AWWA WQTC55123

This paper discusses the sample sizes necessary to be able to detect differences between treatments designed to inactivate Cryptosporidium parvum. This analysis is a byproduct of a study initially designed and implemented to compare in-vitro cell culture based assays with animal infectivity assays for determining the viability and infectivity of Cryptosporidium oocysts. While the overall study compared the average percent infectivity of Cryptosporidium at various doses an important byproduct of the study is the variability associated with the two methods. The more variable the percent infectivity for a method, the larger the required sample size will be to detect a specific difference when testing for the effects of inactivation. For example, if a researcher is trying to detect a 20% difference in the inactivation of Cryptosporidium using two different treatments, then an assay with a high amount of variability between replicate analyses will require more replicates to detect the 20% difference than an assay with lower variability. The number of replicates required is a function of the variability, the difference that is meaningful, and the statistical certainty that the researcher desires in reporting that the treatments are different from one another. The function relating these concepts is summarized in a family of equations called power analyses. This paper reports on the findings of the variability between replicates observed for Cell culture PCR and animal infectivity studies. In order to compare cell culture and animal infectivity, a number of cell lines and isolates were used. The experiments were designed to facilitate comparisons between the detection methodology, the isolates and the cell lines. This paper focuses on the sample sizes necessary to detect differences of a certain magnitude using the two methods studied. The higher the variability the more replicates will be required to detect a specific difference. Smaller sample sizes are likely to result in an inability to detect differences where they really do occur. Experiments, therefore, are likely to be completed with no definitive results. These results will help researchers determine the number of replicates required in order to detect differences of a specific size. This paper reports the sample sizes required to detect particular differences in cell culture assays and it compares the variability and differences that can be detected for both cell culture and animal infectivity. Includes reference, table, figure.