Influence of Human Behavior on Indoor Air Quality in a Care Facility for the Elderly in Japan

Volatile organic compounds (VOCs), CO2, temperature, and humidity in a private room in a care facility for the elderly were measured and the behavior of a resident and staff were recorded in order to clarify the effects of the resident’s behavior, especially defecation, on indoor air quality. Average indoor concentrations of total VOCs (in μg m−3) in summer, autumn, and winter were 40.9, 16.7, and 18.8, respectively. Average indoor concentrations of CO2 in summer, autumn, and winter were 813, 761, and 1144 ppm, respectively, revealing a tendency for the concentrations of CO2 to be higher in winter, in contrast to the VOC concentration. The concentrations of VOCs and CO2 were 1.1 to 1.5 times higher when the resident was present in the room than when the resident was absent. This result suggests that one of the main sources of VOC and CO2 emissions in indoor air was the resident. Acetic acid, 1-butanol, propanoic acid, hexanoic acid, and phenol, which are contained in human sweat, exhaled air, and excrement, were the predominant VOCs in the air of the room regardless of the season, and these five components accounted for more than 90% of the total VOCs. The concentrations of these components were higher when the resident was present in the room, suggesting that the resident was the main source of these components. Based on the changes in the VOC and CO2 concentrations over time and the records of the resident and the staff, it was noted that VOC concentrations decreased, in some cases, before and after diaper changes. Our research suggests that certain aspects of the behavior of residents can be inferred by monitoring changes in indoor air quality.


INTRODUCTION
The number of elderly people in Japan continues to increase.As of June 2021, there are 36.34 million elderly people aged 65 or over in Japan, accounting for 29% of the total population (SBJ, 2021).This number is expected to reach 39.21 million in 2040, accounting for 35.3% of the total population (IPSS, 2017).As the number of elderly people increases, so does the number who need assistance and care.Those aged 65 or over who were certified as needing care rose from 4.91 million in 2010 to 6.45 million in 2018 (Cabinet Office Japan, 2021).As a result, the number of elderly people moving into nursing homes and using home care services has been increasing year by year, and this trend is expected to continue (MHLW, 2021).
In Japan, nursing homes and other facilities for the elderly are required by law to have at least one caregiver for every three residents (MHLW, 1999).Caregivers provide a wide range of services to residents, and many caregivers complain of physical disability, especially back pain (Shogenji et al., 2006;Onuki et al., 2004).In addition to bathing (Kawahara et al., 2010), diaper changing is a common cause of back pain among caregivers (Shogenji et al., 2006;Onuki et al., 2004;Japan Institute of Personnel Administration, 2000).
More than 50% of residents of care facilities for the elderly in Japan use diapers (Ogata et al., 1997), and caregivers spend about 20% of their work time on toileting care, which includes changing diapers (Kooka et al., 2013).In many cases, toileting care is provided in response to requests from elderly residents, but in some cases, there has not been any excretion or defecation, in which case caregivers lose work time.This loss of time may translate into there being insufficient time for the caregiver to perform care that is actually needed by the residents.In other words, if caregivers can anticipate when elderly residents need care, they will be able to provide higher quality services to them at the appropriate time.One potential way to achieve this is to use sensors and other devices to better understand the behavior of elderly people and identify services that are appropriate for their situation.Many studies have already been conducted on the use of sensors to watch over elderly people.For example, Nakano et al. (2016Nakano et al. ( , 2015) ) developed a method to predict the behavior of elderly people by monitoring the electricity usage in their homes.Takama et al. (2017) analyzed logs acquired from various sensors installed in care facilities for the elderly and from sensors worn by care staff, and they reported that the work of care staff and the residents' sleep status could be estimated.In addition to the above, other examples of using force sensors (Hatsukari et al., 2012;Matsuoka et al., 2012) and camera images (Oda and Fujiwara, 2010) to estimate the behavior of the elderly have been reported.
There are thus many reports on understanding the behavior of elderly people based on the physical indicators of both the indoor facilities and the people living and working in them.On the other hand, only a few reports have sought to better understand the behavior of elderly people based on changes in the concentrations of chemical components in the indoor environment (e. g., Oyabu et al., 2001;Sawada et al., 2001;Suzuki et al., 2001).However, since chemical substances are emitted through human activities such as excretion, it is possible to estimate human activities by understanding the changes in their concentrations.
Volatile organic compounds (VOCs), which account for most of the components emitted as a result of human activities such as excretion, are diverse (Li et al., 2020) and their amounts and compositions can vary greatly depending on the type of activity and the environment (Vardoulakis et al., 2020;Nazaroff, 2013;Milner et al., 2011).In addition, sensors that can monitor VOCs in real time are commercially available, but such sensors typically display only the total VOC (TVOC) concentration and cannot measure individual VOCs.The purpose of this study was to clarify the relationships between human behavior and indoor air quality by measuring indoor air quality, especially VOCs, as well as to gain a better understanding of the behavior of elderly residents and staff in a care facility setting.This study examines relationships between changes in VOC concentrations as a result of defecation and the timing of diaper changes, which are a significant work burden for caregivers in care facilities for the elderly.

