New tool could ‘help UK doctors spot high-risk Covid patients in seconds’ | Coronavirus outbreak

Join Hafta-Ichi to find out the article “New tool could ‘help UK doctors spot high-risk Covid patients in seconds’ | Coronavirus outbreak”

A risk calculator that takes seconds to produce a score indicating a Covid-19 patient’s risk of death could help clinicians make care decisions soon after patients arrive in hospital, according to a large study conducted by a consortium of researchers across the UK.

As UK Covid-19 cases rise, schools reopen and the weather gets colder, doctors at UK hospitals are expected to see an influx of coronavirus patients.

Patients with Covid-19 behave very differently to patients with other conditions such as flu and bacterial pneumonia, said Dr Antonia Ho of the University of Glasgow, one of the study’s authors, and it is very challenging for doctors managing this unfamiliar disease to accurately identify those who are at high risk of deterioration or who can ride out their illness at home.

“So having a tool that … can help clinicians at the front door to accurately group patients who are coming in with Covid-19 into four distinct risk categories – low, intermediate, high and very high risk – is hugely valuable,” she added. “Having an accompanying low-risk score will provide that doctor with increased confidence that the vast majority of people, patients with that low-risk score, will come to no real harm.”

To develop the tool, researchers collected routine data from 35,463 adults, with a median age of 73 years, with Covid-19 who were admitted to 260 hospitals across England, Scotland and Wales between February and May this year.

The researchers focused on eight metrics that play a key role in determining mortality risk – age, sex, number of underlying conditions, respiratory rate, blood oxygen concentration, level of consciousness, urea, and C-reactive protein (a chemical linked to inflammation).

These factors were combined into a model to give a score ranging from 0 to 21 points. Age, they concluded, is the biggest predictor, accounting for a third of the score. The calculations suggested patients with a score of at least 15 had a 62% chance of mortality compared with 1% mortality for those with a score of three or less. The tool, which is easily accessible on a smartphone or computer, takes seconds to generate a score and is expected to be rolled out in the NHS this week.

The researchers then validated the tool on a further 22,361 patients admitted to the same hospitals between May and June, and found that the tool performed similarly. The tool also compared favourably to 15 pre-existing models, including machine learning techniques, they found.

“The accuracy is almost 100% in terms of predicting the expected versus the predicted mortality,” said another of the study authors, Stephen Knight.

A fifth of the patients in the study were from minority ethnic backgrounds, which represents the UK population as a whole, noted study author Prof Ewen Harrison.

“We have tested the tool quite extensively, specifically in ethnic minority subgroups, and found that it performs as well or better than the white subgroup.”

But the tool does have some limitations. Since it was tested in a hospitalised elderly population, it is not applicable for paediatric patients or for use within the community. “It remains to be seen whether this score will be as accurate to predict the risk of death in younger patients and… also in populations outside of the UK,” Ho added.

Dr Amitava Banerjee, an associate professor in clinical data science at University College London who was not involved in this study, said as a clinician he would be keen to use the tool.

“Of course, with these kind of risk scores a lot of people feel that it is telling them what they know already,” he said. “But I still think it’s useful to have this built into our approach.”

Source: The Guardian
Keyword: New tool could ‘help UK doctors spot high-risk Covid patients in seconds’ | Coronavirus outbreak

Leave a Comment

Your email address will not be published. Required fields are marked *