Predicted and measured concentrations of pharmaceuticals in hospital effluents. Examination of the strengths and weaknesses of the two approaches through the analysis of a case study

This study deals with the chemical characterization of hospital effluents in terms of the predicted and measured concentrations of 38 pharmaceuticals belonging to 11 different therapeutic classes. The paper outlines the strengths and weaknesses of the two approaches through an analysis of a case study referring to a large hospital. It highlights the observed (and expected) ranges of variability for the parameters of the adopted model, presents the results of an uncertainty analysis of direct measurements (due to sampling mode and frequency and chemical analysis) and a sensitivity analysis of predicted concentrations (based on the annual consumption of pharmaceuticals, their excretion rate and annual wastewater volume generated by the hospital). Measured concentrations refer to two sampling campaigns carried out in summer and winter in order to investigate seasonal variability of the selected compounds. Predicted concentrations are compared to measured ones in the three scenarios: summer, winter and the whole year. It was found that predicted and measured concentrations are in agreement for a limited number of compounds (namely atenolol, atorvastatin and hydrochlorothiazide), and for most compounds the adoption of the model leads to a large overestimation in all three periods. Uncertainties in predictions are mainly due to the wastewater volume and excretion factor, whereas for measured concentrations, uncertainties are mainly due to sampling mode.