1 Monitoring Site
The indoor air was monitored in a private room of a three-story reinforced concrete facility for the elderly located in the Kanto region of Japan.An outline of the private room is shown in Fig. 1.The room has an area of about 18 m 2 and a volume of about 43.2 m 3 , and is equipped with an air conditioner as well as a sash window and a natural ventilation port.In both the private room and the building, heating and cooling are provided by electricity, and no combustion heating is used.The resident, a male in his 80s, is unable to walk on his own and always uses a wheelchair for mobility.He needs assistance to stand up and sit down.He continually wears diapers and does not use the toilet in his private room.He eats all three meals in the communal dining room and takes a bath once every three days.Except for having meals and bathing, he spends most of his time lying in bed.In collecting personal information, in-formed consent were given to the resident and the care facility in advance, and their consent was obtained.In addition, this study was conducted in accordance with the "Ethical Guidelines for Research Involving Human Subjects" set forth by the Central Research Institute of Electric Power Industry, to which the corresponding author belongs.

2 Sample and Data Collection
Sample and data collection was conducted on three consecutive days in summer ( July 29-31, 2020), autumn (November 9-11, 2020), and winter (February 8-10, 2021).A temperature, humidity and a CO 2 sensor (2JCIE-BU01, OMRON) were placed on a shelf at a height of about 1 m, and data were acquired every 10 s (Fig. 1).In addition, adsorbent tubes filled with Tenax TA ® (3.5 in.× 0.25 in.outer diameter, 60/80 mesh, COMSCO), which were previously conditioned at 300°C for 1 h, were mounted on a tube sampler (MTS-32, Markes International).The tube sampler was placed on a shelf about 1 m high, and the room air was aspirated at 0.1 L min -1 for 1 h for collection of each sample (Fig. 1).The sampler was used to collect samples continuously for about 48 h from around noon on the first day to around noon on the third day during each sampling campaign.The adsorbent tubes were tightly sealed and stored in a cool, dark place after sample collection.An infrared camera was installed on the wall (Fig. 1) at a height of about 1.8 m to capture images of the room every 10 s in order to monitor the behavior of the resident and staff, the opening and closing of the sash window, and the use of the air conditioner.

1 Overview of Indoor Air Quality
An overview of the measurement results for each sampling period is given in Table 1.The average concentration (μg m -3 ) of the sum of the 18 VOCs targeted in this study (TVOC) in the private room for each period was 40.9 in summer, 16.7 in autumn, and 18.8 in winter, thus indicating that the TVOC concentration was highest in summer.However, the average concentration for indoor CO 2 (ppm) was 813 in summer, 761 in autumn, and 1,144 in winter, showing a tendency for higher CO 2 concentrations in the winter, in con- trast to the TVOC concentration.The average indoor temperature was 25.3°C in summer, 23.4°C in autumn, and 22.8°C in winter, with little difference between the seasons owing to the use of air conditioning and heating in the room.The average relative humidity in the room was 68.3% in summer, 36.9% in autumn, and 26.3% in winter; thus, the humidity was highest in the summer and lowest in the winter.The relatively higher concentrations of TVOCs in summer compared with the other seasons may be due to the high emissions of VOCs caused by perspiration from the resident and staff.Although the room in the facility was air-conditioned and the average indoor temperature was maintained at 25.3°C even during summer, this was hotter than in the other seasons, which may have encouraged perspiration.However, the high relative humidity during summer may have contributed to the increase in the VOC concentrations.Markowicz and Larsson (2015) reported that an increase in relative humidity indoors promotes desorption of VOCs from surfaces.The high VOC concentrations during summer in this study may have been partly due to the high relative humidity.In addition, the relatively high concentrations of CO 2 in the winter were attributed to decreased ventilation efficiency, since neither the private room nor the facility used combustion heating.Further, the reason why the average concentration of CO 2 was relatively lower in the autumn may be because, unlike the other seasons, there were times when the windows of the private rooms were open during the observation period.This is consistent with the fact that the CO 2 concentration was low, at around 500-600 ppm, even though the resident was present during the time when the windows were open.
A summary of the effects of the residents' presence or absence on indoor air quality is given in Table 1.This table shows that the TVOC and CO 2 concentrations were higher when the resident was present in the room in all seasons, with values ranging from 1.1 to 1.5 times higher than those when the resident was absent.This result indicates that the presence of the resident in the room is one of the main sources of emission of VOCs and CO 2 in the indoor air.Mitova et al. (2020) evaluated the effect of the presence and absence of people on indoor concentrations of chemicals and found that the concentrations of VOCs such as isoprene and formaldehyde increased when people were in the room.Our results were similar to those of Mitova et al., although the components targeted in the present study were different.However, the temperature and relative humidity remained nearly constant regardless of the presence or absence of the resident.This is thought to be because the facility's air conditioning unit was always in operation, thus maintaining a relatively constant temperature and humidity.Standard deviation. 2)Ratio of the concentration when the resident was present (P) to that when the resident was absent (A). 3) Total volatile organic compounds. 4) Relative humidity.

