The aim was to analyze the prognostic value of ventilation/ gas exchange dissociation markers to predict the treatment length in patients with the acute phase of COVID-19.Materials and methods: The analysis was performed using a database of 384 randomly selected patients from December 2021 to May 2022 with a confirmed diagnosis of the acute phase of COVID-19, aged 61±16 years. Spirometry of maximal inspiratory pressure (MIP) and maximal expiratory pressure (MEP) was performed on 2,2±0,2 day of hospitalization with a portable MicroRPM device (CareFusion, UK) and measurement of surrogate oxygenation index (SpO2/FiO2), ROX index (SpO2/ FiO2)/respiratory rate) was carried out in all patients. Log-regression models were used (STATISTICA 10) to determine cutoff values for these functional variables and their ability to predict the patients length of hospital stay from the date of examination (≤ 7 vs > 7 days).Results: The lowest sensitivity (SE) and specificity (SP) were shown in models using only the respiratory muscle strength markers MIP (SE=54%, SP=70%) and MEP (SE=73%, SP=47%), that were the same in the combination of MIP and MEP (SE=65%, SP=58%). The areas under ROC were equal to 0.6 in all three models. The models based on hypoxia markers had a higher classification power (AUC 0,7) compared to the previous three, and the sensitivity value was higher in the model using the ROX index (SE = 58% and SP = 78%), but the specificity was better in the surrogate oxygenation index model (SE = 48% and SP = 88%). A complex model based on a combination of the two integral hypoxia indices, as well as a diaphragm strength marker (SpO2/FiO2+MIP*ROX+MIP) had the best sensitivity (67%) and specificity (84%), and the area under ROC reached 0.8.Conclusion: Indicators of respiratory muscle strength and oxygenation are suitable markers for combined models and algorithms for predicting hospital length of stay in COVID-19 patients.