TY - JOUR
T1 - Microbiological surveillance of private water supplies in England - The impact of environmental and climate factors on water quality
AU - Richardson, Hopi Yip
AU - Nichols, Gordon
AU - Lane, Chris
AU - Lake, Iain R.
AU - Hunter, Paul R.
PY - 2009
Y1 - 2009
N2 - A passive surveillance system captured information on 34,904 microbiological samples from 11,233 private drinking water supplies within England as well as the associated constructional, climatic and environmental variables. Escherichia coli was detected in 6588 (18.87%) of samples and at least one positive sample was detected from 3638 (32.39%) of sites. However, this estimate of supplies failing to meet the European drinking water E. coli standard was probably an underestimate as the more samples taken per supply, the more likely the supply was to fail. A multivariable model of private water supplies data showed a strong seasonal impact, with samples between January and May being significantly less contaminated with E. coli than samples between June and December. Samples from springs (OR 2.5, CI 2.0-3.1) or surface waters (OR 2.4, CI 0.8-7.0) were more likely to fail than groundwater sources, as were supplies with no effective treatment (OR1.8, CI 1.5-2.3). Commercial supplies were less likely to fail than domestic supplies (OR 0.63, CI 0.48-0.83) and the probability of failure was linearly associated with the density of sheep in the area and rainfall on the previous day. A Monte Carlo modelling approach was used to estimate that, had sufficient samples been taken, 54% (95% confidence intervals 49-59%) of all private water supplies in England were likely to be unsatisfactory. These findings will be able to inform risk assessments of private water supplies prior to microbiological results being available.
AB - A passive surveillance system captured information on 34,904 microbiological samples from 11,233 private drinking water supplies within England as well as the associated constructional, climatic and environmental variables. Escherichia coli was detected in 6588 (18.87%) of samples and at least one positive sample was detected from 3638 (32.39%) of sites. However, this estimate of supplies failing to meet the European drinking water E. coli standard was probably an underestimate as the more samples taken per supply, the more likely the supply was to fail. A multivariable model of private water supplies data showed a strong seasonal impact, with samples between January and May being significantly less contaminated with E. coli than samples between June and December. Samples from springs (OR 2.5, CI 2.0-3.1) or surface waters (OR 2.4, CI 0.8-7.0) were more likely to fail than groundwater sources, as were supplies with no effective treatment (OR1.8, CI 1.5-2.3). Commercial supplies were less likely to fail than domestic supplies (OR 0.63, CI 0.48-0.83) and the probability of failure was linearly associated with the density of sheep in the area and rainfall on the previous day. A Monte Carlo modelling approach was used to estimate that, had sufficient samples been taken, 54% (95% confidence intervals 49-59%) of all private water supplies in England were likely to be unsatisfactory. These findings will be able to inform risk assessments of private water supplies prior to microbiological results being available.
U2 - 10.1016/j.watres.2009.02.035
DO - 10.1016/j.watres.2009.02.035
M3 - Article
VL - 43
SP - 2159
EP - 2168
JO - Water Research
JF - Water Research
SN - 0043-1354
IS - 8
ER -