Figures & data
Table 1. Descriptive statistics of the 36 APIs detected >LOQ (limit of quantification) in at least 25% of samples. All API measurements are given in ng/L. n>LOQ = number of samples in which the API in question was detected (i.e. >LOQ). Low 95%CI = Lower limit of the 95% confidence interval. High 95%CI = Higher limit of the 95% confidence interval. <LOQ = LOQ = samples in which the API was not detected (i.e. <LOQ) were considered to be equal to LOQ in the statistical analysis. only >LOQ included = only samples were the API was detected (i.e. >LOQ) were included in the statistical analysis. P values <0.05 marked as bold.
Table 2. Descriptive statistics of ratios ‘pCure installed/without pCure installed’. Ratios were calculated for means or medians for each API, and analysis of ratios for different subsets of APIs are summarized in the table. APIs were divided into subsets depending on the number of samples in which APIs were detected (i.e. >LOQ (limit of quantification)); >50%, >75%, and >90% of samples. The same calculations were performed using only the APIs for which Pharem Biotech state that pCure is effective. Samples <LOQ were either considered as = LOQ, or excluded from the analysis (treated as missing values). Ratio = 1 indicates no effect, ratios <1 suggest that pCure installation decreases API levels and ratios >1 suggest that pCure installation increases API levels. p Par = p-value for t-test (a parametric test). p NPar = p-value for signed rank test (a non-parametric test).