AI and cloud technology help fight opioid addiction
Fighting the opioid epidemic is my mission. More than 115 Americans die each day from opioid overdose.
The misuse of and addiction to opioids—which includes prescription pain relievers, heroin and synthetic opioids—is a serious crisis in the United States and beyond, affecting public health, societal well-being and economic welfare. According to the U.S. Centers for Disease Control and Prevention, the total economic burden of prescription opioid misuse in the United States alone is $78.5 billion a year.
With overdose deaths at epidemic levels, the field of addiction medicine and recovery has become fragmented and uncoordinated. Doctors, therapists, law enforcement officials, recovery coaches, sponsors and families cannot keep up. Consequently, people in dire need are falling through the cracks.
While discussing this crisis over lunch one day in 2012, TryCycle co-Founder Ken House and I decided something new needed to be done to break the cycle of addiction.
Bringing our idea to fruition
Ken and I concluded that recovery occurs within a network of relationships, and that with the right technology, the network could be activated and coordinated to magnify impact, reduce economic burden and better support the needs of both the patient and their providers.
The system we envisioned would create a relationship between a person in recovery and their treatment team and use algorithms to evaluate human input and behavioral data to predict the risk of relapse. With these real-time analytics, healthcare providers could then make more informed decisions for those at risk.
We started working with a software company to build a solution… but after two years we’d made only nominal progress. Concerned about slow development, we changed tacks and started working with IBM. Once we partnered with IBM, we made more progress in two months than the previous two years.
Using AI and cloud to help with recovery
Our system, TryCycle, is an innovative technology designed to incite behavior change for a person in recovery. Using a mobile device, the person in recovery receives regular prompts through push notifications to submit journal entries by answering a set of predefined questions on how they’re doing. Their input is processed using an algorithm and shared with their treatment clinic.
In addition to this manual input, we have the person in recovery enter audio responses to assess sentiment. Together, this active data—analyzed by IBM Watson Natural Language Understanding, Tone Analyzer and Watson Studio—helps identify in real time the possible risk of relapse for each person.
Passive data is also gathered from the mobile device to complement active input. For example, the phone’s accelerometer can show if the person has stayed isolated at home or moved about during the day.
Data is double blind at the client level, making the system less obtrusive and therefore more likely to be used. The results are presented in a dashboard, providing continuous monitoring and allowing for human-based decision making.
Expanding application—from individual to group and other social concerns
While our system focuses on recovery at the individual level, as more people use TryCycle, insights can be extrapolated from the aggregate. Predictive analytics and machine learning will help us understand human performance and behavior, enabling us to learn our way through the opioid epidemic.
Additionally, this concept can potentially be applied to other social concerns, aiding those who might be suicidal or affected by post-traumatic stress disorder (PTSD), or helping to promote healthy habits.
While our system does not prevent first use and cannot cure the disease of addiction, it can help people committed to recovery. It’s critical that therapists and prescribers know when their patient’s triggers are firing. In moments when emotional or physical triggers begin to fire, and when cravings feel insurmountable, TryCycle can be a lifeline.