Bioaerosol Data Distribution: Probability and Implications for Sampling in Evaluating Problematic Buildings

Abstract

Airborne fungal contamination in the indoor environment is a substantial contributor to indoor air quality (IAQ) problems, yet there are no set numerical standards by which to evaluate air sampling data. Intuitively appealing is the operational model that the indoor air should not be significantly different from the outdoor air, but determining what is "significant" as well as where to sample and how many samples to collect to determine significance have not been firmly established. The purpose of this study was to determine the number of samples and their locations necessary to determine significant differences in airborne fungi between the ambient and indoor environments. Sampling results from several hundred air samples for culturable fungi from various sites were used to derive a probability of detection in the outdoor air for problematic or "marker" fungal species. Under the assumption that indoor fungal growth results in an increase in the probability of detection for a given fungal species, mathematical probability dictates the number of samples necessary in the indoor (target zone) and in the outdoor (reference zone) air to demonstrate significance. Ultimately, it is the sparse distribution of the problematic species that drives the number of required samples to demonstrate a significant difference, which varies depending upon the level of significance desired. Therefore, the number of samples in each zone can be adjusted to reach a target difference in detection frequency, or an investigator can assess a sampling scheme to identify the differences in detection frequency that show significance.

Verifying Interpretive Criteria for Bioaerosol Data Using (Bootstrap) Monte Carlo Techniques

“…The single-sample BMC simulations that model the use of the respective descriptors for identification of "localized" contamination (for example, within a room) with a single sample further underscores unreliability with small sample sizes and/or single-sample comparisons. (7,8,22,23,32) More fundamentally, the fact that, in general, descriptors tested do not consistently improve with increasing sample size indicates inherent unreliability with the particular descriptor/criterion rather than limitations due to small samples.…”

“…(1−3,5−7,9−21) Consequently, a fundamental contributor to the wide variability in guidelines for bioaerosol data interpretation stems from differences in the nature of the data generated by a Journal of Occupational and Environmental Hygiene February 2008 85 particular sampling methodology, as well as how data may be truncated and/or processed to allow use of a particular descriptor. (5,16,22,23) As an example, an investigation comparing a suspect indoor environment with the outdoor air using differences in the ratio of phylloplane to nonphylloplane fungi (as discussed herein) must first define (limit) the types of fungi to be considered phylloplane and nonphylloplane fungi. Data for the designated species must then be combined into the respective grouping and each resulting group treated as a single contaminant.…”

Concentrations of airborne culturable bacteria in 100 large US office buildings from the BASE study

“…It is generally assumed that environmental measurements display either normal or lognormal distributions. However, little is known about distributions for bioaerosol measurements, especially indoor samples (Eudey et al., 1995; Luoma and Batterman, 2000; Spicer and Gangloff, 2003), because of the limited number of samples collected in previous studies or failure to identify the proper data distribution prior to data summary and interpretation. Sofuoglu and Moschandreas (2003) reported that the data for airborne bacteria from the first 41 BASE buildings (1994–1996) were log‐normally distributed, but the authors did not describe how they reached this conclusion and it was not replicated in the current analyses for the complete data on 100 buildings.…”

Establishing Site Specific Reference Levels for Fungi in Outdoor Air for Building Evaluation

“…The number of samples required in each zone can be reduced for fungal types detected more frequently in the outdoor air (i.e., greater than 0.05) and/or if an investigator can accept a greater degree of fungal amplification in the test environment. (31) An extension of the issue of bias in the timing of reference samples is the relative contribution of laboratory variability to the extreme nature of bioaerosol data. There are two aspects of the results from this study that shed some light on the issue.…”

“…Further, a format and approach used to objectively quantify differences between the indoor and outdoor environments for any given fungal type is dictated by the mathematical nature of the data. (2,7,11,12,(24)(25)(26)30,31) Culturable airborne fungal spore sampling at five building sites during 2002-2003 provided a bank of outdoor data to characterize the nature of fungal populations in the outdoor air during the day and to identify significant differences between the morning and afternoon hours. In turn, this provided an opportunity to gain insight into outdoor air sampling necessary to establish unbiased reference levels against which to compare the indoor environment.…”