Artificial intelligence overcomes laboratory testing for pathogen detection
Current point-of-care capacity for detection of respiratory infectious diseases is limited, often failing to discriminate between bacteria and viruses. As a result, physicians prescribe antibiotics which are often inappropriate, or even unnecessary, incurring a high economic burden and increasing the risk of antimicrobial resistance. In the COVID-19 era, the ability to eliminate COVID-19 infection for mass populations remotely and in seconds is a game changer for containing the pandemic and optimising medical resources.
AI at front line diagnosis
From a clinical perspective, providing the right treatment within the first 24-48 hours of symptoms is essential. With this in mind, scientists from the EDAS HEALTHCARE project devised a decision support system for physicians that relies on Big Data and machine learning algorithms to diagnose infectious diseases instantly and remotely. “Our patented technology uses artificial intelligence (AI) and anonymous historical lab data to accurately predict infection cause based only on simple demographic parameters,” explains Gil Mildworth, chief business officer at EDAS Healthcare Ltd. The system relies on epidemiology and location-based diagnosis and feeds the information into a combined virology, epidemiology and demographic engine. On a daily basis, the system accumulates information, which is presented as geographical heat maps, showing the current state of infection per pathogen down to the street level. The system considers very few inputs, such as the patient’s age, gender and address, to provide instant and remote diagnosis without a visit to the clinic or hospital. Importantly, the AI technology relies on fact-based data from laboratory tests and can thus be implemented in several clinical workflows. This technology was clinically validated at the Hadassah Medical Centre in Jerusalem, Israel in tens of thousands of patients, enabling the detection of practically all common respiratory pathogens, such as Bordetella pertussis, Haemophilus influenzae, Mycoplasma and Streptococcus pneumoniae. It can also detect viruses that can cause respiratory diseases, including SARS-CoV-2, adenovirus, human metapneumovirus, influenza and respiratory syncytial virus. In terms of performance, EDAS HEALTHCARE results demonstrated over 97 % accuracy in excluding infecting pathogens and over 70 % accuracy in predicting the actual cause of respiratory infection. For COVID-19, accuracy reached 80 %, all without the need for any equipment. Mildworth notes: “The accuracy of excluding a pathogen as the cause of a given respiratory infection is usually 99 %, which is key for providing optimal treatment and coping with COVID-19.”
Implementing the technology in the clinic
Based on market research and interviews with leaders and stakeholders from the United Kingdom healthcare market, EDAS HEALTHCARE partners will focus on primary care and soaring telemedicine services. The technology enables remote accurate diagnosis, which can minimise pathogen spread, as infected patients do not interact with medical staff and other patients in the clinic. Implementation of the EDAS HEALTHCARE technology will lower healthcare costs by increasing primary care capacity, eliminating the need for patient return visits and additional treatments, and reducing unnecessary drug costs and workload in clinical laboratories. Importantly, it can help fight the growing medical challenge of antimicrobial resistance by avoiding unnecessary prescription of antibiotics. In light of the COVID-19 epidemic, the EDAS HEALTHCARE tool enables remote diagnosis without compromising possible isolation. Additionally, it can be used by health organisations to screen mass populations and thus prevent further spread of the virus in the community, thereby increasing the efficiency of existing medical resources. According to Mildworth: “Instantly eliminating COVID-19 with 99 % accuracy is crucial for helping economies return to normality, offering employers the opportunity to scan their employees on a daily basis while enabling molecular labs to double their capacity and optimise test priority. This also means that disease control agencies can react ahead of time and take appropriate containment measures.”
Keywords
EDAS HEALTHCARE, diagnosis, pathogen, respiratory infectious disease, artificial intelligence, AI, COVID-19, coronavirus, machine learning