Getty Images

How Is Artificial Intelligence Shaping Patient Engagement Tools?

Artificial intelligence is finding its place with patient engagement tools, helping to personalize an efficient patient experience.

Although patient engagement principles are rooted in timeless concepts like good interpersonal skills and patient motivation, the technologies used to push engagement are indeed at the cutting edge. Most recently, healthcare organizations have tapped patient engagement tools that leverage artificial intelligence (AI) to drive a personalized and convenient healthcare experience.

As the medical industry continues to embrace the patient as a healthcare consumer, providers and technology developers alike have sought to create a personalized healthcare experience. AI has proven to fill that gap, helping to make sense of patient data and needs and generate action items that can ideally activate the patient.

And patients are primed for this new type of technology, research has shown. In 2018, Accenture’s Consumer Survey on Digital Health showed that patients are welcoming of AI technology in healthcare. And again in 2019, patients expressed a strong interest in AI for helping to manage their own health and drive convenient care.

Healthcare organizations are beginning to respond to that need, implementing tools that leverage AI, machine learning, and natural language processing (NLP) to aid in not just clinical operations, but in patient engagement efforts, as well.

Using chatbots for automated communication

Chatbots have begun to emerge as essential tools for providing convenient access to care. As in other service sectors, healthcare consumers want medicine that is easy to access, seamless, and frictionless. They want it to be easy to book an appointment, simple to get to their provider offices, and intuitive to engage in and manage their own care.

Except healthcare has proven anything but simple.

Medical providers have begun to adopt AI chatbots to begin to streamline the patient experience. Healthcare organizations want to create the same experience consumers have in other service sectors, even if that feels next to impossible.

“I've often said, we need that Disney-like experience,” Kevin Pawl, MS, FACMPE, senior director of Patient Access at Boston Children’s Hospital, said in a previous interview.

Patients want to access their medical care quickly and easily, have all of the relevant information for getting to the hospital, and have a good interaction with their providers all in one.

“You're not calling a hospital with the joy of going on vacation or going on a trip or staying in a hotel,” Pawl acknowledged. “But it would be awful nice if the service and the experience were similar. That's what we were trying to achieve as we tap into artificial intelligence.”

AI chatbots work on the frontlines of the patient experience. For Pawl, and many other organizations, chatbots help filter out patient phone calls and refer patients to relevant resources. A patient might call the hospital looking for parking information, and instead of being placed on hold for several minutes to an hour, they will get the answer right from a chatbot.

More recently, healthcare organizations and technology developers have put AI chatbots to work to sift through patients’ clinical needs. The chatbot might field some patient symptoms and produce recommendations for care.

Other organizations, like Intermountain Healthcare, utilized AI in a similar fashion as a COVID-19 symptom checker.

Simply installing a chatbot isn’t a patient experience cure-all, however. Patients may get fed-up with the technology, which could actually have a negative impact on satisfaction, most experts agree.

In 2020, as more organizations began deploying the tools as part of their COVID-19 hotlines, evidence emerged that chatbots that seemed human-like, competent, and referred patients to resources that actually answered their questions were rated more highly.

Organizations should thus be judicious about which AI chatbots they implement. Additionally, they should test which functions are well-served by a chatbot and which require human interaction. This may vary based on typical call center requests, community needs, and patient demographics.

Tapping risk assessments for patient outreach

Similar to the artificial intelligence used as part of symptom checkers, the tools can also work for improving patient outreach. Healthcare organizations conduct patient outreach based on health data and key risk factors. Patients who fit certain risk factors may warrant outreach about chronic disease management plans or referrals to social determinants of health services.

But again, sifting through that patient data can prove a daunting task, leaving room for AI. Some organizations are deploying the technology to first risk stratify patients and then push out automated patient outreach messages either via text message, email, telephone, or a combination of the three.

CommonSpirit Health, for example, recently adopted AI to identify high-risk patients, predict their healthcare needs, and push out targeted text message outreach. The tool from Docent Health will help address medical needs outside the four walls of the hospital, according to CommonSpirit’s System Vice President of Population Health and Innovation Policy Alisahah Cole, MD.

“It initially started out as a pilot in one of our markets that had a pretty significant amount of patients with Medicaid as well as Medicare, and was really looking at how we help patients navigate through different clinical episodes of care,” Cole previously told PatientEngagementHIT.

“If someone was coming in for a right knee replacement or somebody was pregnant and having a baby those are considered episodes of care. So how do we help patients navigate through not just our own healthcare system, but also through all of the resources that they may need in order to have a successful clinical outcome?”

In Illinois, similar AI technology helped Medical Home Network, a managed care organization, identify members who were high-risk for COVID-19 infection and refer them to testing sources.

For both organizations, understanding the right risk factors, medium for patient outreach, and messaging proved crucial. According to Cole, text message outreach was preferable because most Americans own a cell phone, although not everyone has a phone that lets them download apps.

Additionally, offering patient outreach in multiple languages and incorporating considerations for social determinants of health also made patient messaging more actionable for both organizations.

Utilizing NLP for patient health literacy, patient portal use

As more patients gain access to their patient portals and clinical notes—something mandated by the 21st Century Cures Act and going into effect come April—medical providers have expressed concern about patients understanding that data.

Those concerns are not entirely unfounded. Patient health literacy is generally low, so it might be hard for a patient to see her clinical notes and know exactly what they mean. Natural language processing, a form of artificial intelligence, may be able to fill in that gap.

In 2017, researchers wrote in the Journal of Medical Internet Research indicated that NLP could sift through the medical language in a clinical note or patient portal and give a patient a layperson translation.

In a follow-up 2018 report, researchers found that NLP programs need to offer a good display for patients and improve the way they define medical terms that vary depending on the context. Additionally, NLP systems need to account for clinicians who use abbreviations, misspell words, or have illegible handwriting.

As medicine becomes more digitized—and as it answers calls for more personalization—AI and related technologies will become more important. This technology allows providers to not just understand more about a patient’s health, but to target interventions and outreach to drive patient engagement.

Dig Deeper on Patient data access

xtelligent Virtual Healthcare
xtelligent Rev Cycle Management
xtelligent Healthtech Analytics
Close