How AI can improve patient experience and patient engagement

Hospitals and health systems, especially larger ones, are using AI in myriad new ways – including leveraging it for improved patient experience, and for better patient engagement with the goal of reducing health inequities.

Tom Kiseau, digital leader and chief innovation officer at Chartis, hosts a regular series of AI roundtables alongside colleagues and experts from other practices in the industry. We spoke with him about what he’s learning from these conversations.

He offered perspective on how physicians can use AI models to translate complex jargon and concepts into a message that’s easy to understand, how AI-tailored engagement can enable patients to better adhere to their care plans and medications and how AI can be applied to facilitate the critical role of genetics and family history, identifying information that may already be stored in a hospital’s system.

Q. How did you originally see AI being a fit for improving the patient experience?

A. Much of what degrades today’s patient experience is caused by friction across the journey and the associated patient confusion related to next steps and communication. From finding the appropriate and accessible provider to forging a trusting and attentive personal relationship with patients during and after their clinical encounter, healthcare providers make it difficult for the patient to know when and where it’s appropriate to reach out.

From the patients’ perspective, it’s pretty simple. They want an experience that is accessible, responsive and personalized to them and their needs. Artificial intelligence, with its ability to understand and synthesize, automate operational processes, and predict what will/should come next, is an increasingly powerful tool to help minimize friction and the associated patient frustrations along the patient journey.

For instance, AI is making online scheduling more meaningful for the patient by aggregating past patient preferences, current position in their broader care journey, provider specialty relevance, appointment availability and geography data.

Together, these benefits optimize the patient’s decision about accessing the most appropriate and convenient care for them. During the visit, AI enables a more engaged patient-provider experience by using ambient listening tools to transcribe the clinical encounter and translate that discussion into structured notes captured into the electronic health record, presenting the opportunity to automate follow-ups and next steps, such as referrals and future appointment scheduling.

On an ongoing basis, AI is also being used to better customize care plans and support resources for each patient’s unique and constantly evolving health needs.

For example, AI is increasingly being used by provider organizations to develop personalized risk profiles for certain diseases based on family history, genetic and biomarker mapping, social influences, as well as activities and behaviors that may have emerged since the interaction with the care team/provider.

Harnessing the power of broad population health data to uncover new and more accurate predictors of future health risks enables patients to have a more personalized view into their health status and become more engaged and empowered to follow a customized care plan that’s tailored to them and their unique needs and preferences.

Q. You suggest that to help improve the patient experience, physicians can use AI to translate complex jargon and concepts into a message that’s easy to understand. Please elaborate on how this would work, and the outcomes.

A. Artificial intelligence is a powerful language tool for providers to use to more effectively engage their patients through written communications that are provided to the patient through any growing number of channels, both digital and physical (for example, text message, portal, hard copy post-appointment plan, etc.).

Patients bring diverse levels of understanding across numerous dimensions, including language, clinical expertise, generalized interest and overall reading comprehension. The ability to adapt content and communications can be truly transformative to fostering greater patient understanding, engagement and, ultimately, adherence to their care plans over time.

Recent versions of generative AI, such as ChatGPT, have clearly shown the ability to modify a given set of content to a reading level that is most appropriate for any individual. This allows providers to efficiently capture their notes with the technical jargon that they are trained and most comfortable using, and it allows the AI to take that jargon and translate it to more digestible language for the patient’s unique needs.

Not only does this allow the physician to be far more efficient, but the end product received by the patient is also generally perceived to be better quality and more empathetic.

AI also is driving the translation of medical jargon into different languages, with increasing accuracy in the nuances of expressions and dialects that makes the communication more accessible for patients. In a recent systematic review, 75% of studies showed that patients with limited English proficiency experienced better outcomes when they received care plans in their native languages.

In another recent study, 91% of Spanish-speaking parents preferred to receive translated copies of their children’s written discharge instructions to improve their understanding and adherence to those plans. AI is important to the success of these translation services given the complexity of medical terms to accurately translate into another language, and to layer in the ability to moderate the communication to varying and often simpler reading levels for patient comprehension.

Generative AI tools also are demonstrating the ability to summarize vast amounts of information to the foremost points, allowing a multi-page discharge plan, that many patients will be unable to read and comprehend, to be more understandable within a paragraph or two.

Together, these capabilities make it more likely for the patient to understand their healthcare situation, the actions they need to undertake, the rationale for these actions, and next steps to expect from their care team.

