Indoor penetration of ambient particulate pollution in a hospital maternity ward in Manila, Philippines: perspectives towards holistic city-level air quality management

Abstract

Several studies have analyzed and reported the relationship between particulate matter (PM) in the air and its adverse health effects, primarily on fetal development and subsequent early childhood. This study aims to understand how outdoor air made up of mainly PM, influences indoor air quality in a naturally ventilated maternity ward in an urban hospital setting. The data collection site in this study was the Dr Jose Fabella Memorial Hospital, a maternity hospital located in Manila, Philippines. Indoor and outdoor PM_2.5 levels from November 2021 to June 2022 were investigated. A strong positive correlation (r² ranging from 0.78 to 0.98) was observed between the daily outdoor and indoor PM levels. While the median concentrations were above the World Health Organization (WHO) air quality guidelines, they were below the Philippine National Ambient Air Quality Guideline Values (NAAQGV) at the time of data collection. These results underscore the importance of updating guideline values. Indoor-to-outdoor diurnal ratios (I/O), ranging from 0.77 to 1.33, with peak times (indoor-source-dominated) between 12:00 and 13:00 and trough times (outdoor-source-dominated) between 04:00 and 05:00, offered insight into the times of the day dominated by indoor versus outdoor sources and highlighted the need for continuous air monitoring while providing additional protection in indoor spaces, such as clear indoor air quality guidelines combined with indoor ventilation and filtration requirements. These results highlight the need for a holistic air quality management approach which focuses concurrently on both ambient and indoor air quality in healthcare facilities. Naturally ventilated hospitals must be included as a priority monitoring site, as they are a critical in improving air quality in the context of public health protection.

---

Environmental significance

This study underscores several important points: (1) the interconnectedness of indoor and outdoor air quality, (2) the vulnerability of sensitive populations such as patients in hospital maternity wards, who often have a lower socioeconomic status and may experience poor health compounded by cumulative effects, and (3) the need for robust and consistent air quality monitoring and updated air quality standards to protect public health. In alignment with the special theme of “Air Quality in Emerging Economic Regions”, this paper discusses the capacity to measure indoor and outdoor air quality using air sensors in developing regions. These regions often have sparse air quality monitoring data relative to their population density. The paper suggests coupling data from these sensors with interventions to gain a better understanding of effective air quality management practices.

Introduction

Air pollution remains the top global environmental risk for premature mortality, linked to 8.1 million deaths in 2021.¹ More than half (58%) of the premature deaths are due to ambient (outdoor) air pollution, while 38% of the mortality is due to household air pollution, caused by fine particulate matter (PM_2.5), which has aerodynamic diameters ≤2.5 micrometers. The specific short- and long-term health effects of air pollution exposure on the lungs, heart, tissues, and other body parts are well documented in the literature^2,3 together with the emphasis that infants and children, the elderly, and pregnant women face the highest share of the health burden.¹

Health risks related to air pollution are dependent on the exposure of the public to poor air quality. The majority of people spend time indoors, mostly at home, then in the workplace or at school. In Metro Manila, Philippines, activity data show that people spend 84% of their time in indoor environments.⁴ Ensuring good air quality is therefore of utmost importance, with guidelines provided by the World Health Organization (WHO),⁵ the US Occupational Safety and Health Administration (OSHA), and other similar agencies per country. Household air pollution is mainly linked with inefficient cookstoves and the use of solid fuels and biomass, but there are other sources of indoor air pollution that must be addressed.⁶ In addition, indoor air eventually gets contaminated with outdoor air, causing indoor air to contain both indoor- and outdoor-generated pollution.^4,6–8

Populations that are most vulnerable to air pollution health impacts can be found in health facilities such as hospitals, clinics, and health centers. While out-patient check-ups may only require a few hours, patients seeking consultations are likely already experiencing illnesses that can be aggravated by air pollution to a greater extent even at acute exposures. This study focuses on a birthing hospital because maternal exposure to PM_2.5 and ultrafine particles during pregnancy creates a window of vulnerability for fetal development and children's long-term health.⁹ Low birth weight has also been correlated with PM_2.5 exposure and depending on the period of gestation during which the mother was exposed, the odds ratio can be as high as 1.018 (highest in the first trimester) per μg m⁻³ increase in the mother's PM_2.5 exposure.¹⁰ Pregnancy-induced hypertensive disorders and preeclampsia are 57% and 31%, respectively, more likely to occur in women per 5 μg m⁻³ increments of ambient PM_2.5 they are exposed to (ref. 11). Infants exposed to particulate matter can also develop respiratory and neuropsychological health effects.^12–14

In the Philippines, air quality is monitored under the Philippine Clean Air Act. In line with this, air quality status reports are released by the Department of the Environment and Natural Resources (DENR). In the DENR's National Air Quality Status Reports, the annual nationwide and geometric PM_2.5 and PM₁₀ mean (per region) were reported. Manila City is in the National Capital Region (NCR), which has 32 air quality monitoring stations. Table 1 summarizes the annual nationwide and geometric PM_2.5 mean for the NCR and the Philippines from ref. 15 and 16.

