September 2, 2025
PERK: An R/Shiny application for predicting and visualising Pharmaceutical Pollution and Risk in aquatic environments
Introduction: Pharmaceutical pollution in rivers is a pressing environmental issue, with growing concerns over its impact on aquatic ecosystems and water quality.
Problem: Pharmaceutical pollution in rivers is a growing concern. Every day, people take medications, and much of it is excreted, passing through wastewater treatment plants (WWTPs) before reaching rivers. While some drugs, like acetaminophen, break down quickly, others, such as atorvastatin and tramadol, persist, potentially threatening aquatic life. Understanding how much pollution comes from pharmaceuticals and whether it poses a risk is crucial—but challenging.
While measuring pharmaceutical concentrations in rivers provides valuable data, it’s not a one-size-fits-all solution. Sampling is resource-intensive, requiring specialized equipment and frequent monitoring to capture seasonal or flow-related variations—making it impractical for widespread, continuous use. These challenges highlight the need for predictive models that complement measurements, allowing for broader assessments and better-informed environmental management strategies. Several factors complicate accurate predictions: localized pharmaceutical consumption patterns, individual metabolic differences, variations in WWTP efficiency, and environmental conditions such as river flow and temperature.
Method: To address this, we developed PERK (Predicting Environmental concentration and RisK), a user-friendly R package with integrated web interface that predicts pharmaceutical pollution using WWTP catchment specific prescription data, WWTP efficiency, and river dilution. It helps compare predicted concentrations with real-world measurements and identifies where predictions need improvement.
Results: We applied PERK to estimate pharmaceutical concentrations in a river catchment in South-West England. We tracked 41 pharmaceuticals, including ibuprofen, across five locations. Our predictions were impressively accurate, matching measured values within 0.7–1% in wastewater and rivers. However, notable discrepancies emerged, particularly for atorvastatin and tramadol, highlighting the challenges of predicting environmental concentrations. Prediction accuracy declined in rivers, reflecting the complexity of downstream processes.
PERK also assesses risk using the Risk Quotient (RQ), where values above 1 indicate potential harm to aquatic life. Our findings flagged ibuprofen and naproxen as significant concerns in the studied locations. Encouragingly, the alignment between RQ values based on measured (RQMEC) and predicted (RQPEC) concentrations was strong, with 29 out of 34 compounds consistently classified, validating PERK’s utility.
Future Directions:
Key challenges remain, such as missing data on over-the-counter drugs, regional variations in wastewater treatment, and the need to better account for drug-specific properties and removal processes. Future improvements to PERK will focus on:
- Refining removal estimates using real-time data
- Incorporating drug-specific properties such as degradation rates and therapeutic use patterns
- Enhancing predictive accuracy for pharmaceuticals with high environmental persistence
By bridging the gap between predicted and measured concentrations, PERK sets the stage for better-informed strategies to mitigate pharmaceutical pollution and protect aquatic ecosystems.
Paper (s):
Partners: Wessex Water
Funders: EPSRC (EP/V028499/1), Wessex Water Services Ltd and EPSRC Impact Acceleration Account (Project number: EP/R51164X/1, ENTRUST IAA) and UKWIR IPC project (Innovative Pathways Control).
