Senthil Kumaran is CTO of virtuwell by HealthPartners. He’ll talk about why virtuwell uses data to forecast demand and how it produces better outcomes.
What is virtuwell by HealthPartners all about?
Launched in 2010, virtuwell was designed to disrupt the traditional healthcare visit. Our online clinic has sent half a million treatment plans to consumers seeking convenient care. We’re beating standard measures for quality and affordability, without sacrificing the patient’s experience. The medical practitioners at virtuwell diagnose and treat a range of common illnesses remotely, saving patients a trip to the clinic or hospital. Diagnoses are based on answers gathered in an online interview, including the ability to securely upload photos within the interview.
What kind of advanced technology is required for a platform like this?
We use AI tools to predict follow up needs of patients and route the patient information to the right provider. I like to think of it as Augmented AI –using technology to help human clinicians provide the patients the best care.
To be able to provide quality care quickly and at a low price point, we have to accurately forecast demand for virtuwell and staff accordingly. Demand changes throughout the day but we always need the right specialists, licensed for the states we serve, available to meet it. As the business has grown and forecasting has gotten more intricate, we have adapted to use a time series prediction to staff our clinicians. This is the subject of my MHTA talk.
What approach did you take for time series prediction?
In the session I’ll explain how we work with our own historic data, information about our ongoing customer advertising, and external factors like weather and Google search data to drive predictive and prescriptive analytics and forecast our demand. You’ll learn how by matching a few data points, we’re able to more efficiently staff our online clinic, keep treatment times low, and manage our staffing costs.
How do you know your model works?
Our clinic operation partners are using the model today to staff our practitioner shifts more efficiently. They help us tweak the model to make it better, and it continues to improve as we feed more data into it. Throughout this we continue to maintain a 98% satisfaction rate with our customers, something that wouldn’t be possible if we weren’t delivering high quality care in a timely manner.
What’s next for virtuwell, technologically speaking?
Telehealth is an almost $6 billion business and is growing fast, thanks to the rise of internet of things devices and the explosion of wearables. As consumers gain access to devices that can read heartrate, blood sugar, and other data points that have traditionally required lab tests and physical contact, telehealth companies like us will be able to treat more and more conditions, and save our consumers time and money. Companies like virtuwell are vital in connecting patients with the right clinician and getting them the care they need, which is why it’s essential that we get the tech right.
Senthil’s session is one of many focused on personalization at the conference. Check out these related sessions on our conference web site:
- Building an end-to-end data science practice in an Agile framework Presented by Susan Van Riper, senior product capability manager for Machine Intelligence / Data Science at Best Buy
- Collaborative personalization at scale and AI-enabled experiences, presented by Domingo Huh, lead visual UX designer at Thomson Reuters
Sidebar: Overview of Senthil’s session, AI and telehealth
In this session we will see how Artificial Intelligence is changing the medical diagnostic industry. Only a few telemedicine platforms are using a virtual triaging engine for checking symptoms, then using AI to diagnose and create treatment options for providers. This session will explore how virtuwell, a leading telemedicine platform from St. Paul, is deploying predictive analytics to reduce bottle necks and improve patient flow. Total healthcare costs are reduced by funneling more relevant patient and provider data to electronic medical records. Improvements to customer satisfaction in a virtual setting are made by analyzing the patient data set, systematic follow ups, and connecting patients with the right clinician at the right time. Infusing telemedicine platforms with machine learning algorithms will mean better diagnoses with less human effort, not only on acute conditions but on some chronic conditions as well. We will see how AI is used in cardiovascular diagnosis and lung cancer detection in London hospitals, and learn who is working on Smart IOT technology that will decrease the cost of delivering healthcare services while improving quality of life for patients.