Table 1 Annual average PM_2.5 and PM₁₀ concentrations for the National Capital Region (NCR) and Nationwide average PM_2.5 and PM₁₀ concentrations in the Philippines from 2013 to 2021 as reported in National Air Quality Status Reports. PM₁₀ concentration values were gathered from the Philippines' 2016–2018 report, while PM_2.5 concentrations were gathered from the 2021 report^15,16

Year	PM2.5		PM10
	NCR annual mean^a (μg per NCM)	Nationwide annual mean^a (μg per NCM)	NCR annual mean^a (μg per NCM)	Nationwide annual mean^a (μg per NCM)
a Philippine standards express concentration in μg per NCM, where “NCM” stands for normal cubic meter—the volume of dry gas which occupies a cubic meter measured at twenty five degrees Celsius (25°) at an absolute pressure equivalent to seven hundred and sixty (760) mm Hg.
2013	28.67	—	69	51
2014	35.25	35.25	51	52
2015	27.2	23.19	50	46
2016	26.85	20.36	46	39
2017	36.36	20.71	53	41
2018	20	20.36	47	39
2019	27	24.13	45	35
2020	—	14.78	—	27
2021	—	15.16	43	26

---

The Philippine national ambient air quality guideline values (NAAQGV) for criteria pollutants are defined and implemented by the Department of Environment and Natural Resources. The NAAQGVs are based on the WHO air quality guideline values and interim targets. Table 2 summarizes the WHO 2021 guideline values and the Philippine NAAQGVs available during data collection for this study.

Table 2 Guideline values for ambient air quality^15,16

Parameter	WHO (2021) guidelines		Philippine national ambient air quality guideline values
	Averaging time	Air quality guideline (μg m^−3)	Averaging time	NAAQGV (μg m^−3)
PM10	Annual	15	Annual	60
	24-hour	45	24-hour	150
PM2.5	Annual	5	Annual	25
	24-hour	15	24-hour	35

---

The majority of tertiary hospitals in the country have wards that are naturally ventilated (open windows and no air conditioning), highlighting the importance of understanding the contribution of outdoor air pollution to indoor air quality, as discussed in previous indoor air quality studies in the Philippine General Hospital wards.^17,18 These previous studies viewed air quality improvement in the context of hospital facility management and occupational safety, but as national and local governments expand air quality monitoring networks and develop measures towards air quality improvement, locations such as hospitals with vulnerable populations should be prioritized for air quality monitoring and action. In this study, we aim to understand the influence of outdoor air pollution on indoor air quality in a birthing hospital, specifically in its maternity ward, through parallel hourly PM_2.5 measurements in the naturally ventilated ward and right outside the hospital building. We investigated temporal trends as well as indoor–outdoor ratios and compared them to observed behaviors and activities. From these findings, we propose interventions to reduce indoor air pollution at the hospital management level and beyond. We also provide perspectives on how hospital air quality could be a model and should drive overall air quality management efforts at the national and local government levels.

Methods

Outdoor and indoor air monitoring

The establishment of a hybrid air quality monitoring network in the City of Manila was prioritized in the Asia Blue Skies Program to provide baseline information on the status of air in the city which can guide the clean air action planning process. Manila, with a population of over 1.85 million,¹⁹ only had one² continuous air quality monitoring station managed by the Department of Environment and Natural Resources – Environmental Management Bureau (DENR-EMB). Non-reference air sensors provide an opportunity to fill data gaps that cannot be covered by reference monitoring stations that require more investments, technical capacity, and operational requirements. This monitoring station was complemented by the project's air quality sensors which were installed in different locations in the city. Identifying health facilities as a hotspot of vulnerable populations, Dr Jose Fabella Memorial Hospital, a maternity hospital, was prioritized as a monitoring site for both outdoor and indoor PM_2.5.

Hourly PM_2.5 concentrations were measured outside and inside the hospital using Clarity Node-S sensors which are solar-powered IoT air quality monitoring systems. They measure PM through laser light scattering and transmit data through global cellular communications. The range for PM measurements is 0–1000 μg m⁻³, with a 1 μg m⁻³ resolution. Other technical specifications of the sensor can be accessed through https://www.clarity.io/products/clarity-node-s. Data from the sensors were accessed from the manufacturer's data dashboard and corrections were applied to ensure accuracy, as explained in the next section. For the outdoor measurements, a summary of measurement sites is given in Table 3.

Table 3 Location siting details for the outdoor and indoor sampling

Location	Device ID	Height from ground (m)	Start date	End date	Parameters
Outdoor (ER entrance area)	AXNYHPCH	3.5	November 17, 2021	June 29, 2022	T, RH, number, and mass concentration of PM2.5
Indoor (2nd floor, ward 4)	APRY1ZBH	14	December 29, 2021	September 5, 2024	T, RH, NO2, number and mass concentration of PM2.5

---

For the monitoring site in the hospital, the sensor was installed in a pole near the emergency room entrance, following standard siting protocols (i.e., minimum of 3 meters height from the ground, inlet clear of obstructions, far from exhausts, etc.) adjusted for exposure monitoring using portable sensors. In the indoor monitoring site, the main priority was the exposure of vulnerable populations, but additional considerations include proximity to an electricity outlet since the solar panel of the sensor will not work indoors, availability of a post suitable for sturdy sensor installation, and locations of electric fans for air not to be directly blown into the inlet. The maternity ward on the second floor was selected for the measurements. This ward had an area of 608 m², (6544 ft²), a ceiling height of 3.3 m (10.8 ft), and an average occupancy of 300 (combination of mothers and newborn infants). The ward has a high turnover rate—mothers and babies stay for only 24 hours after giving birth, except for those with complications or needing additional hospital care. This ward always had natural ventilation, i.e., open windows all the time and had multiple fans that blow air around. The air change rate (expressed in air changes per hour or ACH) was not measured due to sensitivities of patients, but Qian et al. estimated that the ACH for such conditions was between 19 and 69 ACH, in which case the indoor air is replaced once every 0.9 to 3 minutes.²⁰

