Artificial
Intelligence (AI) in Medicine. Many positives, some negatives, but can we benefit?
Dr. C.V. Alert,
MB BS, DM. FCCFP.
Family
Physician.
Artificial Intelligence (AI) is a rapidly evolving field of
computer science that aims to create machines that can perform tasks that
typically require human intelligence. AI in medicine
and healthcare is rapidly transforming the field by enhancing diagnostics,
treatment plans, and patient care. AI algorithms can analyze medical data
to identify patterns and insights that may be missed by humans, leading to
earlier and more accurate diagnoses, personalized treatment strategies, and
improved patient outcomes.
Here are some
areas in which AI is being used to improve healthcare:
1. Diagnosis
of complex cases.
Very recently,
it was announced that a new AI tool (from Microsoft) can diagnose complex
illnesses more accurately than most doctors. This AI tool blends the strengths of
both generalists and specialists. Most doctors either have a broad base of
knowledge (e.g. family physicians) or deep expertise in a particular field
(e.g. specialists), but this AI combines both. In testing, this new tool
correctly solved over 80 percent of the specifically selected challenging cases,
compared to just 20 percent by human doctors. Even more impressive, it managed
to do so while ordering fewer tests, suggesting a more efficient and
cost-effective approach to diagnosis.
The system’s
ability to update its diagnosis in real time, as more data is added, means it
can adapt quickly and provide up-to-date recommendations. This is a big deal,
especially in complex cases where every detail counts. Researchers say the AI’s
clinical reasoning skills are beyond what any single doctor can offer, thanks
to its ability to process vast amounts of information from different medical
fields at once.
Despite these
promising results, Microsoft is claiming that this is a pathway to ‘medical
superintelligence’, but is not positioning this AI tool as a replacement for
doctors, at least, not yet. The tool is still being tested in real clinical
settings to see how it performs outside of controlled experiments. Safety
checks and clinical validation are ongoing, and the company says it will only
move forward with broader use once these hurdles are cleared. Stay tuned.
2. Personalized
Treatment Plans:
AI can analyze
patient data to predict individual responses to different treatments, allowing
for the development of more targeted and effective treatment
plans. AI-powered tools can also help optimize medication dosages,
minimizing side effects and maximizing therapeutic benefits.
On the other
hand, many of the ‘newer’ classes of medications, like the monoclonal antibody
treatments for a wide variety of diseases, many newer anti-cancer drugs, and
the new classes of anti-obesity medications, are outside the price range of
many Caribbean patients and most Caribbean National Drug Formularies, so we are
unlikely to see any of these medications in the near future, or at least until
patents run out, and/or effective and affordable generics become available.
3. Disease
Prevention and Management:
AI can analyze
population health data to identify individuals at risk for certain diseases,
enabling proactive interventions and preventive measures. AI can also help
manage chronic diseases by providing personalized recommendations for lifestyle
changes and medication adherence.
On the other
hand, there seems to be limited ability and enthusiasm, if not resources, for
Caribbean Ministries of Health (+ Wellness)
to generate and analyze data on local populations, At a regional level, the
World Health Organization (WHO) and the Pan American Health Organization (PAHO)
Caribbean countries are linked with Latin American countries, but the
populations of tiny islands are dwarfed by the millions of persons in Central
and South America with whom we share little apart for geographic closeness. It
seems likely that Caribbean islands will need to develop Caribbean-specific AI
algorithms, but it is unknown whether any concrete steps have been taken or contemplated
in this regard.
4. Ethical
Considerations and Challenges:
While AI
offers tremendous potential, it is crucial to address ethical considerations
related to data privacy, bias in algorithms, and the potential for job
displacement. It is also important to ensure that AI is used responsibly and
ethically in healthcare, with proper oversight and regulation. Again this
is an area in which Caribbean specific algorithms may be needed. Clear
oversight and regulations would be needed to ensure ethical use of AI in
healthcare.
To most
independent observers, it would seem that many Caribbean islands are
experiencing a mental health pandemic, although this has not been officially
recognized. The triggers for this include both violence and drugs, often
interconnected, and harsh financial conditions which impact on employment
opportunities and even education. While AI is being used to develop tools to
assist in mental health diagnosis, treatment and ongoing support, and can
assist in scenarios where mental health professionals are in short supply, as
we have here, again Caribbean specific (or even island specific?) algorithms
will need to be developed before we can use these technology advances to
improve health care.
AI companions that combine conversational
AI, cognitive behavioral therapy (CBT), and mood tracking are used to support
users’ mental wellness. The best of them simulate real conversations,
prioritize user privacy, and deliver interventions grounded in psychological
research. However, they are not a substitute for a mental health therapist,
especially in severe cases.
But there is a possible downside to all of this new
technology. AI programs can be configured to
routinely answer health queries with false information that appears
authoritative, complete with fake citations from real medical journals. Without
better internal safeguards, widely used AI tools can be easily deployed to
churn out dangerous health misinformation at high volumes. If a technology is vulnerable to misuse, malicious
actors will inevitably attempt to exploit it - whether for financial gain or to
cause harm,
There are many areas in medicine where AI is
likely to make, and is already making, significant valuable contributions. These include areas like the development of
new drugs, where AI can analyze vast amounts of data at record speeds. The
discovery and production of a variety of vaccines against Covid-19, for
example, fall in this category. AI will be here when the next infectious
disease pandemic arises. AI
can automate some tasks, potentially alleviating the burden on healthcare
professionals and helping to address workforce shortages. Robotic
systems can assist surgeons with complex procedures, improving precision and
minimizing invasiveness. AI
can analyze medical images (like X-rays, CT Scans and MRIs) with greater
accuracy and speed than humans, aiding in earlier disease detection. By automating tasks and analyzing data, AI can
minimize human errors in diagnosis and treatment, leading to safer and more
effective care. By improving accuracy,
efficiency, and resource allocation, AI can help reduce healthcare costs and
address workforce shortages
Finally,
AI can be used to train healthcare professionals and enhance
patient education.
Like any other
tool, whether AI can be a blessing, or a curse, depends on the use to which it
is put, or if it is put to use at all.
There is no doubt that, in the ‘real world’, AI is being increasingly
used to improve global health. The question is, do we wish to advance Caribbean
health care services, and the health of our people? In the Caribbean we have
rapidly rising health costs, yet decreasing population health. The gap between the population’s need for health
care, and the ability of the health services to supply this care, seems to be
widening. At least AI seems to provide
one avenue for us to adopt a new business model: ‘business as usual’ is not
working. AI should be used to complement human intelligence, but is not likely
to be useful in situations where human intelligence seems to have run out of
ideas.
Comments
Post a Comment