Artificial intelligence — or AI — is quickly affecting every facet of our lives.
Verily, a health-care-focused Google partner, is using AI to screen for diabetic eye disease in India, thwarting preventable cases of blindness and improving quality of life for underserved populations abroad.
Tech giant IBM is leveraging the tool to improve personnel retention and determine — with stunning accuracy — who could be planning to leave the job.
Breweries are even using it to enhance beer production.
It’s no small wonder that AI pioneer and venture capitalist Kai-Fu Lee told “60 Minutes” earlier this year this technology could “change the world more than anything in the history of mankind. More than electricity.”
What if we could bring this same disruptive force to bear in the fight against the devastating opioid epidemic, which continues to claim 130 lives each day in the United States alone?
Earlier this year, the Journal of the American Medical Association released findings showing that machine learning — a type of AI — is actually better at predicting patients’ risk of an opioid overdose than other methods currently in use.
The study showed that machine-learning-based algorithms could better avoid targeting patients who were not truly at risk of misuse and concluded that this technology “appear(s) to be a valuable and feasible tool” in the fight against this deadly public health crisis.
That is the same idea behind one of our chief policy goals at the Opioid Safety Alliance: something we call a “Prescription Safety Alert System.” The idea is simple:
Today, when doctors and pharmacists are writing or filling a prescription for opioid medications, they lack a complete, well-working tool to make sure their patient has not received the medication elsewhere in recent days.
States have relegated clinicians to Prescription Drug Monitoring Programs — or PDMPs — to check for past opioid fills. While an important tool, they are far from perfect. PDMPs often operate outside of clinicians’ workflow, are not updated in real-time, and often do not work across state lines, among other issues.
In 2019, facing an emergency of this caliber, we must do more.
Our Prescription Safety Alert System, delineated in bipartisan legislation called the ALERT Act of 2019, complements the work of PDMPs by leveraging the power of AI to stem this awful tide of opioid misuse.
Introduced by Congresswoman Annie Kuster, D-New Hampshire, and Congressman Markwayne Mullin, R-Oklahoma, the bill directs the Trump administration to contract with the private sector to establish an alert system populated with existing prescription and dispensing data — this is where AI comes in to play.
Each time an opioid is dispensed, the system would review the patient’s prescription history and develop a clinical risk score that can inform dispensers when a patient could be pharmacy shopping or otherwise at risk of harm — all before the medication crosses the pharmacy counter.
The input is human-led, the output is true AI; using 21st-century technology to analyze patients’ opioid histories without aid and determine their risk of misuse in real time.
Importantly, this data would be made available only for clinical use and would be subject to HIPAA protections, thereby upholding patients’ privacy. We have proposed that patient privacy could be further strengthened by applying blockchain technology to this alert system, ensuring that sensitive information remains secure.
The federal government has devoted nearly $11 billion to the opioid epidemic in the last two years alone. Americans can agree this is a worthy cause but, to date, much of this funding is allocated for resources that reach patients after opioid misuse takes its toll.
Our Prescription Safety Alert System — and the ALERT Act that would bring it to fruition — treats this crisis at its source; allowing for a clinical intervention before an inappropriate fill of opioid medication ever crosses the pharmacy counter.
This is about much more than showcasing recent technology. It is about saving human lives. At the Opioid Safety Alliance, we believe that’s a good use for AI if there ever was one, and we’re betting Congress feels the same.