Data corrections

The Clarity Node-S has been shown in the past to be strongly correlated with reference PM_2.5 but may experience a mean absolute error (MAE) of up to about 14 μg m⁻³ in an environment with mixed ambient sources.²¹ As part of the establishment of the air quality sensor network in the city, a collocation experiment was done with the DENR-EMB reference station in Mehan Garden, Manila, which measures PM mass concentrations through a Teledyne 640 analyzer (https://www.teledyne-api.com/products/particulate-instruments/t640). The sensor was installed in the railings of the reference station housing, around 1 meter from the inlet of the Teledyne instrument. The correction equation for this study was based on collocation data from August 29, 2020 to May 23, 2022 (total of 633 days), excluding periods where the reference instruments were under maintenance (total of 39 missing days). Sensor data were complete for the whole period, demonstrating high data capture above the 75% minimum required by the DENR-EMB. The comparison of Pearson coefficient (r²), root mean square error (RMSE) and MAE shows the need for a multiple linear regression (MLR) correction of raw sensor data using temperature and relative humidity data to ensure higher accuracy of air quality measurements. Comparing the uncorrected and corrected data, the r² improved from 0.64 to 0.81; the RMSE from 11.03 to 2.75; and the MAE from 8.20 to 2.15 μg m⁻³. The relative accuracy of the sensor measurements with respect to the reference station ranged from 91.6 to 97% for the whole duration of the collocation study. The MLR correction was thus applied to all raw sensor measurements prior to use. All data presented herein have already been corrected using R; all plots shown were created using Wolfram Mathematica.

Results

Daily and diurnal trends

Ambient daily 24-hour PM_2.5 averages are monitored by the DENR-EMB and checked with respect to the Philippine National Ambient Air Quality Guideline Values (NAAQGV) stipulated in the Philippine Clean Air Act of 1999, which also states that the basis of the guidelines is the WHO Air Quality Guidelines (WHO AQG). The daily 24-hour PM_2.5 averages from January to June 2022 in the hospital indoor and outdoor sites are shown in Fig. 1, with respect to 24-hour PM_2.5 NAAQGV and WHO AQG. Blue gridlines indicate the NAAQGV (light blue) and WHO standards (dark blue). While only 2.2% and 3.9% of the indoor and outdoor daily 24-hour average values for indoor and outdoor, respectively, exceeded the Philippine NAAQGV standard of 35 μg m⁻³ PM_2.5, comparing the same daily indoor and outdoor daily 24-hour average values to the WHO standard of 15 μg m⁻³ PM_2.5 increases that number to 69.4% and 73.3%, for indoor and outdoor respectively. In other words, only 30.6% and 26.7% of the indoor and outdoor PM_2.5 values are within the threshold set by the WHO. In a separate study in another hospital in Manila, 24-hour indoor PM_2.5 averages in naturally ventilated indoor locations were observed to be between 24.6 and 32.8 μg m⁻³, slightly lower than the NAAQGV standard.¹⁸ In 2020, a study by Enoveso observed indoor PM_2.5 in an unnamed hospital in Manila surrounded by mobile pollution sources and reported mean indoor PM_2.5 concentrations ranging from 21.27 to 31.11 μg m⁻³.¹⁷ Because these studies reported 24-hour PM_2.5 averages, the instantaneous concentrations at peak indoor PM_2.5 concentrations during the daytime, such as outliers in Fig. 3, might be missed.

Fig. 1 Daily 24-hour PM_2.5 average for all months. The solid horizontal blue gridline is the WHO AQG (15 μg m⁻³ PM_2.5); the dashed horizontal blue line is the Philippine (PH) NAAQGV at the time of the study (35 μg m⁻³ PM_2.5). The shading signifies a 95% confidence interval around the mean. On an hourly comparison, there is more scatter than averaging the PM_2.5 values in a day, i.e., the average indoor PM_2.5 concentrations measured in a day agree well with the average measured outdoor PM_2.5.

Fig. 2 shows day-of-the-week, diurnal, and diurnal-per-month averages for the campaign. For the diurnal-per-month, only January to June data are shown. There is no stark difference between weekends and weekdays, but Wednesday and Saturday show peak average PM_2.5 concentrations for indoor and outdoor averages. On the other hand, the diurnal shows multiple peak periods—for the indoors, 06:00 to 09:00 and 09:00 to 12:00 and a smaller peak around 19:00. For the outdoor diurnal, the morning peaks are the same, albeit opposite in trend, and another peak later in the day about the same time as the indoor peak. These peaks are also apparent in the per-month diurnal averages (Fig. 2(c)), albeit with less of a difference between the AM and PM peaks for both May and June. Separating the diurnals by months (ESI Fig. 7†) shows a similar diurnal trend—all indoor peaks at or close to noon are high for all months (January–June) and a morning diurnal that could be characteristic of outdoor PM emissions caused by vehicle traffic is also observed. A peak is also evident as an afternoon peak, albeit slightly lower than the morning peak. This is supported by a spatiotemporal analysis of human mobility in 2018 by Liu, K., where the author mapped hourly density of human mobility in Manila from 00:00 to 24:00 and noted inflow and outflow trips in four time periods: (1) 00:00–08:00; (2) 08:00–12:00; (3) 12:00–16:00; and (4) 16:00–24:00. In this study, the population group that goes to school traveled more in the 00:00–08:00, while older groups travel at 08:00–12:00.²² Moreover, the peak-time of mobility was noted to occur at 00:00–08:00 (morning commute) and 16:00–24:00 (commute to return home).²²

Fig. 2 (a) Day-of-the-week and (b) diurnal average PM_2.5 concentration for the whole campaign, which lasted from November 2021 to June 2022. (c) The diurnal average for all months (January to June 2022). November and December are not shown since they do not have a consistent indoor–outdoor pair of measurements. All bars represent 95% confidence intervals around the mean.

