Facebook and New York City-based NYU Langone Health have developed machine learning models that, using X-rays, can help physicians predict how patients' COVID-19 condition will progress, the social media company said Jan. 15.
There are three models: one that predicts patient deterioration using a single X-ray, one that predicts patient deterioration based on a sequence of X-rays and one that predicts how much supplemental oxygen a patient might need using a single X-ray, Facebook said.
Facebook said its sequential X-ray model can predict up to four days in advance if a patient may need intense care solutions, and that its model's predictions outperform the predictions of human experts generally.
These models may be able to help hospitals make resource distribution decisions as COVID-19 continues to spread, Facebook said.
More articles on artificial intelligence:
MUSC creates AI tool to predict patients' risks for COVID-19 complications
3 recent studies exploring medical AI tools' efficacy
Children’s National Hospital, NIH launch COVID-19 diagnostics AI challenge