With the growing fields of predictive medicine and genomics, personalized medicine has
become increasingly possible in healthcare. But this brings about ethical concerns in how
genomic data could be used. The current technological developments in artificial intelligence
(AI) sparks a new realm of ethical concerns beyond what already exists.
Predictive medicine aims to combine research on biomarkers with laboratory derived
genetic testing to understand an individual’s probability of developing disease. There are,
however, factors beyond genetics that can influence the probability of disease or treatment
outcome that are less measurable such as diet, activity, etc. This necessitates a system by which
probabilities of patient outcomes with specific treatments and how various treatments will
interact with a patient based on their genome can be analyzed.
With the increasing interest in artificial intelligence, mechanisms for integrating social traits and life experiences in combination with genetics to determine factors of health and illness has become increasingly possible. Thus far, many hospitals have made moves to incorporate AI for streamlining
electronic health records to increase the efficiency of hospital operations and patient
management. This process allowed for the standardization of electronic medical records into
datasets that have since been used in clinical research studies for insight on drug development.
Hospitals in Korea have already begun to use AI to incorporate individualized characteristics of
diet, health history, and daily life patterns into diagnosis and treatment mechanisms for patients.
Integration of genomic data in these algorithms can allow for the assessment of potential future
health risks based on genetic biomarkers for curating diagnoses and treatments specific to a
patient.
The machine learning aspect of AI can allow for the analysis of high volumes of patient
data at high speeds to streamline predictive genomics at the individual level. It can also identify
genetic patterns present among a population dataset that could be indicative as a biomarker for
disease. This has the potential for better representation of minority groups that have historically
been underrepresented in US healthcare by utilizing electronic medical records throughout the
world.
However, there is also potential for the creation of race-based patterns that are not
scientifically based. An example of this would be that black patients are “healthier” than white
patients of equal sickness due to the lower healthcare costs they generate, but the reality of that
situation has to do with the fact that black patients are less likely to go to the hospital for
treatment than their white counterparts. Furthermore, similar scenarios of incomplete datasets
can create bias that may result in the overgeneralization and overdiagnosis of certain patients.
This prompts the need for sufficient regulations on the extent of the usage of AI in healthcare.
One of the largest issues with the reliance of AI on electronic medical records is the lack
of standardization between the records of different physicians/hospitals that may be assessed at
differing degrees. This could be the overanalysis in differentiation between how “minimal” and
“minute” is perceived by the algorithm. Furthermore, every instance of sickness may also not be
in the record if it did not necessitate hospitalization, creating “holes” in the full history from
patient to patient, which these machine learning algorithms have to “ignore” to curate recommendations for a specific patient creating significant room for error.
The ethical concerns involved in using AI for predictive medicine lie in the fact that patient consent is not explicitly required for their medical records to be utilized in genetic analysis programs and algorithms, so patients might be wary of the risk of their data being sold, aggregated, and commodified at the discretion of the hospital/provider. This necessitates the creation of regulations involving the monetization of patient health data and sufficient patient privacy at the institutional level.
A basis for the testing of the accuracy of AI algorithms in utilizing genetic sequencing, biomarker identification, and social factors for making clinical recommendations is also required, which public health policy looks to establish. The future implications for ethical regulations on the usage of AI in predictive genomics involve obtaining informed consent for data use, algorithmic biases, patient safety, and data privacy in order to address the current pitfalls seen from the usage of AI in healthcare.
Aarathi Manchikalapudi is a rising 2nd Year student at the University of Virginia. She hopes to major in neuroscience and global public health, and is currently involved in tissue engineering research.
Sources:
Ai in healthcare: How ai is Revolutionizing Healthcare with predictive analytics, personalized
medicine, and disease detection. Managed IT Services, Microsoft Gold Partner, & IT
Help Desk. (n.d.). https://www.datalinknetworks.net/dln_blog/ai-in-healthcare-how-ai-is-revolutionizing-healthcare-with-predictive-analytics-personalized-medicine-and-disease-detection
Jen, M. Y. (2022, October 17). Predictive medicine. StatPearls [Internet].
Lee, S., & Kim, H.-S. (2021). Prospect of artificial intelligence based on Electronic Medical
Records. Journal of Lipid and Atherosclerosis, 10(3), 282.
Naik, N., Hameed, B. M., Shetty, D. K., Swain, D., Shah, M., Paul, R., Aggarwal, K., Ibrahim,
S., Patil, V., Smriti, K., Shetty, S., Rai, B. P., Chlosta, P., & Somani, B. K. (2022). Legal
and ethical consideration in artificial intelligence in Healthcare: Who takes
responsibility? Frontiers in Surgery, 9. https://doi.org/10.3389/fsurg.2022.862322
Phillips, S. P., Spithoff, S., & Simpson, A. (2022). Artificial Intelligence and predictive
algorithms in medicine. Canadian Family Physician, 68(8), 570–572.
Predictive medicine. Predictive Medicine - an overview | ScienceDirect Topics. (n.d.).
Comments