Ventilation and laboratory confirmed acute respiratory infection (ARI) rates in college residence halls in College Park, Maryland

Abstract

Strategies to protect building occupants from the risk of acute respiratory infection (ARI) need to consider ventilation for its ability to dilute and remove indoor bioaerosols. Prior studies have described an association of increased self-reported colds and influenza-like symptoms with low ventilation but have not combined rigorous characterization of ventilation with assessment of laboratory confirmed infections. We report a study designed to fill this gap. We followed laboratory confirmed ARI rates and measured CO₂ concentrations for four months during the winter-spring of 2018 in two campus residence halls: (1) a high ventilation building (HVB) with a dedicated outdoor air system that supplies 100% of outside air to each dormitory room, and (2) a low ventilation building (LVB) that relies on infiltration as ventilation. We enrolled 11 volunteers for a total of 522 person-days in the HVB and 109 volunteers for 6069 person-days in the LVB, and tested upper-respiratory swabs from symptomatic cases and their close contacts for the presence of 44 pathogens using a molecular assay. We observed one ARI case in the HVB (0.70/person-year) and 47 in the LVB (2.83/person-year). Simultaneously, 154 CO₂ sensors distributed primarily in the dormitory rooms collected 668,390 useful data points from over 1 million recorded data points. Average and standard deviation of CO₂ concentrations were 1230 ppm and 408 ppm in the HVB, and 1492 ppm and 837 ppm in the LVB, respectively. Importantly, this study developed and calibrated multi-zone models for the HVB with 229 zones and 983 airflow paths, and for the LVB with 529 zones and 1836 airflow paths by using a subset of CO₂ data for model calibration. The models were used to calculate ventilation rates in the two buildings and potential for viral aerosol migration between rooms in the LVB. With doors and windows closed, the average ventilation rate was 12 L/s in the HVB dormitory rooms and 4 L/s in the LVB dormitory rooms. As a result, residents had on average 6.6 L/(s person) of outside air in the HVB and 2.3 L/(s person) in the LVB. LVB rooms located at the leeward side of the building had smaller average ventilation rates, as well as a somewhat higher ARI incidence rate and average CO₂ concentrations when compared to those values in the rooms located at the windward side of the building. Average ventilation rates in twenty LVB dormitory rooms increased from 2.3 L/s to 7.5 L/s by opening windows, 3.6 L/s by opening doors, and 8.8 L/s by opening both windows and doors. Therefore, opening both windows and doors in the LVB dormitory rooms can increase ventilation rates to the levels comparable to those in the HVB. But it can also have a negative effect on thermal comfort due to low outdoor temperatures. Simulation results identified an aerobiologic pathway from a room occupied by an index case of influenza A to a room occupied by a possible secondary case.

Keywords: Acute Respiratory Infection (ARI); Airborne infection control; College dormitory rooms; Infectious bioaerosols; Multi-zone model; Ventilation rate.

Fig. 1

(a) Spatial relationship of HVB and LVB, and (b) local wind conditions.

Fig. 2

Sensor deployment plan for the 5th floor at the north wing of the LVB.

Fig. 3

Multi-zone models for HVB and LVB: (a) floor plan for the 4th floor in the HVB; (b) the 4th floor in the HVB model; (c) floor plan for the 2nd floor in the LVB; (d) the 2nd floor in the LVB model. Green solid circle in (d) highlights the internal wall leakage pathway between the bathroom (RM4 in Fig. 12) and its adjacent bedroom (RM3 in Fig. 12). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Airflow paths in a room of the HVB: (a) closed airflow paths; and (b) open airflow paths.

Fig. 5

Measured CO₂ concentrations in the dormitory rooms in: (a) HVB; and (b) LVB, where the red diamonds denote the rooms with ARI cases.

Fig. 6

Hourly averaged CO₂ concentrations in dormitory rooms during: (a) weekdays in the HVB; (b) weekends in the HVB; (c) weekdays in the LVB; and (d) weekends in the LVB.

Fig. 7

Comparison of measured and model-calculated mean CO₂ concentrations during 00:00AM and 6:00AM on March 27th, 28th and 29th: (a) HVB; and (b) LVB.

Fig. 8

Room ventilation rates in: (a) HVB; and (b) LVB, where, SGL, DBL, TRP, and QDP denote single, double, triples, and quadruple rooms, respectively.

Fig. 9

Individual ventilation rates in: (a) HVB; and (b) LVB, where, SGL, DBL, TRP, and QDP denote single, double, triples, and quadruple rooms, respectively.

Fig. 10

Ventilation rate under different window/door opening scenarios in 20 LVB dormitory rooms, where horizontal lines represent averaged values.

Fig. 11

Influenza A virus concentrations in room RM3 and RM5 during February 5th.

Fig. 12

Concentration distribution of bioaerosols carrying influenza A viruses 8 h after the release in the source room RM5: (a) airflow map with a linear scaling; and (b) concentration distribution of influenza A virus with logarithmic scaling, where DBL denotes double room, BTH denotes bathroom, JAN denotes the janitor’s closet, * denotes the room with influenza infections, F1 and F2 are two exhaust fans with an airflow rate of 72.2 L/s and 24.1 L/s, respectively.