Community Health Centers around the country are consistently testing Artificial Intelligence (AI) and tech tools for creative outreach. At the National Association of Community Health Center’s (NACHC’s) CHI and EXPO in August, we learned how San Ysidro Health successfully uses new AI and tech for patient engagement, care plan compliance, and Medicaid re-enrollment.
AI Increasing Patient Engagement at the San Ysidro Call Center
Vice President of Population Health at San Ysidro Health, Sonia Tucker, told us how her multi-disciplinary team researched different AI tools to support its call-in center as it struggled with higher call volumes and fewer agents. Sonia explained:
“In 2021 patient engagement was dropping, abandonment rates were rising, and patients were frustrated. We needed to do something about it.”
San Ysidro chose Kore.ai to deliver HealthAssist – an automated approach to call-center conversations for health providers. With the right training data from Uniform Data System (UDS), their Electronic Health Records, and language learning sources in English and Spanish, the San Ysidro team felt Kore.ai could help them reduce operational costs and improve call-center response rates, patient scheduling, and appointment reminders. They also used recorded call center data to identify themes and create a “rules engine” to train Kore.ai. The abandonment rate and number of unassigned patients have dropped substantially, and now they’re exploring how AI can help increase patient compliance. While there are “kinks” in these AI systems, overall, they’ve been happy with the ways AI supports their call center operations.
Since these practices are still new, Sonia left us with a few “AI pro tips”: “When training your AI, think carefully about your prompts– consider the AI like a young Intern you’re training.” She also said, “Hallucinations are a real problem. It’s critical to monitor AI communications and make immediate corrections, constantly.”
AI Tapped to Inspire Care Compliance
San Ysidro is also testing AI tools like ChatGPT for bi-directional text messages with patients. That tool is used for appointment reminders, follow-up conversations, redetermination campaign messages in all languages, surveys, and even one-on-one texting with the nursing team. San Ysidro staff are training the AI to create personalized messages for different service needs – using Azara to gather Social Drivers of Health – overlayed with zip code heat mapping to identify the location of patients they’d like to reach.
Ana Melgoza, Vice President of External Affairs at San Ysidro shared that, “tech is helping us provide information in different languages based on a better understanding of the reality of patients’ lives as we try to meet them where they are and help them with patient compliance”.
Take, for example, their “El Zapaton” program. This program was developed with incentives (shoe giveaways, food baskets, books, personal protective equipment, socks) based on analyzing patient data. The program’s aim is to improve quality measures for things like vaccine uptake, follow-up care, and medication adherence.
Hootsuite is a social media management tool, also powered by AI, that enables San Ysidro to post messages and short-form videos on multiple social media sites in multiple languages in support of the campaign and patient outreach. They can also reach patients directly through text messaging, call reminders, QR code promotions, digital media in waiting rooms, and provider support.
Traditional low-tech, in-person activities then clinched the deal with patients and helped San Ysidro develop trust and relationships. Atypical partnerships help as well. For example, many patients work late so Uber Eats became a partner; or they have pets, so local pet stores donated dog food or leashes. Favorite activities have included a “walk with doctors” event, cooking classes at the local university, and affinity groups.
Additional AI tools to help with Patient Outreach
Technology and AI tools that use generative learning are now ubiquitous, and health centers are researching different ways to use these tools. For clinical support, AI tools are advertised for ambient scribing, predicting no-show patterns, improving data quality research, helping with chronic care management and follow-up reminders, and even augmenting decisions for medical diagnoses.
For communications and outreach, this list of AI Tools for Communicators shares just a few of the many messaging, design, and outreach resources available.
Board members also want to understand “big data” and desire training to understand how AI could be useful for their health center.
What tools are you using?
While NACHC does not endorse any of the AI and Tech tools listed in this article, we are trying to understand how and why they’re being used by health centers. Please share what you’ve learned!
Send your stories to: firstname.lastname@example.org
NACHC recognizes that as AI resources expand, data stewardship, data privacy, and thoughtful data governance policies will be more essential than ever, and we’re working to support health centers. Health Information Technology – NACHC