He explained how the interpretation of X-rays and other diagnostic modalities will be faster and possibly more accurate than the human brain; and how even an experienced pathologist will be replaced by AI in interpreting complex histological slides, especially when it comes to very rare diseases.
This insightful and cutting-edge interview prompted me to wonder how AI may possibly help some of the most commonly seen conditions in pediatric offices across North America.
I have also found that many parents are leery about these medications because they project — meaning they assume that because a family member reacted adversely to a medication, it will cause the same problems in their child or teenager.
ADHD is a common condition in both children and adults. The most prescribed medications are methylphenidates, mixed amphetamine salts, atomoxetine and guanfacine. Each of these medications has distinct pharmacokinetic profiles. When it comes to methylphenidate and atomoxetine, we have discovered quite a bit about the enzymatic activities and how variations affect the speed of responses, based on how fast or slow the medications are metabolized.
Cytochromes such as CYP2D6 and CES1 play a big role in the metabolism of these medications. Some patients are poor metabolizers and when atomoxetine is prescribed, they experience up to a 10-fold higher exposure compared to normal metabolizers; with methylphenidate, there could be a 2.5 higher exposure in slow metabolizers compared to normal metabolizers.
Researchers are having a feast trying to gain knowledge in the field of pharmacogenetics and a recent scientific article in Pub Med (pubmed.ncbi.nlm.nih.gov/36068472/) concludes by saying “Having this pharmacogenetic information available may aid clinicians and patients when choosing medications and doses to treat ADHD.” It sounds good but how real is it at this time?
Another area which costs the health-care system billions of dollars is polypharmacy. In many chronic disease managements and complex neurological cases, a number of medications are combined and there is always a risk that these multiple medications may interact with each other and cause side effects in patients.
So far, effective tools to address this problem have remained elusive.
In another recent article, sourced from the National Library of Medicine, titled “Implementation of Artificial Intelligence Tools for Chronic Disease Management in Primary Care Settings” the authors paint a picture of how their AI model provides exciting changes ahead in primary care; changes which will prevent drug interactions and changes which will make current alerts via electronic health records of possible drug interactions look ancient and out-dated.
They admit that a major obstacle is the lack of communication between the departments of genetics, pharmacology and clinical medicine, although that is rapidly changing across many campuses in North America. The way reimbursement will unfold and the cost of paying for this research are two other obstacles. It is a complex topic but at the same time very interesting to try and read the main ideas behind the potential of AI. For those interested in more detail, the link to this research can be found at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8224063/
I often think of how relatively simple it is to treat urinary tract infections compared to ADHD. All one has to do is collect the urine, do culture testing, find the bacteria and pay attention to the reports which tell us which antibiotics to use and which ones to avoid. Hopefully, AI will make the choice of medications for ADHD similarly easier one day, but for now, it is difficult to predict how long this will take.
It sounds very promising but, as famous New York Yankees baseball player Yogi Berra used to say, “It is difficult to make predictions, especially about the future.”
Dr. Nieman is the founder of Centre 70 Pediatrics. He has written for the Calgary Herald since 1999. For more information, see www.drnieman.com