A more robust way of looking at the PM_2.5 measurements is by looking at the five-number summary (box-and-whisker plots), as shown in Fig. 3. A five-number summary is a more informative way of looking at the averages without being skewed by the presence of outliers; however, looking at the extreme points is also warranted. Because of the lack of contextual information and knowledge of events surrounding the extreme concentrations during this period of time, an inference cannot be made about the possible sources of pollution during these very high periods. This calls for a dedicated framework where activity data are collected in conjunction with air quality monitoring, so that interventions can be designed for extreme events and instantaneous periods of high PM_2.5 concentrations. Fig. 3 shows that all median values are at or above the WHO's 15 μg m⁻³ PM_2.5 AQG but below the Philippine NAAQGVof 35 μg m⁻³ PM_2.5. The 75% daily quantile is also below the Philippine NAAQGV. However, all outliers and far outliers are above the NAAQGV. It is important to note that Fig. 3 shows all hourly points (N = 5365), while Fig. 1 show a daily 24-hour PM_2.5. Thus, instances of exceedances are higher on an hourly basis. A 24-hour average is not sufficient to show acute personal exposure, especially in highly polluted and dynamic areas such as densely populated urban areas; thus, it is recommended that PM_2.5 be reported at a more granular frequency than diurnal averages.

Fig. 3 Five-number summary (box-and-whisker) plots for day-of-the-week indoor and outdoor PM_2.5. The solid horizontal blue gridline is the WHO AQG (15 μg m⁻³ PM_2.5) and the dashed blue horizontal line is the Philippine (PH) NAAQGV (35 μg m⁻³ PM_2.5), which were the applicable standards at the time of data collection. The black and gray points are outliers and far outliers, respectively.

Indoor–outdoor PM_2.5 relationships

Linear fits using ordinary least squares regression were found to explain the indoor–outdoor relationship well, as exhibited in Fig. 4(a) and (b); the fit (solid black) line lies close to the 1 : 1 (dashed) line. A previous study by Lu et al.²³ showed an exponential relationship between indoors and outdoors; however, this was not observed here. The difference in the nature of the relationship between indoor and outdoor between the Lu et al. study and this study may be explained by the very high exchange rate in this study, as opposed to a lag of the outdoor to indoor entrainment exhibited in Lu et al. Another reason may be the sampling height difference.

Fig. 4 (a) Hourly and (b) daily average PM_2.5 data comparison for each month.

Fig. 4 shows a strong correlation between the indoor and outdoor mass concentrations across the whole mass concentration range. However, looking at the indoor-to-outdoor mass concentration ratio (I/O) would give insight into the relative contributions of indoor versus outdoor sources. Fig. 5(a) and (c) show that the I/O ratio is notably highest at noontime; there is a consistent diurnal trend regardless of the day of the week (e.g., weekend vs. weekday). I/O trends are not expected to follow a diurnal pattern, but the presence may indicate a trend related to external factors such as human activities or circadian patterns. Moreover, the range of I/O ratios usually observed is usually greater than 0.5 but less than 3, indicating the strength, level of relationship, and interdependencies of indoor and outdoor PM.^24–27 In the Manila hospital study by Lomboy et al., previously described above, the researchers reported a similar maximum I/O (1.34) and much lower minimum I/O (0.26) in a similar indoor environment (i.e., a naturally ventilated hospital specialty ward in the department of pediatrics; the hospital was situated in a similar ambient environment).¹⁸

Fig. 5 (a) Indoor/outdoor (I/O) ratio, (b) diurnal average and (c) diurnal I/O by day-of-the-week. The shading on (b) signifies the 95% confidence interval around the mean.

During the daytime (08:00 to 14:00), the I/O is more than 1, indicating the strength of possible indoor sources; conversely, from 15:00 until before 08:00, the I/O is less than 1, indicating the strength of outdoor sources, which may be counterintuitive to human activity patterns outdoors but may be explained by human activities indoors that were either observed, reported by hospital staff during this study, or reported in the literature in other studies.^17,18 During the daytime, there is an increase in human activity indoors (e.g., walking) that may contribute to the resuspension of indoor dust and thus contribute to PM_2.5 directly. There could also be an increase in gaseous emissions indoors (volatile organic compounds or VOCs from consumer products that the persons wear, skin emissions, cleaning products, etc.^28–33) that can react with outdoor ozone coming in and may contribute to secondary PM formation; however, because this study was limited to measuring PM, and no measurements of VOCs and ozone were made, the nature of secondary sources of PM cannot be determined and can only be hypothesized based on indoor sources studied in the past as well as anecdotal knowledge of human activities in the hospital. With that said, it has been shown that the entrainment of ozone, even at low concentrations (less than ten ppb), can generate up to 40 μg m⁻³ for particles in the 1.2–500 nm size range.³⁴ However, with very high ACH, it seems more likely that the primary indoor source would be mechanical resuspension,^35–39 unless the indoor VOC sources are in the hundreds of ppb, in which case PM can form quickly (e.g., in the order of minutes).

