyle changes and help people be alert to symptoms of conditions for which they are at higher risk," it says.
"The use of AIGHP to improve understanding of how an individual might respond to a given drug or medication could allow for better prescribing practices, reduce waste, improve outcomes and avoid harmful side effects. If AIGHP enabled even marginal improvements, this could still be significant given the huge burden placed on the NHS by the ineffective use of drugs and widespread adverse drug reactions."
The technology can also be used by healthcare providers to shape policy, for example, by prioritising screening and planning resources.
In common with other AI developments, the Ada Lovelace Institute points out that AIGHP comes with privacy, discrimination and societal concerns if the technology was widely adopted. These concerns, the report says, stem from uncertainty about the effectiveness of the technology, while the use of sensitive personal data could spark new forms of discrimination based on genetics, such as in health insurance.
The report also warns that there is currently no information on public attitudes towards the use of AIGHP, which could lead to structural challenges. The report, therefore, recommends that the government and the NHS: "ensure that there is convincing scientific justification for the use of AIGHP and that appropriate safeguards are in place".
This report, Predicting: The Future of Health is the second of two by the collaboration in the future of healthcare.
