Enhancing Confidence in Drinking Water Quality for Improved Risk Decisions

ABSTRACT

Reliance is often placed on “indirect” indicators to detect changes in water quality in real time. However, there are challenges in using these indirect indicators, as they cannot replace standardized laboratory tests. As an alternative, reliance on three successive test cases is examined, based on “joint and conditional probabilities” of hazard (contaminant level exceeding acceptable level) that could enhance confidence in monitoring results: (i) a simple and quick screening analysis sensor (primary analyzer, having type-1 error, i.e., α = 0.10 and type-2 error, i.e., β = 0.05), (ii) a repeat of the analysis with the same rapid analyzer is feasible, and (iii) further analysis of the sample with a highly reliable capability is feasible (secondary analyzer with α = 0.05 and β = 0.02). This three-level monitoring protocol for obtaining the posterior probability of hazard is examined and a method is proposed for taking improved risk management decisions by accounting for sensitivity and specificity of monitoring instruments. The revised probability of hazard will ensure that a positive harmful detect is proven false or real with much greater certainty and series of action (or no action) is taken at the earliest.

Key Words:

• water quality monitoring
• false positives
• water distribution
• risk management