2 Composition of Indoor VOCs
The average concentration and standard deviation for each VOC in the room are given in Table 2.For all seasons, among the 18 VOCs targeted, acetic acid had the highest concentrations, accounting for more than 50% of the TVOCs.Acetic acid and five other VOCs, namely, 1-butanol, propanoic acid, hexanoic acid, and phenol, were predominant regardless of the season, accounting for more than 90% of the TVOCs.Table 2 also shows the average concentration ratio for each component when the resident was present and absent.From the table, it is clear that the concentrations of most VOCs were higher when the occupant was in the room.Short chain fatty acids (SCFAs) such as acetic acid are produced as a result of fermentation of indigestible foods by gut microbiota (e.g., Lee and Zhu, 2021;Baxter et al., 2019;Raninen et al., 2016).The formation of SCFAs is influenced by various factors such as the pattern of food intake, antibiotic treatment, and microbial populations (Lee and Zhu, 2021).Phenols are produced by intestinal microorganisms from tyrosine, one of the main amino acids (Smith and Mac-Farlane, 1996), and their production is correlated with dietary protein intake (Geypens et al., 1997).These components are found in exhaled air as well as in sweat (Casas-Ferreira et al., 2019;Meijerink et al., 2000) and excreta (Casas-Ferreira et al., 2019).Therefore, one of the main sources of these VOCs in the indoor air is considered to be the resident.This is consistent with the fact that the VOC concentrations were higher when the resident was present in the room than when he was absent.
The correlations among the components of indoor VOC concentrations in each season are shown in Fig. 2, indicating that the correlations among VFAs and phenols tend to be relatively high regardless of the season.However, the correlations between components not of the same type were not very high, especially for alcohols and ketones.The high correlation between VFAs and phenols may be attributable to the fact that they are produced as a result of a series of fermentation reactions and microbial metabolism in the body.In contrast, alcohols and ketones have production mechanisms in the body Standard deviation. 2)Ratio of the concentration when the resident was present (P) to that when the resident was absent (A). 3) Detected when the resident was present but not when the resident was absent. 4)Not detected regardless of the presence or absence of the resident. 5)No data.
that are different from those of the other components, and thus their correlations with the other components were low.

