Texas Hospital Taps AI at the Edge to Build Safe, Intelligent Virtual ICU

The global pandemic motivates Houston Methodist Hospital to roll out a virtual ICU. Using software that collects and analyzes data and uses AI at the edge, clinicians can treat patients while reducing exposure to infectious disease.


Article Key

Image credit: Sickbay/Houston Methodist Hospital

In the early days of the pandemic, medical workers in the U.S. were watching their counterparts in other countries fall ill, and even perish, as they cared for patients who contracted the highly infectious COVID-19. Medical facilities immediately began researching new ways to handle patient care and protect their essential workers.

New protocols for patient visits emerged, and the use of telehealth soared as doctors treated 50 to 175 times more patients via telehealth than they did before COVID-19. The acceptance of telehealth—by providers, patients, and insurance companies—has paved the way for even more innovation in patient care and treatment. Witness the rise of the virtual intensive care unit.

In January, as the pandemic spread across international borders, Houston Methodist Hospital had created a technology hub to explore how commercially available products and technologies could be used to improve patient care. The timing was fortuitous. When the hospital began seeing COVID-19 patients in March, Houston Methodist was able to pivot quickly. According to Roberta Schwartz, chief innovation officer at Houston Methodist, the hospital rolled out its virtual ICU in two weeks.

Aggregated Patient Data

Houston Methodist’s virtual ICU allows medical staff to monitor multiple patients remotely using the FDA-approved Sickbay Precision Medicine Platform for Critical Care from Medical Informatics Corp (MIC). The software platform collects real-time patient data from all types of bedside machines, including cardiac monitors and ventilators. 

Using artificial intelligence, machine learning, and predictive analytics, the Sickbay platform can anticipate or identify patient risk factors. The data is presented to medical staff on a dashboard that displays near real-time waveform data from all machines; it also ties in the patient’s electronic health record, labs, medications, and other related information for the duration of the patient’s stay. Staff receive alerts when immediate concerns arise.

The Sickbay software runs on Intel® Edge AI technology and Cisco’s Digital Network Architecture (DNA) technology. Using Intel® Xeon® processor-based platforms, Sickbay provides graphical images and data analytics to health care providers, who can access the information via Web-based apps. That enables doctors and nurses to monitor patients from a command center, office, or home using a computer, smart phone, or tablet. Medical staff view a dashboard that displays multiple patients’ data in a single screen.

Hospitals can use Sickbay’s apps or create customized apps using open APIs and SDKs to gain more insight from the patient data. The Sickbay platform is vendor-agnostic, which is important in hospital settings where critical care machines are sourced from different vendors or when hospitals with disparate systems merge.

Flexible Critical Care Monitoring

Houston Methodist can monitor 180 COVID-19 patients remotely—onsite or off. The hospital has trained 900 providers, and virtual visits have grown from 5 percent pre-COVID to 80 percent. In one two-month span earlier this year, the hospital staff responded to 400 virtual alert calls to the bedside and made 7000 remote connections to patient rooms using audiovisual equipment.

The virtual ICU enables medical staff to reduce exposure to infectious diseases, such as COVID-19, and it brings other benefits as well. Constant monitoring means that health care professionals can react more quickly to patient emergencies. Instead of waiting for a nurse, who might be tending to a patient in another room, the person monitoring from the command center can react to an alert immediately and communicate with a patient directly to assess the situation. More urgent cases can be prioritized.

different screen types showing ICU stats

Image credit: MIC

In addition, it improves efficiency. Remote access saves much-needed personal protective equipment for those who need it. Doctors on call or at home can assess patients without having to go into the hospital. Consultations are easier because the information is available online. Machine learning and AI can track trends across patient data, which can lead to more effective care.

In the case of Houston Methodist, patient rooms can become part of the virtual ICU when needed, and in just a few minutes, allowing the hospital to quickly ramp up critical care services. The virtual capabilities also allow Houston Methodist to provide medical specialists to nearby rural hospitals that lack resources or need immediate access to specialized expertise.