Confidence intervals for the parameters estimated from simulated O2 uptake kinetics: effects of different data treatments

New Findings • What is the central question of this study? Is the accuracy and/or precision of the O2 uptake kinetics parameters, estimated by non‐linear regression, affected by different data treatments applied to reduce the noise of breath fluctuations? • What is the main finding and its importance? Simulations showed that, even after the averaging of more repetitions, the width of the asymptotic confidence intervals was narrower than the real ones, in particular when the O2 uptake responses were resampled at time intervals shorter than the average breath duration (e.g. 1 s). The reasons for this discrepancy were investigated, allowing us to identify simple methods to obtain the correct confidence interval of the O2 uptake kinetics parameters. The behaviour of pulmonary O2 uptake following a moderate‐intensity step exercise increment is usually described by a first brief increase, followed by a second exponential time course reaching the new steady state (phase II). The parameters describing the phase II kinetics are investigated by applying different data treatments to the acquired O2 uptake data to reduce the effects of their noise before running a non‐linear regression procedure. The effects of different data treatments (nothing, resampling at various time intervals or averaging of more repetitions) on the precision and/or accuracy of the kinetics parameters estimated by non‐linear regression with a simple mono‐exponential model were investigated by artificially generating 105 simulated responses with average breath duration of 3.5 s. The simulations showed that, whatever the explored data treatment, the average estimated parameters were close to the theoretical ones. Nevertheless, in all the explored conditions, the non‐linear regression provided narrower asymptotic confidence intervals than the real ones. In particular, when the responses were resampled at 1 s time intervals, the width of the asymptotic confidence interval for the time constant was 50% of the real one, even after the averaging of more repetitions. The reasons for this discrepancy were investigated further, allowing us to identify some methods to obtain the correct confidence interval of the O2 uptake kinetics parameters. The simplest method to obtain an asymptotic confidence interval close to the real one is to average more responses resampled to a time interval slightly longer than the average breath duration.