Human activities during the daytime at or surrounding the hospital that can contribute to resuspension and mechanical generation of dust include activities within hospital wards; conduct of medical rounds; construction and repair activities around the area; dusting or changing of beddings that can resuspend bioaerosols; delivery of food to patients using trolleys; collection, movement, and transport of garbage before noon. On the other hand, there are also specific activities that can generate particles smaller than PM_2.5, including fire incidences in nearby areas; phototherapy (usage of blue light); usage of UV light for sanitization of areas during the height of the COVID-19 pandemic; and the use of mist sprays.

Since Dr Jose Fabella Memorial Hospital operates 24 hours daily, high mobility in the facility even beyond midnight may be caused by emergency medical cases or patient admissions. Healthcare workers with different shifting schedules may also contribute to this. Moreover, there are nearby public utility vehicle terminals with jeepneys and tricycles that operate even during late hours and as early as 04:00. The surrounding residential and commercial establishments could also contribute to the increased pedestrian and vehicular traffic around the hospital vicinity. Peak measurements for PM_2.5 which start at 06:00 can also be attributed to residual nighttime pollutant concentrations, low planetary boundary layer height, and increase in vehicular traffic emissions as part of the morning rush.⁴⁰

Discussion

Failing to address the effects of air pollution can lead to a broad range of health complications for affected populations. As such, actions need to be taken, and interventions informed by observed air pollution trends should be designed to improve the air quality and health of the general public.

For example, many studies have suggested air filters improve respiratory health and indoor air quality.^41–45 These can be installed within a building's ventilation or centralized air conditioning systems or using standalone air purifiers. A hospital's heating, ventilation, and air conditioning (HVAC) system, if it exists, can be optimized by regularly inspecting the area, exhausts, and vents. Localized fans with filters, e.g., Corsi-Rosenthal boxes,⁴⁶ may filter the resuspended PM_2.5 from human activity. Other physical adsorbents may also be applicable (a few studies have explored other air-purifying technologies, such as electrostatic precipitators and negative ion generators, both of which operate by producing ions that bind to particles, promoting deposition, e.g., ref. 47 and 48) Because the addition of fans can also work oppositely, i.e., the presence of wall fans in the hospital maternity ward can introduce outdoor air to the indoors faster, periods of high outdoor PM (I/O less than 1, e.g., midnight to around 08:00 in Fig. 5(a) and (c)) can lead to a quick entrainment of outdoor to indoor PM; thus, it is important to ensure that an efficient filtration system for any air intake (either natural or mechanical ventilation) is present. The hospital's outdoor environment has mobile sources of air pollutants, e.g., tricycles idling or waiting for passengers, and area sources, e.g., eateries (hole-in-the-wall restaurants) along one side of the road. Indoor exposure to outdoor pollution may be reduced by closing the windows, especially when the I/O is less than 1 (i.e., from 15:00 until before 08:00) when the outdoor source overwhelms the indoor sources. In an urban area like the Dr Jose Fabella Memorial Hospital location, vehicle emissions may be a major source; thus, reducing vehicular emissions, such as reducing NO_x by filtration or addressing traffic flow management within the hospital vicinity may be greatly beneficial in improving the quality of air that penetrates the indoor environment. For example, mapping out the route of emergency vehicles to minimize idling near wards or identifying peak vehicular hours in the nearby roads will help identify and further characterize the outdoor emissions that make their way to the indoors.

When outdoor PM is low, indoor PM is also reduced when I/O is below 1, so interventions such as closing the windows or employing filtration systems may be less necessary. However, during hours of high indoor activity, the indoor PM_2.5 compared to the outdoor PM_2.5 slightly increases. For such scenarios, one intervention may not be enough. To reduce outdoor PM_2.5 coming indoors, windows must be closed, and outdoor air brought to the inside must go through a high-efficiency filter. Staying indoors and closing windows also protect against air pollution. These levels of protection vary with the type of indoor environment depending on the opening of windows, the use of air conditioners and air purifiers, and the building's location, age, and condition. A study by Lin et al. (2009) showed that higher concentrations of indoor PM_2.5, blood pressure, and heart rates occur with open windows than with closed ones.⁴⁹ Another study by Reisen et al. (2019) reached a similar conclusion, proving that leaving windows and doors closed leads to reduced PM_2.5 concentrations.⁵⁰ In recent years, improvements to indoor air (such as construction of a building with better air conditioning and air filtration systems) have been underway in Fabella hospital as an intervention to address this issue.

During sensitive hospital operations and procedures (e.g., surgeries, chemotherapy, etc.), proper ventilation and an air-cleaning device that removes radicals or ozone are recommended, but only clear the air using adsorbents or HEPA filters. Ventilation ensures good air recirculation and pollutants are not localized within a person's breathing space. In addition, a filtration or air cleaning device ensures that even if polluted outdoor air is brought inside, it gets filtered through and adsorbed rather than recirculated.

Frequent floor cleaning with non-PM forming agents is recommended. Lomboy et al. (2015) reported that in a similar hospital in Manila, sodium hypochlorite was used as a cleaning agent.¹⁸ Chlorine-based cleaners are known to initiate chlorine radical chemistry that can quickly form secondary PM.^51,52 Chemical-free floor scrubbing technologies and vacuums equipped with HEPA filters are also recommended for cleansing and sanitizing.⁵³ In many cases, technologies that employ UV-C irradiation (such as disinfecting lamps) are typically used for germicidal disinfection and decontamination; however, it is recommended that this is used in conjunction with other disinfection and air filtration methods to reduce unwanted prolonged exposure to UV-C.⁵⁴ It is also recommended to use only a specific wavelength (e.g., 222 or 254 nm) instead of a wide-range UV lamp to prevent unwanted radical reactions that can lead to atmospheric oxidation and secondary PM generation.

