Analysis of Temporal and Spatial Variability in Disinfection By-product Concentrations for Exposure Assessment Applications

Abstract

PP-30-118 Background/Aims: Most epidemiological studies of disinfection by-products rely on aggregate water system average total trihalomethane (TTHM) and total haloacetic acid concentrations as surrogate measures for individual-level exposure. It is critical that these surrogates accurately reflect temporal and spatial variability in disinfection by-product concentrations to reduce exposure misclassification. Methods: We assessed temporal and spatial variability in routinely collected TTHM monitoring data for 201 public water systems (PWSs) in Massachusetts from 1995 to 2004. Results: The annual TTHM average across 46 PWSs with complete quarterly data remained relatively unchanged during the study period (1995: 41.3 μg/L; 2004: 42.4 μg/L). However, average TTHM among systems disinfecting with chloramination (n = 5 PWSs) increased considerably (1995: 23.6 μg/L; 2004: 43.0 μg/L), while TTHM levels increased slightly among chlorinated systems (n = 30 PWSs). The quarterly averages across 46 PWSs showed consistent seasonality with a 30%–50% higher summer average TTHM compared to winter. High spatial variability (ie, ≥1 sampling date with TTHM between-location differences was ≥30 μg/L) was noted in 49% of PWSs. Over half of these PWSs had high spatial variability for at least 25% of their sampling dates. We also examined spatial variability based on different disinfection practices for 4 PWSs that used chloramination at some point during the study period. Among these systems, 64% of the chlorinated sample dates had high spatial variability compared to 1% of the chloraminated sample dates reported. Conclusion: Our seasonal variability findings emphasize the need for quarterly sampling dates to adequately capture short-term temporal trends. We found considerable spatial variability across several water systems, which could result in misclassification bias if system averages are used to estimate individual-level TTHM and THAA exposures. Therefore, exposure assessment approaches should target systems with minimal spatial variability and sensitivity analyses should examine potential exposure misclassification in larger epidemiological studies based on water systems using different types of disinfectants.