Using AI to Improve the Patient Experience
Think about everything that influences your health: your job, where you live, how you eat, etc. You and your health are unique and special. Now imagine the challenge of caring for 85 million unique people, each with their own set of expectations, perceptions, and circumstances. That’s the challenge that we handed our enterprise data science team two years ago.
The challenge was steep, since thousands of members use our broad services in support of their unique healthcare needs every day. Additionally, they share their experience with us through multiple channels like email and social media, to name a few. Using innovation and advanced analytics, we have been able to better understand their unique experiences in order to learn and ultimately transform how we deliver service and provide the best possible experience.
It’s through an investment of $2 million over four years in artificial intelligence and, specifically, natural language processing and machine learning, that we’ve been able to quickly improve our members’ experience and zero in on what really matters.
To elevate our service capabilities, our enterprise data science team built a proprietary tool – Express Scripts PULSE, that uses natural language processing to digitize and organize unstructured data from multiple sources — emails, website feedback, phone calls with our pharmacists and customer service agents, physician notes in electronic health records, even social media engagement with the company. With it, we convert all of these interactions with members and providers into consumable data sets and then apply algorithms to assign sentiment, reveal member preferences and create solutions for any problems that arise.
The team that actively uses PULSE is able to address both broad themes and the needs of individual members in real-time. Some of the issues that we’ve been able to uncover and resolve are:
- A briefly unhooked shopping cart – A coding error unhooked the shopping cart from the Express Scripts website, preventing members from ordering new medications or refills. Through proactive listening, we were able to quickly find and correct the error, instead of waiting until a larger number of members couldn't get their prescriptions.
- A hard-to-find logout button – By sorting through member feedback and finding that they were having issues logging out, we were able to redesign the website to make it more intuitive.
- Problematic claims – Claims that needed to be adjusted and corrected were identified and proactively remediated, before causing a nuisance for the member
PULSE is also improving the member experience by allowing us to engage with members according to their contact preferences. For example, if they reach out to us on social media, we’re consistent in interacting with them there. If they call us, we respond with a call, etc. In the future, by using this tool, we’ll have built knowledge of how individual members’ prefer to communicate, which will allow us to train models on how to engage with patients for clinical needs as well.
Artificial intelligence continues to evolve and we continue to learn more about our members and how to meet their needs, not just as a population but as a unique individual. It is a powerful tool that’s helping our members and leading to better outcomes for patients in ways that untamed feedback from 85 million members wouldn’t.