Water Journal : Water Journal December 2011
microbiology refereed paper technical features 86 DECEMBER 2011 water Where the levels of micro-organisms in different water types are being compared, the minimum number of samples required depends upon the variation in microbial levels in each water type or location. Recognising that this is the case in advance of conducting a monitoring program is important and allows for consideration of a two-stage monitoring program. Using such an approach, investigation of temporal variation in water quality for each water type or location would precede the main monitoring program designed to quantify differences in microbial water quality between different water types/ sample locations. In this way, some planned monitoring programs may be revealed as overly ambitious based on the high variability in the parameters of interest, requiring reconsideration of the program objectives and budget amounts. Step 8: Reporting Data The usefulness of data is extended by the way it is reported. It is important that not only details of the method(s) employed are recorded (including operating characteristics such as recovery efficiency, method sensitivity, measurement uncertainty, equivalent sample volume analysed, etc) but that any assumptions made in designing the program are detailed. It is also important that characteristics of the water type are recorded (e.g. physico- chemical properties, etc) together with details about pollution inputs and the conditions under which sampling was undertaken (e.g. high-risk versus low-risk pollution events), where relevant. As budgetary constraints, lack of available methods for the micro-organism of interest, method shortfalls (e.g. in terms of sensitivity and/or inability to distinguish between viable and non-viable organisms) and/or limited data about the relationship between surrogates and target micro- organisms may all constitute reasons for compromises in monitoring program design, it is important that the rationale behind decisions are documented. This information is important to current and future users of data generated from the monitoring program (so that they can understand data limitations and the context), and to those seeking to address methodological shortfalls through targeted research. Proper reporting of data is also important if data are to be 'admitted' to a data repository, as some data repositories may have minimum data reporting requirements. Conclusion Water quality monitoring programs present multiple dilemmas for those charged with their design and implementation, especially in current times when economic imperatives can assume a primary role in deliberations. Compromises are inevitably required when designing water quality monitoring programs; however, it is important that the principle of generating high-quality data influences decision making, as does fitness for purpose. In addition, effective communication between those commissioning water quality testing and laboratory personnel is a key element in deciding the most appropriate method to be employed for monitoring and, also, in determining how methodological shortfalls and other data gaps might be best addressed through targeted research. The 'User's guide to microbiological program design and collection of exposure data' provides an aid to enhance communication. The Author Dr Joanne O'Toole (email: joanne.otoole@ monash.edu) has over 20 years' experience as a microbiologist and consultant for government and private industry organisations. She has experience in laboratory management and accreditation and has represented government and water industry bodies on Australian Standards committees. She was awarded her PhD in 2009 and a National Water Commission Fellowship in 2010. She currently holds an NHMRC Training Fellowship at the Melbourne School of Land and Environment, University of Melbourne. She also undertakes research at the Infectious Diseases Epidemiology Unit at Monash University. Acknowledgements This research was funded by the National Water Commission. Case studies highlighting dilemmas that may be encountered by water quality managers, consulting engineers and health regulators in monitoring program design using 'real world' examples are contained in the full NWC report which can be accessed from the NWC website (Waterlines Report #50). References ISO/IEC (2005): ISO/IEC 17025. General requirements for the competence of testing and calibration laboratories. Geneva, International Organization for Standardization. NATA (2009): Biological Testing. Supplementary requirements for accreditation. ISO/IEC 17025 Field Application Document. NATA. Step 7: Summary • What type of data analysis is planned (e.g. use in stochastic or deterministic modelling; pass/fail determination; comparative analysis)? • Where complex analysis is to be performed, consult a statistician for advice. • Can categorical data be employed or is continuous variable data required? • Is a pilot monitoring program required to obtain information about the likely prevalence and/or levels of the target micro-organism? • Determine minimum number of samples required. • Is the budget sufficient to allow for the monitoring of an adequate number of samples? If not, review the budget and/or scope of monitoring program. Step 8: Summary Record the following aspects of the monitoring program: • Purpose (including the context of data collection (e.g. absence of any data; limitations in existing data etc). • Method characteristics, including operating characteristics, sample volume analysed. • Assumptions made in the selection of micro-organism(s), method(s), sample numbers, sampling frequency. • Characteristics of water type being tested, pollution inputs, climatic conditions etc. • Raw data and not just final results; this includes raw counts in combination with the equivalent volume analysed and recovery efficiency data. Retain supporting evidence where applicable (e.g. reference to scientific literature and/or existing monitoring or pilot study data upon which monitoring program may be based).
Water Journal April 2012
Water Journal November 2011