HOSP: Decision Support Tool
Shir Mnuchin - Team Leader
Chen Agassy - R&D
Vered Rosengarten - Backend developer
Vladislav Krupalnik - Scientific consultant
Uri Marom - Data scientist
Lior Pollack - Solutions Architect
Gal Saada - Client side developer
Shaked Giladi - Testing
Matan Zutta - Client side developer
David Feldstein - Solution Architect
Yotam Shiner - Medical Consultant
Gabriella Goranov - UX Designer
Rotem Kellner - Frontend Developent
Idan Arbesman - UX/UI
Orr Zeitler - Frontend Developent
Hersh Ravkin - Data scientist
Yair Bier - Medical advisor
Paz Agassy - Graphic Designer
Mark Erlich - Data Scientist
Current practice dictates that patients diagnosed with COVID-19 are either hospitalized at a medical center, admitted for hotel-based monitoring or placed under home-stay monitoring. To date, such decisions are made during a phone consultation with the patient’s physician. This laborious process demands that physicians make dozens of calls per day resulting in significant bottlenecks and backlogs. An additional issue is that the pressure to make instant and fast decisions, with no clear data-driven protocol, leads to an excess of patients being hospitalized due to physicians’ fear of making a wrong decision. This creates a huge and unnecessary overload on the healthcare system.
The challenge here was to develop a system to help medical staff make rapid, responsible, and data-driven triage decisions regarding COVID-19 patients, in order to address the bottlenecks faced by healthcare systems.
The HOSP: Decision Support Tool, developed in collaboration with SurgiSphere, is based on a smart algorithm informed by data gleaned from experienced medical staff. By submitting clinical parameters, preconditions, medical history, and patients' mobility abilities, physicians receive an instant score for each patient which represents a researched, data-driven system by which to make a rapid and reliable patient-care decision.
Promising initial research shows that by deploying this tool, decision-making time can be slashed by approximately 80%.
Web based platform development commenced mid-March 2020, followed by IRB approval (Helsinki) at the Assuta Ashdod Medical Center. We started training our statistical model on retrospective anonymous data of COVID-19 patients (PI: Dr. Tal Tova Patalon). Clinical trials were later expanded as we were granted approval for research on retrospective anonymous data of COVID-19 patients at the Sheba Medical Center, the largest medical center in the Middle-East, ranked as one of the leading research hospitals worldwide.
Feedback from community physicians to date has been extremely enthusiastic with the simplicity of the tool and the rapid feedback it provides being cited as the biggest advantage.
We have been selected as one of the leading research groups, thus granted with access to the Israeli Ministry of Health's Data-Lake, containing retrospective anonymous data collected from all COVID-19 patients in Israel (approximately 16,000 patients, dated early May 2020). Research is due to take place in the coming weeks and will help the data scientists refine the model and ensure maximum levels of accuracy.
The team also plans to launch the app in the US and eventually, worldwide. Early discussions with potential users in other countries have been promising.