Evaluation of a Prediction Model for Influent Pharmaceutical Concentrations

Abstract

This research attempts to evaluate a previously developed model for predicting influent concentrations of prescription pharmaceuticals and understand why some drugs are more accurately predicted than others. It also investigates the impact of demographics on influent drug concentrations in municipal wastewater treatment plants (WWTPs). Samples from two WWTPs in the southeastern United States indicate that the model generates concentration estimates that are accurate to within at least one order of magnitude for a diverse set of drug compounds. Additionally, strategically timed sampling at college-town WWTPs, wherein 19–25% of the population comprises students ages 18–22, reveals that the influent concentrations of several drugs (atorvastatin, simvastatin, rosuvastatin, and metformin) generally prescribed to older-than-average persons increase while students are absent on holiday, whereas the influent concentration of one drug (