Differences in detection frequency as a bioaerosol data criterion for evaluating suspect fungal contamination

Abstract

With no health based numerical standards for evaluating airborne fungal spore data, sampling for environmental fungal spores is conducted by a variety of non-standardized methods to test the hypothesis of similarity between indoor and outdoor airborne fungal populations. Numerically based criteria, to include fixed fungal spore levels and various ratios, have been utilized to assist in interpreting data, but the lack of standards also precludes verification relative to a “true” airborne concentration. However, using the bootstrap version of Monte Carlo analysis (BMC), the false negative and false positive error rates of criteria can be approximated by the frequency indoor and outdoor fungal data sets are correctly or incorrectly determined to be similar. An alternative criterion for airborne fungal data analysis, using differences in frequency of detection (Δ f d) greater than the combined median, treats individually detected fungal types as separate contaminants; the mathematical description of differences between indoor and outdoor fungal populations is the calculated probability that Δ f d greater than actually observed could randomly occur. Culturable and spore trap sampling data at various sites from 2004 to 2008 provided a source of data by which to test the performance of Δ f d. Probability values estimated via BMC were close approximations to direct calculations based on Δ f d, and strongly support Δ f d as a criterion. As a building performance indicator, analysis via BMC demonstrates the appropriate measure for differences in “mold levels” is defined by the frequency with which a particular type of mold is detected relative to the combined median.