Few studies have examined HEPA filtration in a classroom environment. This study aimed to assess if air cleaners using HEPA filters reduced classroom indoor air pollution exposure using a block randomized crossover trial in elementary schools in Los Angeles County that were already using existing HVAC systems with MERV 13 filtration due to the COVID-19 pandemic. In July 2022 before the placement of portable air cleaners, there was no statistically significant difference between PM concentrations in the treatment and control classrooms, indicating successful randomization with regard to baseline PM concentrations. Our findings show that HEPA classrooms had 39.9% lower average annual PM2.5 than non-HEPA classrooms (0.887 µg/m3 compared to 1.468 µg/m3) during the 2022–2023 school year. Outdoor PM2.5 infiltration was lower in the HEPA classrooms compared to the non-HEPA classrooms.

A previous study that used a similar study design was conducted from April to December 2021, Mengzhou city, Henan Province, China [15]. PM2.5 concentrations were measured in classrooms, living rooms, and outdoor environments during the study period. MicroPEM samplers were also used in the students’ living room. The living room and classroom interventions contributed to 42.31% and 21.34% reductions in personal PM2.5 exposure, respectively. Participants with living room and classroom air purification interventions had the lowest PM2.5 levels, with an average of 45.9 ± 44.4 µg/m3, followed by participants with only living room intervention (62.0 ± 51.5 µg/m3), participants with only classroom intervention (73.4 ± 54.1 µg/m3), and participants with no intervention (89.0 ± 61.4 µg/m3). PM2.5 levels in this study were much higher than levels in Los Angeles.

Another air filtration intervention study conducted from 2015 to 2020 in urban elementary schools located in the Northeastern United States utilized a factorial randomized trial with a four-arm design [16]. This study examined treatment with and without air cleaners with HEPA filters and with and without school-wide integrated pest management (IPM). The median PM2.5 classroom exposure at baseline was 5.5 µg/m3 with HEPA filtration and no IPM and 6.1 µg/m3 with control filtration. After the intervention, the median PM2.5 in classroom exposure was 3.1 µg/m3 and 5.3 µg/m3 for HEPA and control filtration, respectively. These PM levels are closer to those measured in the current study and the control classrooms were 70% higher than HEPA treatment classrooms, similar to what we observed (65%).

A third study also studied urban elementary schools from the Northeastern United States from 2013 to 2014, using a pilot randomized controlled trial [17]. Treatment classrooms received HEPA filtration while control classrooms had the filters replaced by a sound device to mimic the noise from the air filtration. Prior to randomization, baseline mean classroom levels of PM2.5 were 6.3 μg/m3 with no statistically significant differences between the control and treatment classrooms. In the control group, mean PM2.5 concentrations decreased from 6.4 μg/m3 at baseline to 4.8 μg/m3 and 5.0 μg/m3 at the first and second follow-up visits, respectively. In the treatment group, mean PM2.5 concentrations decreased from 6.2 μg/m3 at baseline to 2.4 μg/m3 and 2.6 μg/m3 at the first and second follow-up visits, respectively. The intervention group had greater reductions in PM2.5 levels compared to the control group, corresponding to a 49% and 42% reduction, respectively. This is similar to the reduction of 29.7% observed in the current study, even though the classroom PM2.5 levels we observed were much lower. Like the previous results, portable HEPA filter air cleaners were effective in improving short-term air quality in classroom environments. It is unclear whether the classrooms in any of the three previous air cleaner studies had pre-existing HVAC filtration.

Compliance limits the effectiveness of the air cleaners. Since our study used intention-to-treat analysis, its treatment effect estimates are more conservative (i.e., likely to be underestimated if there is imperfect compliance with treatment assignments, such as teachers installing their own HEPA filters in non-HEPA air purifiers). Although our study team noted that the majority of the air cleaners remained powered on with the correct assigned filters in place at the end of the first 6 months, when the filters were replaced, we could not monitor and ensure their proper use throughout the entire year, making ITT analysis the most appropriate choice. However, heterogeneity of results is more likely from mixing non-compliant and compliant data into the final analysis. Another limitation is that the outdoor monitors were installed prior to this study so we do not have calibration data associated with these monitors. The indoor monitors were collocated, and agreement was confirmed prior to placing them in the classrooms.

Interpretation of our results should be made with some consideration. During the study period, the ambient air quality was generally good, ranging from 7.8–10.1 µg/m³ with no wildfire smoke events, so we cannot assess the performance of the additional air filtration during extreme pollution episodes. Furthermore, indoor levels were very low due to the existing HVAC systems with MERV 13 filters in operation. While there are statistically significant reductions in PM levels with the portable HEPA air cleaners, we cannot say if these small reductions have any meaningful health benefits to the students or teachers.

In summary, our results are consistent with previous studies showing that, when properly maintained, portable air cleaners with HEPA filters are effective at removing PM2.5 from classrooms [15,16,17]. The schools in our study were using MERV 13 filters in their HVAC systems during the entire length of our intervention study due to the COVID-19 pandemic, which prevented a considerable about of infiltration of outdoor PM2.5 into the classrooms. However, infiltration was still lower in the treatment classrooms with HEPA filters compared to the control classrooms. Our findings provide support for the use of portable air cleaners with HEPA filters in classrooms to reduce PM, even in classrooms with existing HVAC air filtration.