Risk-oriented approach to the quality assessment of water sources of drinking water supply

Introduction. Correct control and assessment of compliance of the composition and properties of drinking water sources with the established safety criteria are necessary conditions for ensuring the quality of life and health of the population in a situation of increasing exposure to chemical factors. Therefore, the most important is to increase the reliability of sanitary and hygienic conclusions. 

Material and methods. To achieve the goal of the work, there was used the Bayesian approach, methods for checking data homogeneity and evaluation of the measure of the statistical homogeneity. It is shown that the evaluation by Bayes method, requiring consideration of limiting factors, provides a correct solution to the problem of the implementation of the established sanitary and hygienic requirements. Representation of Bayesian relations in the form of estimates of the mixture of distributions allows taking into account the degree of statistical homogeneity of the combined samples, which significantly expands the scope of the method. Although the samples selected for consideration in the simple case can be considered homogeneous in the examples of the assessment of hygienic safety of water, it turned out that the transition from the binomial distribution law to a more strict normal one reduces the risk of accidental error of the desired conclusion. This provides the researcher with the opportunity to use, depending on the requirements for the accuracy of the assessment, different algorithms of a risk-based approach to the assessment of hygienic safety. 

Results. On the example of the study of the water quality in the Verkhny Isetsky Vodokanal recreational area and source of water in Northern part of Ekaterinburg the inclusion of the homogeneity of merged data was shown to allow “forgetting” the earlier information (e.g., 2009) and thus to give more weight to recent data (2010). It is also shown that in this case, the combined estimates with the degree of homogeneity of the data series are not worse than the Bayesian estimates. It is concluded that the lower the degree of uniformity, the faster the above “forgetting” occurs. And the risks of violation of hygienic requirements depend on the choice of the law of distribution of controlled indicators.

Conclusion. There was established the possibility of increasing the reliability of sanitary and hygienic conclusions by correctly combining the latest measurement information with outdated data, taking into account their gradual “forgetting”. The developed methods of the used for this risk-based approach can be widely applied to assess the hygienic safety of the population under the influence of a wide range of chemical factors.