Experts at Massachusetts Central Hospital in Boston, along with a colleague from Australia, have created a computer model that determines the likelihood of death in patients with COVID-19 based on the history of illnesses they suffered prior to being infected with the coronavirus. The developers analyzed data from electronic medical records of 16.7 thousand patients of the Mass General Brigham network of non-profit hospitals and identified 46 diseases and 5 demographic factors that, to varying degrees, affect the likelihood of dying from coronavirus infection.
Calculations have shown that the risk of death from COVID-19 is maximized if a person has chronic kidney disease, hypertension, heart failure, atrial fibrillation or an abdominal aortic aneurysm. Also, coronavirus infection is most dangerous for patients with coronary heart disease and other cardiovascular diseases, with impaired cerebral circulation, with various pulmonary diseases and with diseases of internal organs in the later stages. In addition to risk assessments, the specialists used another calculation method, which shows how one factor in the model affects another. It turned out that the likelihood of death is most dependent on age, the presence or absence of a history of pneumonia, type 2 diabetes, heart failure and chronic kidney disease.
As conceived by the authors, the results of the model’s forecasts can be used to identify the most vulnerable population group in order to focus efforts on its protection. For example, to give these people vaccinations as a priority.
Read more in the exclusive material from Izvestia:
Detailed outcome: a method for predicting death from COVID-19 based on medical records has been created