Novel Technologies for Respiratory Virus Identification

NHS services always come under pressure over the winter period as demand for hospital services increase with the arrival of cold weather and the increase spread of respiratory infections.

As hospitals now look beyond the COVID-19 pandemic, the need to meet future hospital service challenges are important. Changes in habit and measures put in place to control the spread of COVID-19 have affected the transmission of other respiratory viruses. Understanding changes in occurrences of respiratory like illnesses during the winter period could help in understanding future winter pressures for healthcare services.

Since viral and bacterial infection symptoms can be similar, broad spectrum antibiotics are often used as the first line of defence even when the cause of infection is unknown. Antibiotics are not effective against viral infections and only work against bacteria. Antibiotics prevent millions of deaths each year however, inappropriate prescription rates and overuse have led to antibiotic resistance that has created a global health emergency in superbugs resistant to these treatments. There is a need for rapid and sensitive tests to diagnose the cause of infection and ensure patients get the most suitable type of treatment.

THE TECHNOLOGY The hospital is working closely with Pictura Bio, who are developing a rapid pathogen identification technology comprising microscope and a computer which is trained to recognise fluorescent patterns from different viruses. The performance of this AI algorithm is highly dependent on the quality of the samples used to train it. It has the potential to be quicker than current gold standard technology and could be used in a patient facing setting including GP surgeries and Pharmacies.

THE RESEARCH STUDY This is a proof of concept study to determine whether the Pictura Bio system is able to detect and differentiate different viruses from nasopharyngeal swabs/aspirate specimens. The data collected will be used to "train" the algorithms to be able to accurately identify respiratory viruses. The accuracy with which the algorithms estimate the test dataset will be monitored at regular intervals during the training dataset collection period. The Pictura Bio system is still under development, which means that it is still "learning". The system needs to see more information so that it can be sufficiently accurate to be used in clinical practice and should become more accurate in identifying these viruses as it sees more and more information from patients.

This study will take place at Portsmouth Hospitals University NHS Trust and aims to recruit 1000 patients. To do this, we will recruit both adults and children who either present to the emergency department or are admitted to QAH with a clinical suspicion of a respiratory viral infection. All participants will have a nose and/or throat swab taken as part of their clinical assessment, and we would ask to take a further nasal swab for the purpose of the study. Research sampling will be combined with routine clinical samples where possible to reduce the frequency of testing. We will use most of the information to teach the system how to become more accurate at identifying respiratory viruses. We will keep the remaining information separate and use it to test how accurate the system is. All of the data will be kept securely. Basic information will be collected including age, gender, results of blood tests taken for clinical review, treatment and outcome data. No results from the swabs taken for the purpose of the study will be available to either the participant or the clinical team and the information will have no effect on patient care.

SHARING OF FINDINGS Results of the study will be disseminated via paper submissions to relevant journals and conferences. A lay summary of the results will be produced in collaboration with the PHT PRA group. The study team will explore with this group other avenues and formats for the dissemination of the study findings to ensure as wide a public audience as possible, for example through co-produced public talks, articles in community communications and social media.