Other interventions that can help improve the air quality inside maternity wards, such as in this study, include adding indoor and outdoor air quality monitoring systems and changing the hospital layout so that the rooms of the vulnerable patient groups are away from the emission sources and the main roads. However, as noted in the results above, fans may increase the recirculation of polluted air indoors; thus, it is recommended to complement this with efficient filters or other materials that adsorb PM and gaseous indoor pollutants. Further investigation of interventions, such as the effect of air conditioning on air quality, is also warranted—for example, the hospital in this study had a 12-person air-conditioned ward besides the ward without air conditioning.

These interventions were formally submitted to the hospital administration in September 2023 for their consideration. In April 2024, the new Dr Jose Fabella Memorial Hospital became operational. Compared to the old facility which had big wards with more than 50 beds each as described in this study, the new hospital has mostly air-conditioned rooms with only an eight-bed capacity. In most tertiary hospitals in the country that have yet to be improved, the indoor–outdoor air quality conditions would likely be similar to that of the old Fabella Hospital. The results and recommendations from this study would thus still be relevant in these health facilities to reduce exposure of patients (especially pregnant women who just gave birth, and newborn babies) and healthcare workers from poor air quality, while emphasizing the message that air quality management requires a holistic approach that involves addressing outdoor and indoor air pollution sources in parallel.

Data availability

Datasets generated during and/or analyzed during the current study are not publicly available due to ongoing government collaborations, but may be available from the authors upon reasonable request.

Author contributions

CMFR led the writing and performed the literature review and data analysis; EGT collected, cleaned and analyzed the data and provided the methodology; DLB and MPB contributed material to the manuscript and provided comments. All authors read and approved the final manuscript.

Conflicts of interest

CMFR is affiliated with OpenAQ, an air quality nonprofit organization funded by philanthropic organizations, corporate sponsorships and subcontracts funded by public and private partners. OpenAQ has a non-financial data-sharing partnership with Clarity Movement Co., but this study was done independent of OpenAQ's activities. There are no other competing interests to disclose.

Acknowledgements

This study was conducted as part of the Asia Blue Skies Program activities in Manila City, Philippines, and is supported by 3M Globalgiving. The research and air quality monitoring efforts were done in partnership with the Manila City Government through the Department of Public Services and the Dr Jose Fabella Memorial Hospital management.

