BACKGROUND: Elevated percent positivity (≥30%) of Legionella in hospital domestic water systems has been suggested as a metric for assessing the risk of health care-acquired Legionnaires' disease (LD). METHODS: We examined the validity of this metric by analyzing data from peer-reviewed studies containing reports of Legionella prevalence in hospital water (ie, percent positivity) and temporally matched reports of patients with health care-acquired LD. RESULTS: Our literature review identified 31 peer-reviewed publications reporting matched data. We abstracted a total of 206 data points, representing 119 hospitals, from these articles. We determined that the proposed 30% positivity metric has 59% sensitivity and 74% specificity (ie, a 41% false-negative rate and a 26% false-positive rate). These notable error rates could have significant implications, given that we identified 16 peer-reviewed articles and 6 government guidance documents that referenced the 30% positivity metric as a risk assessment tool. CONCLUSIONS: Environmental sampling of hospital water distribution systems for Legionella can be an important component of risk management for LD. However, the possible consequence of using a percent positivity metric with low sensitivity and specificity is that many hospitals might fail to mitigate when a true risk is present, or might unnecessarily allocate limited resources to deal with a negligible risk.
BACKGROUND: Elevated percent positivity (≥30%) of Legionella in hospital domestic water systems has been suggested as a metric for assessing the risk of health care-acquired Legionnaires' disease (LD). METHODS: We examined the validity of this metric by analyzing data from peer-reviewed studies containing reports of Legionella prevalence in hospital water (ie, percent positivity) and temporally matched reports of patients with health care-acquired LD. RESULTS: Our literature review identified 31 peer-reviewed publications reporting matched data. We abstracted a total of 206 data points, representing 119 hospitals, from these articles. We determined that the proposed 30% positivity metric has 59% sensitivity and 74% specificity (ie, a 41% false-negative rate and a 26% false-positive rate). These notable error rates could have significant implications, given that we identified 16 peer-reviewed articles and 6 government guidance documents that referenced the 30% positivity metric as a risk assessment tool. CONCLUSIONS: Environmental sampling of hospital water distribution systems for Legionella can be an important component of risk management for LD. However, the possible consequence of using a percent positivity metric with low sensitivity and specificity is that many hospitals might fail to mitigate when a true risk is present, or might unnecessarily allocate limited resources to deal with a negligible risk.