Artificial intelligence (AI) knows nothing about the social determinants of health. More specifically, those environmental and behavioral factors outside a patient’s control, such as adequate housing, affordable nutritious food, sustainable employment, cannot be relegated to computerized thinking. Human interaction cannot be taken out of the equation, and the narrative – the patient’s unique story and circumstances – cannot be left to computer algorithms to decipher.
I drew this conclusion and learned about the burgeoning field of artificial intelligence and its impact on healthcare at the 6th Annual Hospital Innovation and Lean Network of New York conference held at Lenox Hill Hospital in New York City November 7 – 8. Knowing very little about artificial intelligence, this was an opportunity to get a good grasp on this topic.
An afternoon session covering artificial intelligence in population health improvement was my first foray into this foreign realm. I am a people person with a penchant for listening to people’s stories and relaying their unique perspectives on life, health, and everything in between. So when a slide displaying dozens of disease conditions to be fed into a computer program to predict the likelihood of 30-day readmission and mortality of patients with acute myocardial infarction (IMF) – heart attack – popped up, my neural pathways short-circuited.
The presenter was super smart, so I strained to listen and learn and find a way to apply what I was hearing to my role as a health communicator. In this role, my goal is to take complex topics – such as artificial intelligence – and distill them into laymen’s terms. I concluded that AI in healthcare does and will play a significant role in predictive analytics, and this will lead to improved care for patients and the ability for healthcare providers to fine tune treatment. Will AI replace some tasks now performed by healthcare providers? I am less sure of that. What I do know is that unless a patient’s story is fully revealed, we cannot effectively treat their conditions. No measure of machine learning can reveal whether a patient can afford their medication, whether they live in a home riddled with mold, or whether violence is present in the home. These are examples of the many struggles that patients face. Each is a barrier to good health.
We now know that these social determinants of health have a bigger impact on a person’s health than clinical care. Eighty percent of an individual’s health is tied to their physical environment, social determinants – where one lives, works, and plays – and behavioral factors – including exercise and smoking. The patient’s narrative must be told and heard in order to provide the most effective and cost-effective care. Hospitals, health plans, large physician practices, and many community-based organizations now incorporate social determinant of health interventions into treatment and prevention programs. These address the many barriers to care that are often hidden in the patient’s narrative.