MyndYou: Helping patients get better care faster with its AI care assistant
Delivers accurate AI solutions in multiple languages
Handles thousands of patient calls a day
Doubles success rate of traditional clinician-led outreach
Expands ability to monitor patients proactively and intervene early
Protects patient data with HIPAA-compliant large language model
With Google Cloud, MyndYou gained the services to handle automated patient calls smartly and in real time, making healthcare more efficient and improving patient outcomes.
With Google Cloud, MyndYou gained the services to handle automated patient calls smartly and in real time, making healthcare more efficient and improving patient outcomes.
With six in 10 adults in the US currently living with chronic diseases, and a nationwide shortage of up to 450,000 nurses predicted by 2025, the staffing gap is now one of the most acute challenges facing US healthcare. It is that gap that MyndYou is helping to close, with its AI-powered virtual care assistant—Eleanor. Trained on thousands of clinical conversations between patients and clinicians, Eleanor can conduct complex clinical phone conversations with patients in their homes, enabling healthcare providers to manage thousands of patients at a time.
Eleanor not only listens to what the patient is saying, but how they are saying it, with the ability to identify clinically relevant events within that conversation. Eleanor ranks this information by urgency and relays it to the care team, enabling them to prioritize their caseload and ensure each patient receives the appropriate level of care.
The way Eleanor has been built with Google Cloud, it’s very easy for us to expand to new clinical use cases. “She” doesn’t get tired and doesn’t need a lunch break, so she can help clinicians reach more patients and do more.
Itay Baruchi, PhD
CTO and Founder, MyndYou
Solving the challenge of real-time speech transcription
For Eleanor to deliver clinically valid information based on patient conversations, it’s important to have conversational assistance that is tuned to the specific use cases for which she was built, and for that, we use Gemini and Google Cloud.
“We’ve done work with Gemini where we bring our data and fine-tune the model so that when patients answer to Eleanor, Gemini returns an accurate value,” says Itay Baruchi, Chief Technical Officer at MyndYou. “For example, based on the American Cancer Society’s screening assessment recommendations, patients' answers regarding their smoking habits can prompt Gemini to determine whether they need to have a lung cancer assessment or not. They may answer in many different ways about number of cigarettes or number of packs smoked a day. We train and fine-tune the results so Gemini can help us.”
Eleanor is built to understand responses, parsed from the many ways a patient may answer, and engage in real conversations. She has pleasurable interactions with patients and that’s critical to her success.
The performance of Google Cloud is great. We use Vertex AI to train and serve models using multiple APIs and with different languages. It’s a very stable service, and the response times are within our requirements, which is extremely important for us.
Noy Carmel
Vice President of Product Management
Making smart decisions to optimize outreach
The opportunity to complement patient care with Eleanor extends beyond using technology to complete conversations. MyndYou is using Google Cloud’s BigQuery to determine the best time to reach patients.
“In order to reach patients at the right time, we are using large-scale analytics to analyze data and identify when is the best time to call based on a variety of parameters including age, gender, and geography,” says Baruchi. Based on the optimizations, Eleanor will shift her call times to maximize the reach rate.
For one program, that means Eleanor is better able to connect patients who have just left the hospital to a nurse, and that nurse can ensure the patient knows their medication plan after leaving the hospital.
With BigQuery, our analysts can optimize the use of our data to make strategic decisions that influence patient outcomes and shape and grow our business.
Itay Baruchi
CTO and Founder, MyndYou
The faster path to better patient care
A key outcome of these inputs is that Eleanor is able to build trust with patients, making them more likely to share the information their care team requires to manage their needs effectively.
Engagement rates with Eleanor are significantly higher than with traditional calls. Not only does Eleanor reach more patients through call optimization and the ability to make hundreds of calls at a time, which would be a significant time drain and expensive endeavor for clinician call operators, but her ability to get patients’ information leads to better health outcomes too.
Tali Segev, VP of Customer Success, cites the example of a customer who was using clinicians to call patients for colonoscopy appointments. “The patient navigators were sitting in an office calling patients and couldn’t get hold of them. And when they did, they didn’t get the action they needed. With Eleanor, we reached patients and we were able to ask them about their barriers to completing their screenings. We doubled their success rate, which means more people coming in for colonoscopies, and we freed up time for the clinicians to do other patient care jobs.”
With another customer, Eleanor performs regular check-ins with high-risk patients who agree to ongoing monitoring – at a frequency they couldn’t support with human staffing. If Eleanor identifies an urgent need, she can transfer a call to a care manager.
Baruchi shares, “Our customers share that their patients are appreciative when Eleanor learns they’re sick or having problems with their medications and can connect them to help. Patients have shown they’re willing to speak with her time and time again and it extends the customer’s ability to deliver care when it's needed. ”
Protecting patient data with HIPAA-compliant processes
MyndYou is continually refining Eleanor to improve the quality of its user interactions and deliver better outcomes for patients and healthcare providers. One of the ways MyndYou addresses ongoing quality is by analyzing calls to understand how effective the questions are, where patients are dropping off, and how conversations could be improved. By feeding the data from thousands of daily calls into the Google AI PaLM 2 large language model (LLM) in Vertex AI, the MyndYou team can query that data to generate natural language responses that summarize user pain points and help it to understand how to improve its service.
Given the sensitivity of the patient information Eleanor handles, it’s essential that MyndYou uses an LLM that secures that patient data in a fully compliant way.
“PaLM is HIPAA-compliant,” explains Baruchi. “That means we can use it with data from our patient conversations to enable continuous improvement and it will be protected. It is not being used to train the AI model further. That is a game changer for us.”
Growing the business to deliver patient care at scale
As MyndYou grows, the ease of use of Google Cloud solutions allows the company to respond with new developments and expanded solutions
“The Vertex AI platform is very friendly to developers and data scientists. It allows us to move fast and integrate our solution fast without putting significant investments in devops and developers. We could move fast into the market using LLMs in a safe and private way,” says Baruchi.
“We use our proprietary data with Google technology so we have something that is unique. We have fine-tuned the data for our clinical use cases and optimized our processes by using BigQuery. Combining our unique data and unique capabilities with the Google APIs we achieve something that nobody else can.”
Alongside this growth, MyndYou is focused on refining Eleanor to ensure it can deliver the best service possible to healthcare providers and patients.
“The more advanced we can make Eleanor’s interactions, the more we can help our customers manage their individual patients’ journeys, and the greater impact they can have on patient outcomes,” Segev explains. “With Google Cloud, we believe we have the right provider to help us.”
The more advanced we can make Eleanor’s interactions, the more we can help our customers manage their individual patients’ journeys, and the greater impact they can have on patient outcomes. With Google Cloud, we believe we have the right provider to help us.
Tali Segev
VP Customer Success, MyndYou
Based in New York with offices in Tel Aviv, MyndYou is run by a group of clinicians, technical experts, and innovators. With its AI-powered virtual care assistant, it is closing gaps in healthcare by helping healthcare providers detect needs and deliver care for patients at scale.
Industries: Healthcare and Life Sciences, Technology
Location: New York, Israel
Products: Google Cloud, BigQuery, Cloud Run, PaLM, Speech-to-Text, Text-to-Speech
