Chloroform Associated Health Risk Assessment Using Bootstrapping: A Case Study for Limited Drinking Water Samples

Chlorine is the most commonly used disinfectant for watertreatment but when in contact with organic precursors it producestoxic compounds known as trihalomethanes (THMs). Chloroform isthe most commonly occurring THM and constitutes approximately90% of the THM concentration. A study was conducted inNewfoundland for estimating the chloroform content in thedrinking water supplies in Clarenville and St. John's. Thesampling program was conducted in two stages, July and October-November. Higher concentrations of chloroform were observed in Clarenville but in St. John's the concentrations werewithin Health Canada guidelines. The chloroform is a probablehuman carcinogen and is categorised as `B2' by U.S. EnvironmentalProtection Agency (U.S. EPA). The human health risk associatedwith consumption of drinking water was estimated by performingprobabilistic analysis. In the first phase bootstrap simulations(B = 250) were performed to calculate the variability in thedata. The normal distribution was found the best-fitted distribution to chloroform concentration data. For eachbootstrapped run 1000 Monte Carlo Simulations (MCS) wereperformed in the second phase. The cancer risk was estimatedusing dose-response relationship of chloroform. Appropriateprobability density functions (PDFs) were defined subjectivelyfor intake rate, body weight, exposure duration, exposurefrequency and averaging time. The excess lifetime human cancerrisk varied from 0.5 × 10-4 to 4.0 × 10-4, and zero to 1.0 × 10-4 for Clarenvilleand St. John's, respectively. The average cancer risks in Clarenville and St. John's communities were 1.2 × 10-4 and 2.5 × 10-5. The estimated probabilitiesof exceedences above 100 ppb (μg L-1), were 100 and 2.5% for Clarenville and St. John's, respectively which emphasised the need for water treatment. Granular Activated Carbon (GAC) was proposed as one of the potential treatment technology for risk reduction in St. John's community. The treatment efficiency of a proposed technology was assumed uniformly distributed and revised risks were calculated. Theaverage risk in St. John's can be reduced to 8.0 ×10-6 after treatment.