3 Relationship between Human Activities and
Indoor Air Quality The temporal changes of indoor VOC and CO 2 concentrations and the behavior of the resident and staff in each season are shown in Fig. 3.In the figure, the timing of the entry by staff into the room is indicated by a small dot given that staff were typically in the room for only 5 min at most.The windows in the room were always closed except in autumn, and the air conditioner was not used except in summer.In addition, for diaper changes, although toileting care was noted in the facility's records, urinary excretion (and not defecation) was noted in all instances and the quantity was unknown.
As shown in Fig. 3, the VOC concentrations in the air of the room tended to increase or decrease within a relatively short period of time.Also, as shown in Table 2, the VOC concentrations tended to be higher when the resident was in the room for all seasons, and the VOC concentrations increased and decreased significantly even when the resident was in the room.This result suggests that the behavior of the resident may have some influence on the indoor VOC concentration.However, no relationship was observed between the timings of staff entry and the VOC concentrations.This may be because, with the exception of a few cases, the staff members were in the room for only about 5 min, a time which did not affect the VOC concentrations given the low temporal resolution of the measurement technology.In contrast, when staff entered the room the CO 2 concentration increased and then decreased after they left, in some cases, suggesting that the staff' behavior could be captured owing to the high temporal resolution of the CO 2 sensor.
The resident in question is paralyzed, wears diapers at all times, and spends most of his time indoors lying in bed.His autonomous physical activity is limited to small movements of the arms and legs.Therefore, the emission sources of VOCs originating from the resident are considered to be limited to exhalation, body odor, and excretion.Of these, VOCs due to exhalation and body odor are considered to be emitted at approximately the same rate when the resident is in the room.However, VOCs derived from defecation are considered to be emitted during the period between defecation and a dia-  , 2-pentanone; 2, 2-butanol; 3, 1-propanol; 4, dimethyl disulfide; 5, 1-butanol; 6, acetic acid; 7, propanoic acid; 8, isobutyric acid; 9, butanoic acid; 10, isovaleric acid; 11, pentanoic acid; 12, hexanoic acid; 13, heptanoic acid; 14, phenol; 15, p-cresol; 16, p-ethylphenol; 17, indole; 18, skatole; 19, total.Summer Autumn Winter per change.Fig. 3 shows that, for most cases, VOC concentrations increased before a diaper change and decreased after a diaper change in summer.However, in aut-umn and winter, VOC concentrations decreased before and after diaper changes in fewer cases than in summer and were more pronounced in winter.These results sug-   4. The decrease in the VOC concentration was more pronounced for VFAs, as represented by acetic acid.Given that VFAs such as acetic acid are frequently detected in feces, it is reasonable that the concentrations of the VFAs decreased after a diaper change.
There were, however, some cases where the indoor VOC concentrations did not change or increased after a diaper change.The reason for this is not clear, but it is possible that if the amounts of excretion and/or defecation were very small, very few VOCs would have been emitted.The resident had his diaper changed every few hours, and it is thought that there were some cases in which the amount of excretion and defecation was extremely small or there was no excretion and defecation.In addition, the resident used relatively thick bedding in the winter, which may have suppressed the volatilization of VOCs from the bedding.In fact, there were few cases where the VOC concentrations increased before diaper change in the winter.In addition, the measurement frequency of VOCs in this study was one sample per hour, and the temporal resolution was not high.Therefore, it is possible that a temporary increase in concentration was not captured.Furthermore, the ventilation rate is not expected to be constant at all times, which would also affect the indoor VOC concentrations.
The hourly changes in the composition of the VOCs for each season are shown in Fig. 5.Although the concentrations of acetic acid and 1-butanol, the major constituents, remained relatively constant throughout all seasons, the composition of the constituent VOCs did change within short periods of time.For example, in summer, the proportion of 1-propanol increased after 19:40 on July 30.Although the reason for this increase is not clear, it may be because the resident had eaten fruit that contained a lot of 1-propanol at dinner that evening shortly before the measurement.In addition, the ratio of VFAs other than acetic acid and phenol temporarily increased in the sample at 12:40 on November 10 (autumn).At this time, two research workers were working in the room for about an hour for sampling.The temporary change in composition may have been influenced by the VOCs originating from the researchers.Overall, the results suggest that detailed information on human behavior, diet, and other aspects can be extracted by analyzing changes in the indoor concentrations and composition of VOCs.

CONCLUSION
To clarify the relationships between a resident's behavior and indoor air quality in a care facility for the elderly, VOC and CO 2 concentrations, temperature, and humidity were measured and the resident's behavior was ascertained from camera images installed in the room.The TVOC concentrations in the indoor air had a tendency to be higher in summer, whereas the CO 2 concentrations were higher in the winter.Given that the concentrations of both VOCs and CO 2 were higher when the resident was present, it was considered that the resident was one of the main sources for these components.Among the 18 VOCs targeted, acetic acid, 1-butanol, propanoic acid, hexanoic acid, and phenol, which are contained in human sweat and excrement, were the most prevalent VOCs, regardless of season, and their concentrations were relatively high when the occupant was present in the room.These findings suggest that these VOCs may have been emitted from the sweat and feces of the occupant.Finally, as a result of evaluating the relationships between the behavior of the resident and staff and the VOC concentrations, the VOC concentrations were found to increase before diaper changes and to decrease after the changes in many cases.This indicates that the timing of defecation can be estimated by monitoring the resulting VOC concentrations.

Fig. 1 .
Fig. 1. Outline of the private room in the care facility.The room has an area of about 18 m 2 , a ceiling height of 2.4 m, and a volume of 43.2 m 3 .

Fig. 3 .
Fig. 3. Temporal changes of indoor VOC and CO 2 concentrations and the behavior of the resident and staff for each season.Some data on the CO 2 concentration in autumn could not be obtained owing to a failure of the sensor.A red dotted line indicates that the VOC concentrations decreased before and after a diaper change.

Fig. 4 .
Fig. 4. Ratio of VOC concentrations after a diaper change to that before diaper change in the case where VOC concentrations decrease after diaper change.x, not detected both before and after a diaper change; y, detected before diaper change but not after; z, detected after a diaper change but not before.

Fig. 5 .
Fig. 5. Hourly changes in the composition of VOCs for each season.

Table 1 .
Overview of the indoor air quality during each sampling campaign.

Table 2 .
Concentrations of indoor VOCs for each sampling campaign.