A Prospective Study of Ventilation Rates and Illness Absence in California Office Buildings

Background – This study investigated the associations of ventilation rates (VRs), estimated from indoor CO2 concentrations, in offices with the amount of respiratory infections, illness absences, and building-related health symptoms in occupants. 

Methods – Office buildings were recruited from three California climate zones. In one or more study spaces within each building, real-time logging sensors measured carbon dioxide, temperature, and relative humidity for one year. Ventilation rates were estimated using daily peak CO2 levels, and also using an alternative metric. Data on occupants and health outcomes were collected through web-based surveys every three months. Multivariate models were used to assess relationships between metrics of ventilation rate or CO2 and occupant outcomes. For all outcomes, negative associations were hypothesized with VR metrics, and positive associations with CO2 metrics. 

Results – Difficulty recruiting buildings and low survey response limited sample size and study power. In 16 studied spaces within 9 office buildings, VRs were uniformly high over the year, from twice to over nine times the California office VR standard (7 L/s or 15 cfm per person). VR and CO2 metrics had no statistically significant relationships with occupant outcomes, except for a small significantly positive association of the alternative VR metric with respiratory illness-related absence, contrary to hypotheses. 

Conclusions– The very high time-averaged VRs in the California office buildings studied presumably resulted from “economizer cycles” bringing in large volumes of outdoor air; however, in almost all buildings even the estimated minimum VRs supplied (without the economizer) substantially exceeded the minimum required VR. These high VRs may explain the absence of hypothesized relationships with occupant outcomes. Among uniformly high VRs, little variation in contaminant concentration and occupant effects would be expected. These findings may provide initial evidence for an upper bound of the range of VRs within which increased VRs provide benefits in reducing illness absence.