Introduction: Viscoelastic models are widely accepted in modern literature to explain the non-linear relationship between the force applied by the rescuer and the chest displacement that occurs during chest compressions (CCs). Various studies have shown that applying CCs for prolonged periods of time changes chest properties, but proposed models have mostly been tested under laboratory settings with few patients or during short CPR intervals. Aim: To retrospectively characterize the mechanical dynamics of the chest during manual CCs in pre-hospital cardiopulmonary resuscitation using a simple viscoelastic model. Methods: Force and acceleration signals were extracted from CPR monitors used during pre-hospital resuscitation efforts by TVF&R (OR, USA) on adult patients (≥ 18yo), from which individual CCs and series of CCs between pauses were identified. The model decomposes the measured force as the sum of an elastic term and a damped term, specifically F = k·x + d·v . We characterized each full cycle with a spring coefficient k (N/cm) to measure the stiffness of the chest; and with separated compression and recoil damping coefficients, d c and d r (N·s/cm), to measure its viscosity. Parameters were averaged for sets of 10 CCs. We characterized the variations of the model parameters within each series. We also assessed the chest recovery between consecutive series by comparing the values obtained at the end of each series with the beginning of the next series. Results: A total of 1,156,608 CCs from 615 patients were analysed. From these, 6,073 series of median (IQR) length 190 (119-222) CCs were selected. All parameters decreased within the series, with decreases of the median after 20 sets of 10 CCs of 9.8%, 7.1% and 11.1% for k , d c and d r , respectively (p trend < 0.05). When examining the first 5 series for all cases, k presented a mean recovery of 7%, d c of 5%, and d r of 5%. Conclusion: The proposed model allows the observation of a reduction in stiffness and viscosity of the chest within CC series. It further allows for evaluating the dynamics of the chest around pauses, showing recovery of its properties during the pause. Overall, the model reveals a gradual reduction in chest stiffness and viscosity during CPR.