Methods. A retrospective observational study design was applied. The study consisted of 100 patients who had been tested positive for COVID-19 infection and their vital signs were monitored. According to the collected data on patients’ physiological parameters and provided responses to the questions related to the infection, prognoses for the state of health were generated by the system. The accuracy of estimated predictions for the health condition was evaluated and compared with the real-time health status of patients.

Results. The results revealed that predictions provided by an algorithm for vital signs, including respiratory rate, systolic blood pressure, temperature and pulse, were quite accurate (> 90%). Oxygen saturation was the only physiological parameter with the lowest precision (72.82%). While comparing the real-time and predicted health condition of patients for today, 90.07% of all generated prognoses coincided with the actual state of health. Nevertheless, the accuracy of the prognosis decreased slightly (84.89%) for the patients’ status of health predictions for tomorrow.

Conclusions. This study indicates that the system for predicting the prospective vital signs and the state of health of a patient is precise and effective. Utilisation of this program could help to enhance the delivery of health care, improve the outcomes for patients in the hospital and ensure the well-being of patients at home.