Researchers at the Mass General Brigham healthcare system in the US have used artificial intelligence to help identify long Covid and track how different symptoms manifest over time. Through the tool they found that 22.8% of the population shows long Covid symptoms.

The algorithm used in the AI ​​tool was developed by drawing on data on nearly 300,000 patients at 14 hospitals and 20 community health centers in the Mass General Brigham system.

The tool analyzes electronic health records and helps clinicians identify more cases of long Covid so patients can get care.

The results, published in the journal Med, suggest that the incidence of the disease could be greatly underestimated. According to the new approach, the percentage of the population estimated to be suffering from long Covid amounts to 22.8%, while previous studies estimated the percentage to be 7%. The authors note that this percentage gives a more realistic picture of the long-term effects of the pandemic.

The researchers found their tool to be more accurate and less biased than other algorithms based on a single diagnostic code or individual clinical encounters, as the people it identified as having long-term Covid reflected the broader demographic makeup of Massachusetts.

The researchers plan to make the algorithm public so that doctors and health care systems around the world can use it.

Limitations of the study include the fact that the health record data used may be less complete than that recorded by physicians in their clinical notes after the medical visit.

Future studies may explore the use of the algorithm in groups of patients with conditions such as chronic obstructive pulmonary disease or diabetes.