September 27, 2020

Conversational AI on the Frontlines of COVID-19 with Montefiore Health System

Aaron Bours

Head of Marketing at Hyro

By mid-March, 2020, the world’s attention — up until that point fixed on a devastated northern Italy — shifted at once to New York which had rapidly become the globe’s new coronavirus epicenter. As the state’s some 20 million citizens hunkered down in their homes, adhering to strict social distancing measures, they turned en masse to their healthcare providers’ digital channels, creating crippling backlogs in medical care and access.

Image for post

Jeff Short, VP and Chief of Staff of Montefiore Health System — which serves hundreds of thousands of patients from the Bronx to New Rochelle — describes both the incredible challenge Montefiore faced at the time and the solution they were hoping to find:

Image for post

That technology was Hyro’s COVID-19 Virtual Assistant, a complimentary conversational AI assistant created to redirect and reduce the pressure on overloaded healthcare professionals and support centers. By scraping and ingesting data provided by Montefiore as well as from freely developed resources by the CDC and WHO, the COVID-19 Virtual Assistant has been able to answer real-time patient queries regarding COVID-19 and guide patients — using Natural Language Understanding (NLU) capabilities — to their requested information or service.

Moreover, the COVID-19 Virtual Assistant provided Montefiore’s patients with a self-risk assessment and screening tool, enabling them to ascertain their current level of risk. By leading patients through a flow of questions regarding their symptoms, recent travel, and possible contact with confirmed cases, the tool produced initial risk assessments ranging from low to high alongside recommendations on how to best proceed.

Adrin Mammen, Associate Vice President, Patient Access Transformation Officer at Montefiore Health System, writes about the immediate results Montefiore experienced upon deployment:

“during the first week of deployment, we were already witnessing hundreds of daily conversations with the COVID-19 screening tool and chat solution.”

Hyro and Montefiore combed through thousands of patient interactions with the COVID-19 Virtual Assistant in its first three months of deployment; March-May, 2020

Here’s a snippet of what we found:

Image for post

Overall, 44% of all patient interactions with the Virtual Assistant were directly related to COVID-19, versus 56% pertaining to other medical services and information such as appointment scheduling, Rx management, and billing. Through patient engagement and interaction with the COVID-19 Virtual Assistant, we soon realized that patients were searching for all manner of services on Montefiore’s website. We promptly responded by identifying and accommodating these key patient demands within the virtual assistant.

“We have been able to direct chatbot users to other key areas on the website, including our FAQ document and our patient portal application” — Adrin Mammen commenting on Hyro’s swift response to burgeoning patient exigencies.

Image for post

Of all COVID-19 related conversations, 41% revolved around the topic of screening and self-risk assessment. Unsurprisingly the majority of patients at 45.9% reported having a cough, perhaps the best-known coronavirus symptom, and a highly common symptom of the seasonal flu, often misleading patients into unwarranted concern. The second most pressing topic of interest for patients was “tests” at 23%.

The following case study demonstrates the potential of conversational AI in action, as a dynamic, laser-focused approach to the complex variables offered by massive-scale crises. By moving fast and cooperating on every turn, Hyro’s partnership with Montefiore gave birth to a COVID-19 Virtual Assistant specifically tailored to their unique patients’ needs.

Image for post

Download the full (free) case study to find out how Montefiore’s patients engaged with the Hyro COVID-19 Virtual Assistant at the height of the pandemic in New York. Further findings in this case study include:

  • Obstacles in deploying typical conversational interfaces
  • Mobile vs Desktop? Which medium did patients prefer to converse with?
  • FAQs, top COVID-19 terms and topics, and other key patients insights
  • Full segmentation of Risk Assessment reports, according to age groups, symptoms, and degree of risk

DOWNLOAD THE FULL CASE STUDY NOW

As always, we remain committed to innovating and crafting solutions for today’s healthcare, through the lens of tomorrow’s problems.

Visit our website to learn more and follow along on Linkedin and Twitter for updates on our latest solutions and advancements.