Abstract
Context: To determine if radon mitigation is needed to protect occupants of multifamily housing, reliable testing procedures are needed. Yet, protocols on how many ground-contact housing units must be tested vary from 10% to 25% to 100%.
Objectives: To estimate the probability of failing to identify a building containing at least one unit with elevated radon level when all ground-contact units are not tested.
Design: Retrospective analysis of previously collected data from licensed (ie, certified) radon measurement professionals using hypergeometric and Monte Carlo statistical methods to estimate the confidence that there are no units with radon levels of 4 picoCuries/liter of air (pCi/L) or more based on various testing percentages.
Setting: Testing data were obtained from 29 US states for 7892 ground-contact units in 687 multifamily buildings, primarily 5 to 20 units per building.
Main Outcome Measure: Probability of failing to identify elevated radon levels in untested units.
Results: About 15% (n = 1163) of the units had radon levels of more than 4 pCi/L (the EPA action level); 59 units had more than 20 pCi/L (maximum of 96 pCi/L). For building sizes of 5 to 20 ground-contact units, the 2018 federal testing protocols that currently require testing of 10% and 25% of ground-contact units in each building failed to identify 47%-69% and 32%-46% of the units, respectively, depending on building size.
Conclusions: Measurement of 90% of the ground-contact units in buildings with 5 to 20 ground-contact units results in up to 4% of the units with elevated radon levels being missed. To achieve 95% confidence that no units in the building have radon levels of 4 pCi/L or more in buildings up to 20 units, 100% sampling is required. For the vast majority of multifamily buildings, all ground-contact units in multifamily buildings should be tested for radon.