References

1. Health Effects Institute, State of Global Air 2024, Boston, MA, 2024, Available from: https://www.stateofglobalair.org/resources/report/state-global-air-report-2024.
2. R. Gheissari, J. Liao, E. Garcia, N. Pavlovic, F. D. Gilliland and A. H. Xiang, et al., Health Outcomes in Children Associated with Prenatal and Early-Life Exposures to Air Pollution: A Narrative Review, Toxics, 2022, 10(8), 458.
3. X. Lei, Y. Zhang, Z. Wang, Z. Lu, C. Pan and S. Zhang, et al., Effects of prenatal exposure to PM2.5 and its composition on cognitive and motor functions in children at 12 months of age: The Shanghai Birth Cohort Study, Environ. Int., 2022, 170, 107597.
4. S. Kecorius, E. G. Tamayo, M. C. Galvez, L. Madueño, G. Betito and M. Gonzaga-Cayetano, et al., Activity Pattern of School/University Tenants and their Family Members in Metro Manila – Philippines, Aerosol Air Qual. Res., 2018, 18, 2412–2419.
5. World Health Organization, WHO Global Air Quality Guidelines. Particulate Matter (PM2.5 and PM 10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide, 2021, Available from: https://www.who.int/publications/i/item/9789240034228.
6. US EPA. Indoor Air Quality, 2024, Available from: https://www.epa.gov/report-environment/indoor-air-quality.
7. R. W. Allen, S. D. Adar, E. Avol, M. Cohen, C. L. Curl and T. Larson, et al., Modeling the residential infiltration of outdoor PM 2.5 in the multi-ethnic study of atherosclerosis and air pollution (Mesa Air), Environ. Health Perspect., 2012, 120(6), 824–830.
8. P. M. Shrestha, J. L. Humphrey, E. J. Carlton, J. L. Adgate, K. E. Barton and E. D. Root, et al., Impact of Outdoor Air Pollution on Indoor Air Quality in Low-Income Homes during Wildfire Seasons, Int. J. Environ. Res. Publ. Health, 2019, 16(19), 3535.
9. N. M. Johnson, A. R. Hoffmann, J. C. Behlen, C. Lau, D. Pendleton and N. Harvey, et al., Air pollution and children’s health—a review of adverse effects associated with prenatal exposure from fine to ultrafine particulate matter, Environ. Health Prev. Med., 2021, 26(1), 72.
10. G. Harris, W. D. Thompson, E. Fitzgerald and D. Wartenberg, The association of PM2.5 with full term low birth weight at different spatial scales, Environ. Res., 2014, 134, 427–434.
11. M. Pedersen, L. Stayner, R. Slama, M. Sørensen, F. Figueras and M. J. Nieuwenhuijsen, et al., Ambient Air Pollution and Pregnancy-Induced Hypertensive Disorders, Hypertension, 2014, 64(3), 494–500.
12. V. L. Tapia, B. V. Vasquez, B. Vu, Y. Liu, K. Steenland and G. F. Gonzales, Association between maternal exposure to particulate matter (PM2.5) and adverse pregnancy outcomes in Lima, Peru, J. Expo. Sci. Environ. Epidemiol., 2020, 30(4), 689–697.
13. C. Wojtyla, K. Zielinska, P. Wojtyla-Buciora and G. Panek, Prenatal Fine Particulate Matter (PM2.5) Exposure and Pregnancy Outcomes—Analysis of Term Pregnancies in Poland, Int. J. Environ. Res. Publ. Health, 2020, 17(16), 5820.
14. M. Veras, D. Waked and P. Saldiva, Safe in the womb? Effects of air pollution to the unborn child and neonates, J. Pediatr., 2022, 98, 27–31.
15. Department of Environment and Natural Resources, National Air Quality Status Report 2016–2018, Quezon City, Philippines, 2020, Available from: https://air.emb.gov.ph/wp-content/uploads/2021/07/NAQSR-2016-2018-FV.pdf.
16. Department of Environment and Natural Resources, National Air Quality Status Report, Department of Environmental and Natural Resources, Central Office, Quezon City, Philippines, 2023, Available from: https://air.emb.gov.ph/wp-content/uploads/2025/01/NAQSR-2021-v2.pdf.
17. R. A. D. Enoveso, M. T. Cruz, M. F. T. Lomboy, G. V. Dalmacion, R. R. Quizon and V. B. Molina, et al., A look at an urban tertiary hospital's indoor air pollutants using source apportionment of PM2.5, Philippine J. Health Res. Dev., 2020, 24, 1.
18. M. F. T. C. Lomboy, L. L. Quirit, V. B. Molina, G. V. Dalmacion, J. D. Schwartz and H. H. Suh, et al., Characterization of particulate matter 2.5 in an urban tertiary care hospital in the Philippines, Build. Environ., 2015, 92, 432–439.
19. Philippine Statistics Authority, 2020 Census of Population and Housing Results, City of Manila, 2021, Available from: https://rssoncr.psa.gov.ph/content/2020-census-population-and-housing-results-city-manila.
20. H. Qian, Y. Li, W. H. Seto, P. Ching, W. H. Ching and H. Q. Sun, Natural ventilation for reducing airborne infection in hospitals, Build. Environ., 2010, 45(3), 559–565.
21. G. Raheja, J. Nimo, E. K. E. Appoh, B. Essien, M. Sunu and J. Nyante, et al., Low-Cost Sensor Performance Intercomparison, Correction Factor Development, and 2+ Years of Ambient PM2.5 Monitoring in Accra, Ghana, Environ. Sci. Technol., 2023, 57(29), 10708–10720.
22. K. Liu, Spatiotemporal Analysis of Human Mobility in Manila Metropolitan Area with Person-Trip Data, Urban Sci., 2018, 2, 3.
23. C. Lu, H. Xu, W. Meng, W. Hou, W. Zhang and G. Shen, et al., A novel model for regional indoor PM2.5 quantification with both external and internal contributions included, Environ. Int., 2020, 145, 106124.
24. D. Massey, A. Kulshrestha, J. Masih and A. Taneja, Seasonal trends of PM10, PM5.0, PM2.5 & PM1.0 in indoor and outdoor environments of residential homes located in North-Central India, Build. Environ., 2012, 47, 223–231.
25. C. Monn, A. Fuchs, D. Hiigger, M. Junker, D. Kogelschatz and N. Roth, Particulate matter less than 10 pm (PM,,) and fine particles less than 2.5 pm (PM&): relationships between indoor, outdoor and personal concentrations, Sci. Total Environ., 1997, 208(1–2), 15–21.
26. J. D. Sinclair, L. A. Psota-Kelty and C. J. Weschler, Indoor/outdoor ratios and indoor surface accumulations of ionic substances at Newark, New Jersey, Atmos. Environ., 1988, 22(3), 461–469.
27. M. Viana, I. Rivas, X. Querol, A. Alastuey, J. Sunyer and M. Álvarez-Pedrerol, et al., Indoor/outdoor relationships and mass closure of quasi-ultrafine, accumulation and coarse particles in Barcelona schools, Atmos. Chem. Phys., 2014, 14(9), 4459–4472.
