Summary The ability to measure and track aerosols in the vicinity of patients with suspected or confirmed COVID‐19 is highly desirable. At present, there is no way to measure and track, in real time, the sizes, dispersion and dilution/disappearance of aerosols that are generated by airway manipulations such as mask ventilation; tracheal intubation; bronchoscopy; dental and gastro‐intestinal endoscopy procedures; or by vigorous breathing, coughing or exercise. We deployed low‐cost photoelectric sensors in five operating theatres between surgical cases. We measured and analysed dilution and exfiltration of aerosols we generated to evaluate air handling and dispersion under real‐world conditions. These data were used to develop a model of aerosol persistence. We found significant variation between different operating theatres. Equipment placement near air vents affects air flows, impacting aerosol movement and elimination patterns. Despite these impediments, air exchange in operating theatres is robust and prolonged fallow time before theatre turnover may not be necessary. Significant concentrations of aerosols are not seen in adjoining areas outside of the operating theatre. These models and dispersion rates can predict aerosol persistence in operating theatres and other clinical areas and potentially facilitate quantification of risk, with obvious and far‐reaching implications for designing, evaluating and confirming air handling in non‐medical environments.

The COVID-19 pandemic raised public awareness about airborne particulate matter (PM) due to the spread of infectious diseases via the respiratory route. The persistence of potentially infectious aerosols in public spaces and the spread of nosocomial infections in medical settings deserve careful investigation; however, a systematic approach characterizing the fate of aerosols in clinical environments has not been reported. This paper presents a methodology for mapping aerosol propagation using a low-cost PM sensor network in ICU and adjacent environments and the subsequent development of the data-driven zonal model. Mimicking aerosol generation by a patient, we generated trace NaCl aerosols and monitored their propagation in the environment. In positive (closed door) and neutral-pressure (open door) ICUs, up to 6% or 19%, respectively, of all PM escaped through the door gaps; however, the outside sensors did not register an aerosol spike in negative-pressure ICUs. The K-means clustering analysis of temporospatial aerosol concentration data suggests that ICU can be represented by three distinct zones: (1) near the aerosol source, (2) room periphery, and (3) outside the room. The data suggests two-phase plume behavior: dispersion of the original aerosol spike throughout the room, followed by an evacuation phase where "well-mixed" aerosol concentration decayed uniformly. Decay rates were calculated for positive, neutral, and negative pressure operations, with negative-pressure rooms clearing out nearly twice as fast. These decay trends closely followed the air exchange rates. This research demonstrates the methodology for aerosol monitoring in medical settings. This study is limited by a relatively small data set and is specific to single-occupancy ICU rooms. Future work needs to evaluate medical settings with high risks of infectious disease transmission.

The COVID-19 pandemic heightened public awareness about airborne particulate matter (PM) due to the spread of infectious diseases via aerosols. The persistence of potentially infectious aerosols in public spaces, particularly medical settings, deserves immediate investigation; however, a systematic approach to characterize the fate of aerosols in most clinical environments has not been reported. This paper presents a methodology for mapping aerosol propagation using a low-cost PM sensor network in ICU and adjacent environments and the subsequent development of the data-driven zonal model. Mimicking aerosol generation by a patient, we generated trace NaCl aerosols and monitored their propagation in the environment. In positive (closed door) and neutral-pressure (open door) ICUs, up to 6% or 19% respectively of all PM escaped through the door gaps, however, the outside sensors did not register an aerosol spike in negative-pressure ICUs. The K-means clustering analysis of temporospatial aerosol concentration data suggests that ICU can be represented by three distinct zones: (1) near the aerosol source, (2) room periphery, and (3) the outside region. These zones inform two-phase aerosol plume behavior: dispersion of the original aerosol spike throughout the room and an evacuation phase where "well-mixed" aerosol concentration in the ICU decayed uniformly. Decay rates were calculated in positive, neutral, and negative modes, with negative-pressure rooms clearing out nearly twice as fast. The aerosol concentration decay followed the trends in the air exchange rates. This research demonstrates the methodology for aerosol persistence monitoring in medical settings; however, it is limited by a relatively small data set and is specific to small-size ICU rooms. Future studies need to evaluate medical settings with high risks of infectious disease transmission and optimize hospital infrastructure.

The COVID-19 pandemic heightened public awareness about airborne particulate matter (PM) due to the spread of infectious diseases via aerosols. The persistence of potentially infectious aerosol in public spaces, particularly medical settings, deserves close investigation; however, approaches for rapidly parameterizing the temporospatial distribution of particles released by an infected individual have not been reported in literature. This paper presents a methodology for mapping the movement of aerosol plumes using a network of low-cost PM sensors in ICUs. Mimicking aerosol generation by a patient, we tracked aerosolized NaCl particles functioning as tracers for potentially infectious aerosols. In positive (closed door) and neutral-pressure (open door) ICUs, an aerosol spike was detected outside the room, with up to 6% or 19% of all PM escaping through the door gaps, respectively. The outside sensors registered no aerosol spike in negative-pressure ICUs. The K-means clustering analysis of temporospatial data suggests three distinct zones: (1) near the aerosol source, (2) room periphery, and (3) immediately outside the room. These zones inform two-phase aerosol plume behavior: dispersion of the original aerosol spike throughout the room, and evacuation phase where “well-mixed” PM decayed uniformly. Decay rates were calculated for 4 ICUs in positive, neutral, and negative mode, with negative modes decaying the fastest. This research demonstrates the methodology for aerosol persistence monitoring in medical settings; however, it is limited by a relatively small data set. Future studies need to evaluate medical settings with high risks of infectious disease, assess risks of airborne disease transmission, and optimize hospital infrastructure.Significance StatementAirborne infectious diseases, including COVID-19, are a major concern for patients and hospital staff. Here, we develop a systematic methodology for mapping the temporospatial distribution of aerosols in ICUs using a network of low-cost particulate matter (PM) sensors. Our method of analysis provides a perspective on the exfiltration efficiency of ICUs as well as the benefits of rooms that have a negative-pressure mode. Our methods could be extended to other public spaces with a high risk of infectious disease to optimize infrastructure and assess the risk of airborne disease.