COVID-19 threatens to overwhelm our healthcare systems. In many cases, healthcare personnel cannot keep up with the influx of patients. Help may be on the way, however. A company has launched a new app that relies on telehealth to screen patients based on their symptoms and concerns.

Decoded Health is a new startup with an impressive lineage; it has the backing of Menlo Park’s SRI International‘s SRI Ventures. Its Virtual Medical Resident app for telehealth plays the role of a resident physician in an emergency department, providing the patient’s first point of contact. A resident’s primary role is to assess patient complaints, vital signs, and symptoms so that the patient is directed to the next stage of appropriate care.

As we’ve written here many times, telehealth helps patients avoid unnecessary trips to a hospital or doctor’s office. This eliminates unnecessary exposure to the coronavirus or other infectious diseases. This also reduces the number of patients that a healthcare facility must physically accommodate and process. The Virtual Resident helps make telehealth screening practical and effficient.

Decoder Health also announced a partnership with Vituity, a physician-owned and operated acute care management and staffing group with a large telehealth division. Vituity was founded more than 40 years ago in northern California and will use the Decoder Health app to boost the efficiency of its telemedicine and in-person practices.

The Decoder Health app performs as a “virtual medical resident” by using explainable artificial intelligence with advanced machine learning. “Explainable AI” means that the system’s decision making process can be traced and understood by human experts. The Virtual Medical Representative system uses natural language processing to identify the key problem from what a patient says or writes.

For example, consider a case where a patient calls and says “I’m glad I could reach someone because I’m in a lot of pain and I’m worried because my wrist hurts.” Natural language processing would identify “wrist hurts” as the most important part of what the caller said. Deep learning algorithms would then guide a process of questions that would lead to diagnosis and action recommendations. There’s a lot more to the process than parsing one sentence. Decoder Health describes its app as “as a force multiplier for physicians” and suggests the current model can increase the primary healthcare contacts by ten times.

With the current fear of healthcare system overwhelm, the promise of the Decoder Health app is welcome and timely. This approach to automating healthcare intelligence could reduce workloads for physicians and other staff, while delivering faster treatment decisions for patients.