28. M. Kruza and N. Carslaw, How do breath and skin emissions impact indoor air chemistry?, Indoor Air, 2019, 29(3), 369–379.
29. N. Wang, L. Ernle, G. Bekö, P. Wargocki and J. Williams, Emission Rates of Volatile Organic Compounds from Humans, Environ. Sci. Technol., 2022, 56(8), 4838–4848.
30. P. Mochalski, J. King, K. Unterkofler, H. Hinterhuber and A. Amann, Emission rates of selected volatile organic compounds from skin of healthy volunteers, J. Chromatogr. B, 2014, 959, 62–70.
31. Z. Zou, J. He and X. Yang, An experimental method for measuring VOC emissions from individual human whole-body skin under controlled conditions, Build. Environ., 2020, 181, 107137.
32. Z. Zou and X. Yang, Skin volatile organic compound emissions from 14 healthy young adults under controlled conditions, Build. Environ., 2022, 222, 109416.
33. Z. Zou and X. Yang, Volatile organic compound emissions from the human body: Decoupling and comparison between whole-body skin and breath emissions, Build. Environ., 2022, 226, 109713.
34. C. M. F. Rosales, J. Jiang, A. Lahib, B. P. Bottorff, E. K. Reidy and V. Kumar, et al., Chemistry and human exposure implications of secondary organic aerosol production from indoor terpene ozonolysis, Sci. Adv., 2022, 8(8), eabj9156.
35. A. Benabed, A. Boulbair and K. Limam, Experimental study of the human walking-induced fine and ultrafine particle resuspension in a test chamber, Build. Environ., 2020, 171, 106655.
36. A. Benabed and A. Boulbair, PM10, PM2.5, PM1, and PM0.1 resuspension due to human walking, Air Qual. Atmos. Health, 2022, 15(9), 1547–1556.
37. A. R. Ferro, R. J. Kopperud and L. M. Hildemann, Source Strengths for Indoor Human Activities that Resuspend Particulate Matter, Environ. Sci. Technol., 2004, 38(6), 1759–1764.
38. N. Serfozo, S. E. Chatoutsidou and M. Lazaridis, The effect of particle resuspension during walking activity to PM 10 mass and number concentrations in an indoor microenvironment, Build. Environ., 2014, 82, 180–189.
39. E. Stratigou, S. Dusanter, J. Brito and V. Riffault, Investigation of PM10, PM2.5, PM1 in an unoccupied airflow-controlled room: How reliable to neglect resuspension and assume unreactive particles?, Build. Environ., 2020, 186, 107357.
40. M. T. Cruz, J. B. Simpas, A. Sorooshian, G. Betito, M. O. L. Cambaliza and J. T. Collado, et al., Impacts of regional wind circulations on aerosol pollution and planetary boundary layer structure in Metro Manila, Philippines, Atmos. Environ., 2023, 293, 119455.
41. E. Cheek, V. Guercio, C. Shrubsole and S. Dimitroupoulou, Portable air purification: Review of impacts on indoor air quality and health, Sci. Total Environ., 2021, 766, 142585.
42. F. J. Kelly and J. C. Fussell, Improving Indoor Air Quality, health and performance within environments where people live, travel, learn and work, Atmos. Environ., 2019, 200, 90–109.
43. N. N. Hansel, N. Putcha, H. Woo, R. Peng, G. B. Diette and A. Fawzy, et al., Randomized clinical trial of air cleaners to improve indoor air quality and chronic obstructive pulmonary disease health: results of the CLEAN AIR study, Am. J. Respir. Crit. Care Med., 2022, 205(4), 421–430.
44. H. J. Park, H. Y. Lee, C. H. Suh, H. C. Kim, H. C. Kim and Y. J. Park, et al., The effect of particulate matter reduction by indoor air filter use on respiratory symptoms and lung function: a systematic review and meta-analysis, Allergy, Asthma Immunol. Res., 2021, 13(5), 719.
45. V. K. Vijayan, H. Paramesh, S. S. Salvi and A. A. K. Dalal, Enhancing Indoor Air Quality –the air filter advantage, Lung India, 2015, 32(5), 473.
46. P. R. Dal, M. N. Kunz, T. Pistochini, R. L. Corsi and C. D. Cappa, Characterizing the performance of a do-it-yourself (DIY) box fan air filter, Aerosol Sci. Technol., 2022, 56(6), 564–572.
47. D. B. Day, J. Xiang, J. Mo, M. A. Clyde, C. J. Weschler and F. Li, et al., Combined use of an electrostatic precipitator and a high-efficiency particulate air filter in building ventilation systems: Effects on cardiorespiratory health indicators in healthy adults, Indoor Air, 2018, 28(3), 360–372.
48. Y. Liu, Electrostatic Precipitation of Fine Airborne Particles and Biological Decontamination: Efficiency and Energy Consumption Tendencies, University of Strathclyde, Glasgow, UK, 2018, Available from: doi:  DOI:10.48730/435a-rg52.
49. L. Y. Lin, C. Y. Lin, Y. C. Lin and K. J. Chuang, The effects of indoor particles on blood pressure and heart rate among young adults in Taipei, Taiwan, Indoor Air, 2009, 19(6), 482–488.
50. F. Reisen, J. C. Powell, M. Dennekamp, F. H. Johnston and A. J. Wheeler, Is remaining indoors an effective way of reducing exposure to fine particulate matter during biomass burning events?, J. Air Waste Manage. Assoc., 2019, 69(5), 611–622.
51. C. M. F. Rosales, Measurements of Atmospheric Radicals in Ambient and Indoor Environments: Outdoor and Indoor Relationships, Indoor Formation Mechanisms, and Instrumental Analysis, PhD dissertation, Indiana University, Bloomington, Indiana, USA, 2021, Available from: https://hdl.handle.net/2022/26746.
52. J. M. Mattila, C. Arata, C. Wang, E. F. Katz, A. Abeleira and Y. Zhou, et al., Dark Chemistry during Bleach Cleaning Enhances Oxidation of Organics and Secondary Organic Aerosol Production Indoors, Environ. Sci. Technol. Lett., 2020, 7(11), 795–801.
53. C. Wetmore, Enhancing indoor air quality in healthcare facilities, Healthcare Facilities Today, 2014, [cited 2023 Feb 10]. Available from: https://www.healthcarefacilitiestoday.com/posts/Enhancing-indoor-air-quality-in-healthcare-facilities--6708.
54. C. C. R. Ramos, J. L. A. Roque, D. B. Sarmiento, L. E. G. Suarez, J. T. P. Sunio and K. I. B. Tabungar, et al., Use of ultraviolet-C in environmental sterilization in hospitals: A systematic review on efficacy and safety, Int. J. Health Sci., 2020, 14(6), 52–65.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ea00160e

---