Hc1 Uses Artificial Intelligence to Uncover Opioid Crisis Insights
As the opioid crisis continues to wreak havoc on the nation’s health and productivity, an Indianapolis-based startup called hc1 is applying artificial intelligence to a vast array of datasets in an attempt to uncover insights aimed at decreasing opioid misuse, abuse, and addiction.
Brad Bostic, CEO of hc1, describes his venture as a healthcare relationship management company, a term he coined in 2011, the same year he started the company. (The hc stands for “health cloud,” he adds.)
With the rapid growth of cloud storage technologies, Bostic founded hc1 to harness the abundance of siloed data at both the patient and provider levels and create holistic consumer profiles that could span providers, and thus improve care. So far, hc1 has amassed 90 million HIPAA-compliant consumer profiles and has more than 1,000 customers that subscribe to its customer-relationship and data-parsing services.
|Want more AI, machine learning, and drug discovery content? Network and engage with industry leaders online this June. Learn more.|
The current state of American healthcare, Bostic says, is “impersonal and appalling.” He experienced this firsthand as he watched his mother fight cancer in the late 1990s, and he doesn’t think much has changed since then. He says hc1 aims to enable healthcare organizations to deliver the “unified, personalized, superior service” that consumers have come to expect from other industries, such as retail.
“Even though my mother was receiving top-quality care at a renowned institution, the experience was awful,” he recalls, because his mother was treated as a number rather than a person. “Nobody knew who she was. Healthcare is not an industry that has done a great job with customer service, while Amazon treats the delivery of a teddy bear as if it’s a life or death situation. It seemed like a problem to be solved.”
Hc1 turned its attention to opioid data late last year after realizing the crisis was ripe for artificial intelligence-driven analysis, Bostic says, because the associated data tends to be broad, in multiple formats, and inconsistent. On hc1’s opioid dashboard, the company is partnering with drug-testing labs from around the country and examining their de-identified data in search of trends and patterns. (The labs conduct drug tests for employers, law enforcement, and pain management practices.) In the past, there has been available data on what’s being prescribed and overdose deaths, but Bostic says “the whole middle of the story” has been impossible to see.
“We use machine learning to generate a view into what’s happening across America,” Bostic says. “It’s an early warning system. We can see leading indicators that allow us to predict where to put resources to lower the misuse of opioids. Federal and state governments are deploying funds to reduce opioid use, but it’s very hard to measure impact. If we can look by county or ZIP code, it really helps to say whether an investment in a population is lowering misuse and abuse.”
In Indiana, where 1,100 drug overdose deaths resulted in more than $1.4 billion in medical costs and lost lifetime earnings in 2014 alone, the state’s Management Performance Hub (MPH) will work with hc1 to get real-time insights into opioid usage trends across the state and make proactive decisions about where to deploy resources.
Bostic says the MPH already had a warehouse of historical data; the role hc1 will play is “bringing in more live, leading indicators so they can inform other state agencies. We’re fortunate to have a progressive home state that understood the importance of this data and wanted to be a subscriber.” He expects other state and federal agencies will become hc1 customers in the future to measure the effectiveness of the healthcare dollars they’ve spent on managing the opioid crisis.
Ultimately, Bostic says, hc1 would like to be able to use machine learning not just to detect existing patterns, but also to predict future behavior—correlating, for example, the number of 911 calls with overdoses in a specific area.
“The forecasts won’t be 100 percent accurate, but it’s better than 0 percent,” he notes.