A predictive model, made in Italy, to calculate the risk of in-hospital mortality for Covid-19

A group of Italian researchers has developed a web app that is available free of charge. The results have been published in the journal Plos One.

The idea to develop the app arose from the observation of the lack of validated tools for predicting in-hospital mortality in Covid-19 patients. A group of Italian doctors from ASST Papa Giovanni XXIII in Bergamo, IRCSS Policlinico San Matteo in Pavia, Policlinico universitario A. Gemelli in Rome and the University of Palermo created a web app and presented the results of their research in a scientific study that was published in the journal Plos One and that you can read here.

The researchers enrolled 2,191 inpatients with Covid-19 from three facilities (1,810 patients from the Bergamo and Pavia units; 381 from the Rome unit). The patients, whose average age was 67 years (45% were 70 years or older), were then classified into three macro profiles.

Seven independent risk factors for hospital mortality were identified: age, male sex, duration of symptoms before hospital admission of less than 10 days, presence of previous diseases such as diabetes, coronary artery disease, chronic liver disease, and lactate dehydrogenase level on admission.

The researchers found that a shorter duration of symptoms before hospital admission was associated with higher hospital mortality, in contrast to those who were admitted after a longer duration of symptoms.

On admission to hospital, fever was present in 85% of patients, dyspnoea in 56% and cough in 44% of patients. At the end of follow-up, 540 patients had died (24.6%), 302 (13.7%) had been transferred to intensive care, 1,358 patients (62.0%) had been discharged and 258 were still in hospital. In the best profile, in-hospital mortality at 7 and 21 days was 5% and 8%, respectively; in the intermediate profile, 18% and 28%; in the worst profile, 52% and 70%.

The Gray competing risks multivariate model (with discharge as a competing event) was used to develop the algorithm for predicting hospital mortality. Discrimination and calibration were assessed by the area under the receiver operating characteristic curve (AUC) and the Brier score in both the derivation and validation cohorts. The AUC was 0.822 (95%CI 0.722-0.922) in the derivation cohort and 0.820 (95%CI 0.724-0.920) in the validation cohort with good calibration.

The web app is available free of charge at the following address.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s