Q. You further suggest that AI-tailored engagement can enable patients to better adhere to their care plans and medications, and help health systems more proactively reach out to patients to address their care needs. Please discuss this point and how it translates into patient experience wins.

A. Beyond the operational and language-based capabilities I previously discussed, AI is augmenting the reach of provider organizations to accurately segment cohorts of patients based not only on their health needs, but also on their preferences to best manage their health.

As health systems expand the number and type of access points for care and engagement, greater attention must be paid to the access points individual patients find the most meaningful to them. AI can be an important tool to inform these decisions.

AI is increasingly valuable for identifying and supporting proactive outreach opportunities, engaging patients in a discussion about their current health status or sending customized reminders about preventive care or recommended next best (clinical) actions.

As healthcare providers increasingly shift toward a value-based orientation, these triggers are critically important to creating an “always on” care environment to ensure patients don’t fall through the cracks or “get lost” in the traditional reactive healthcare system. For example, AI tools can help boost the number of attributed lives through primary care annual wellness visits, particularly among those that haven’t engaged with them in more than a year.

This creates revenue growth for patient populations that (based on their lack of attributable use of healthcare) are often relatively healthy, and therefore high-value patients. Left unengaged, these patients are at-risk for attribution with other providers, and, more importantly, the lack of appropriate preventative care introduces greater risk of unnecessary utilization and, worse, overall patient health status.

Many of these organizations are using AI to not only identify these patients but proactively reach out to them through preferred communication channels and encourage them to schedule their annual wellness visits – typically through streamlined, AI-driven online scheduling options.

Given the fact that only a quarter of Medicare beneficiaries participate in their annual wellness visit each year, and that better health outcomes at lower total costs are the likely result of those visits over the years that follow, proactive engagement is an imperative for any organization to succeed in value-based care.

Recognizing the financial infeasibility of proactively engaging patients using traditional methods, leveraging advanced AI tools and automation will be essential to conducting these initiatives at scale.

Beyond supporting proactive engagement, AI is a critical tool for triaging and responding to patients as they heed the calls to action from the impending, future deluge of proactive healthcare outreach.

Today, a vast majority of health systems, can’t quantify the inbound volume they receive from patients across their myriad channels (phone, web form, portal, text message, email, etc.), let alone how quickly and effectively those patient inquiries are reviewed and addressed, provided they’re reviewed and addressed at all. AI tools are increasingly valuable in ingesting, triaging, stratifying, recommending responses, and, in some cases, responding to certain inbound patient communications.

While powerful on their own, when integrated into an aligned human-mediated operating model, the results can be extraordinary in terms of staff, care team and patient satisfaction.

What’s fascinating is that while many leaders are quick to point out the risks of AI tools “autonomously” responding to patients (an approach we would only suggest in very discrete, defined, low-risk situations versus using the tools to augment and support responses), they’re perfectly happy to operate in a status quo where, in our experience, a majority of the inbound patient messages (for example, non-live agent handled interactions) receive no response.

Q. Finally, you say AI can be applied to facilitate the critical role of genetics and family history, identifying information that may already be stored in a hospital’s system. How would this work?

A. Perhaps one of the foremost trends in healthcare that underscores the importance of the AI revolution is the explosion of big data, particularly in the genetic space. The typical whole genome sequence file can easily top 100GB in size, and while precision medicine applications don’t often require the entire genome to be sequenced for clinical interpretation, it demonstrates the sheer size of this data.

AI is uniquely positioned to process large data sizes like genetic biobanks and uncover previously unknown links between biomarkers as precursors for certain diseases.

Where AI is having the most impact is in data mining to identify clinically relevant patient cohorts and to more efficiently match patients with ongoing clinical trials. On the data mining front, AI is uncovering links between patients with similar genetic profiles and mutations and drawing more accurate conclusions about what treatments are most likely to be beneficial given their biomarker commonalities.

For clinical trials, AI is also scanning patient records within their EHRs to find patients that are most likely to be viable candidates for certain trials and possibly benefit from those innovations.

Where the real power of AI resides, however, is layering in genetic data with other, diverse forms of data that healthcare organizations aren’t accustomed to analyzing – but are becoming increasingly relevant indicators of their patients’ health.

Social influences, such as poverty rates, food deserts or environmental pollution, can all adversely impact a patient’s ability to manage their health issues. AI can quickly ingest, organize, analyze and implicate those data points for providers to better understand and plan for those unique challenges faced by each patient.

Follow Bill’s HIT coverage on LinkedIn: Bill Siwicki
Email him: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.