The promise of AI is unlimited in healthcare, but what can we actually do today?

October 3, 2019

As the hype around artificial intelligence begins to cool, the focus has shifted to "assistive intelligence" as providers realize the value of a helping hand

Healthcare organizations of all sizes continue to invest in new technologies to help optimize their businesses, provide new patient access avenues, and create opportunities for providers to better support patients. The field of artificial intelligence (AI) has been thrust into the limelight as technology that will magically save us from tedium, parse massive amounts of complex information to find answers better than humans, and revolutionize the field of healthcare. However, as with all new technologies, the white-hot hype of AI is starting to cool; the bigger opportunity for health systems is around more “assistive intelligence,” the “AI” that’s already providing value for various stakeholders without losing the human touch that’s central to the delivery of care.

Connecting patients with the right care is a prime example of an area where this type of AI is having an impact today and provider data management is foundational to it. Data management is an ongoing area of focus for healthcare organizations, whether it’s centralizing patient information due to a merger for an EMPI; aggregating lab results, problems, and diagnoses into a clinical data repository; or leveraging tools to turn semi-structured clinical notes into discrete pieces of information for further processing, IT departments are spending time and money wrangling vast amounts of data. While patient data has been the main focus historically, provider data is equally important for enabling effective patient routing though.

To this end, while we wait for the magic of artificial intelligence to offer standalone interpretations and diagnoses, other types of data, including critical provider information, is often left unmined and underleveraged, resulting in missed opportunities for greater efficiency and quality. For example, when it comes to patient access, organizations are placing a growing emphasis on delivering an interface for almost every digital means a patient could use to find providers – be it voice, text, mobile or a traditional computing device – but collecting provider data in a way that serves only this channel is insufficient. The reality is that 58% of patients report still preferring to book appointments by phone.

In this key access channel, call center agents, front desk staff, and care coordinators are attempting to navigate ever-changing provider populations, scheduling rules, and patient preferences, alongside growing service lines. By applying the same rigor of data management to the information these users require, we can leverage assistive intelligence to better support them, and the patient, in finding the right care. Related assistive tools that harness detailed and actionable provider data can help agents match patients effectively based not only on their unique clinical needs but also on their specific preferences (e.g., location, appointment availability). Such tools can allow agents to spend less time looking up information and more time ensuring the patient understands their options and pathways to care. Furthermore, with customer relationship management platforms taking hold in healthcare, there’s a new opportunity to fine-tune patient routing by marrying rich patient data with detailed provider profiles.

Building on this, one approach to assisting patients that is growing in popularity is offering chatbots. Today, these are mainly simple user experiences, often enabling extended navigation support on a website or basic clinical triaging. However, the potential of pairing chatbots with a modern telephony system, CRM system, and provider data management platform offers opportunities for new types of assistive tools to allow patients to complete more of their access journeys via chatbot. Chatbots also have the potential to facilitate a contextual “warm handoff” to an access center when needed, where an agent can pick up the conversation and bring in their own knowledge to direct the patient more seamlessly.

We’re all excited about the true future for artificial intelligence, but it requires deep investment into data, models, and approaches that require years of fine-tuning to effectively replace humans - if that’s even a goal we want to strive for in healthcare. In the meantime, focusing on ways to curate robust provider information and make it actionable in assistive tools can go a long way in giving superpowers to those engaging with patients each day, ultimately helping drive better experiences and outcomes. 

Chris Gervais is chief technology officer for Kyruus.