Artificial Intelligence predicts best treatment for Covid-19 patients

Researchers have used Artificial Intelligence (AI) to predict which critically ill Covid-19 patients might respond to interventions carried out in an intensive care setting, such as proning – where patients are turned onto their fronts to get more oxygen into the lungs. This approach, where comprehensive patient data is analysed day-by-day, could be used to improve Covid-19 guidelines and determine the best clinical treatments.

The researchers, led by a team from Imperial College London and Royal Brompton and Harefield hospitals, for the Covid-ICU Consortium (a national alliance of clinicians researching the most effective treatments for Covid-19 patients), analysed data from 633 mechanically ventilated Covid-19 patients across 20 Intensive Care Units (ICUs) in the UK during the first wave of the Covid-19 outbreak (1 March – 31 August 2020). The team examined the importance of factors associated with disease progression, like blood clots and inflammation in the lungs, as well as treatments given and whether patients were discharged or, sadly, died. They used this data to design and train an AI tool that can identify how patients will respond to certain clinical interventions.

The researchers found that during the first wave of the pandemic, patients with blood clots or inflammation in the lungs, lower oxygen levels, lower blood pressure, and higher levels of lactic acid in the blood (a natural substance that builds up due to a lack of oxygen) were less likely to benefit from being placed in the prone position (on their front). Overall, pronation increased the level of oxygen in the body in only 45 per cent of patients. Going forwards, this information can be used to ensure that patients receive timely sequential interventions and their responses are assessed, and patients that do not respond can be referred for other interventions, such as extracorporeal membrane oxygenation (ECMO) – a life support machine that acts as an artificial heart and lungs to pump blood around the body.

First author and clinical science lead Dr Brijesh Patel, honorary consultant at Royal Brompton Hospital and clinical senior lecturer in cardiothoracic at Imperial College, said: “This AI approach will allow patient care to be streamlined so that the window of opportunity for interventions, such as ECMO, are not missed. ECMO is currently the last resort for Covid-19 patients and over 20 per cent of all patients on a mechanical ventilator were referred to one of the five adult ECMO centres in the UK. However, only 4 per cent received ECMO. This will have been due to a number of reasons, one of which is that some patients were not referred early enough.

“Our analysis shows that patients need to be referred for ECMO as soon as other less invasive interventions, such as proning, have been shown to not work and this AI tool enables clinicians to predict if patients will respond to proning. If they will not, patients may be referred to ECMO sooner. This national evaluation enabled us not only to examine disease course and how our management decisions affected this course, but importantly where we could improve.”

Dr Patel continued: “Most studies look at the health of patients on admission to ICU and whether they get better or, sadly, die. In ICU there is a huge amount of information which we use at the bedside to manage patients on a day-by-day basis. Our study focuses on how patients’ conditions changed daily. It helped focus our attention on which specific parameters matter the most, and how the importance of each parameter changes over time. This dynamic understanding is vitally important when trying to understand a new life-threatening disease and to know when and in whom each intervention works. We hope our findings will help and encourage more research to be undertaken that focuses on the daily needs of patients.”

Senior author and data science lead Professor Aldo Faisal, Director of Imperial’s Centre in AI for Healthcare at the Departments of Computing and Bioengineering, said: “More than one year on, we’re still learning how the course of Covid-19 affects the body, and how this can change day-by-day. Data science and the daily data feeds from ICUs across the country help us learn much faster how best to treat individual patients based on their daily symptoms and needs.”

This is the first study that examines daily clinical patient data, using AI to understand the clinical response to the rapidly changing needs of patients in ICU. It could be applied to potential future waves of the pandemic and other diseases treated in similar clinical settings.

The researchers continue to collect patient data and are currently curating data and analysing findings from the second wave of the pandemic. The study was co-funded by the Imperial College London Covid-19 Research Fund, Royal Brompton & Harefield Hospitals Charity, and NIHR Imperial Biomedical Research Centre, and is endorsed by the Intensive Care Society.

Natural history, trajectory, and management of mechanically ventilated COVID-19 patients in the United Kingdom by Brijesh Patel et al., published 11 May 2021 in Intensive Care Medicine.