A group of research scientists have developed Artificial Intelligence-based models to predict the risk of mucormycosis among COVID patients at the time of discharge from hospitals.
The study was published in the Journal of Infection (an official journal of the British Infection Association) with a high impact factor of 38.63.
Mucormycosis is a rare but life-threatening fungal infection. India has been facing a public health challenge from COVID-19 related mucormycosis.
A collaborative study by various universities was aimed at developing AI-based models to predict mucormycosis among patients at the time of discharge.
The dataset included 1,229 COVID patients and 214 in-patients, COVID positive and infected with mucormycosis. The authors used logistic regression, decision tree, random forest, and the extreme gradient boosting algorithm. All the models were evaluated with five-fold validation to derive a reliable estimate of the model error.
The study determined that the top five variables positively impacted mucormycosis risk thereby impacting obesity, anosmia, de novo diabetes, myalgia, and nasal discharge.
The developed model can predict high risk in patients, thus initiating preventive care or aiding in early detection of mucormycosis. Thus, this study holds potential for early treatment and better management of patients suffering from COVID-19-associated mucormycosis, the authors said.
One of the authors, professor Raja Shekhar Bellamkonda from the School of Management Studies, University of Hyderabad, said that applying technology and mathematical models will be crucial for the growth of the healthcare sector.
Professor Shabbir Syed Abdul from Taipei Medical University, Taiwan, said that appropriate use of AI in healthcare can help us in this challenging pandemic situation. “There is a need for more collaborative research work in the healthcare domain,” professor GVRK Acharyulu of UoH said.
