Estimating Respiratory Parameters Using Intra-Arterial Partial Pressure Measurements

We develop algorithms for estimation of respiratory parameters in ventilated neonates. Such algorithms are important for maintaining the oxygen and carbon dioxide blood levels of preterm infants. We present mathematical and statistical models of partial pressure in the arterial blood. Our mathematical model includes physiological compartments, diffusion exchange and mass transfer. To account for possible heteroskedasticity we present parametric model of the variance and estimate the parameters using maximum likelihood. To implement physiological constrains we use Bayesian estimator applicable when informative priors are available. We evaluate the applicability of our techniques using a real data set of intra-arterial pressure measurements.