Survey of unsatisfactory levels of airborne bacteria in air-conditioned offices

Those involved in property management are under pressure to alleviate the risk of disease spread through buildings, especially via the Heating, ventilating and air-conditioning (HVAC) system. To aid this, the level of indoor bacteria can be used as a referent to identify the performance of the HVAC system. Epistemic assessment of an AC office requires prior knowledge of the probable failure rate of offices in a region, which may be obtained by extensive measurements. This study proposes a statistical model to obtain a rapid estimation of the office probable failure rate over a region, that is, that fraction of offices with an average bacteria level above certain action levels. The model was based on transformation of small sample data to a geometric distribution using two analytical unbiased estimators: the weighted average and corrected sample variance of bacteria levels. These estimators could improve the accuracy of the estimates due to uncertainties of small sample sizes. Model parameters were determined from a dataset of levels in 290 Hong Kong offices (average 703 cfu m-3and 95% range 204-2418 cfu m-3), using Monte Carlo simulations. The model was validated with another district survey of 109 offices in nine major Hong Kong commercial districts (average indoor bacteria level 297 cfu m-3and 95% range was 69-1266 cfu m-3). The observed office failure rate of each district was compared with the corresponding model estimate and consistent results were found.