Incident location at home and emergency medical services on-scene time for older adults in Japan: a population-based cross-sectional study

Abstract

Background

As population aging increases the demand for emergency medical services (EMS), reducing prehospital time has become crucial for improving patient outcomes and operational efficiency. Accessible housing environments may facilitate patient transportation and reduce on-scene time (interval between ambulance arrival and departure from the scene), but evidence is scarce. This study examined the associations between incident locations at home and EMS on-scene time among older adults in Japan.

Methods

This population-based cross-sectional study analyzed 26,058 people aged ≥ 65 years transported from their homes to hospitals by the Higashi-Hiroshima Fire Department’s ambulances between January 2016 and December 2024. We examined the associations between two exposure variables (floor level and in-home location) and on-scene time using generalized linear models with gamma distribution and log-link function. Adjusted mean ratios (aMRs) and 95% confidence intervals (CIs) were calculated. Then, a correspondence analysis using text data from the ambulance dispatch records was conducted to generate further hypotheses regarding the factors characterizing each in-home location.

Results

The median EMS on-scene time was 16 min (interquartile range: 13–21). Higher floors were significantly associated with progressively longer on-scene times. When compared with living rooms, corridors (aMR: 0.95, 95% CI: 0.93–0.97) and gardens (aMR: 0.95, 95% CI: 0.93–0.98) had significantly shorter on-scene times. In contrast, toilets (aMR: 1.08, 95% CI: 1.04–1.12), bathrooms (aMR: 1.09, 95% CI: 1.05–1.13), and kitchens (aMR: 1.06, 95% CI: 1.03–1.10) had significantly longer on-scene times. Correspondence analysis showed that garden and other locations were positioned at the “exogenous” pole along the horizontal axis (representing type of injury or disease). The toilet, bathroom, and kitchen, which were associated with longer on-scene times, were positioned toward the “loss of consciousness” pole along the vertical axis (representing level of consciousness).

Conclusions

Higher floors and confined areas (toilets and bathrooms) were associated with longer on-scene time. Future studies are invited to examine the effect of patients’ consciousness levels to ensure that paramedics can rapidly reach and transport emergency patients regardless of location.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12873-026-01694-6.

Background

Since critical illnesses and injuries are time-sensitive, prehospital time, referring to the duration between an emergency call and patient arrival at a hospital, is an important concern in emergency medical services (EMS). Still, optimal approaches are not dictated exclusively by time, as paramedics must decide on the approach to use according to the patient’s clinical demands, environmental context, and resource availability [1]. It remains that prolonged prehospital time is critical owing to its association with poor prognosis in patients who require critical care, such as those with acute myocardial infarction or severe trauma [2–4].

Population aging is increasing worldwide [5], and so is EMS demand, as older adults are more likely to require emergency care [6]. Indeed, the number of ambulance dispatches is increasing in Japan’s rapidly aging population, with older adults accounting for approximately 60% of all such dispatches [7]. The Japanese people can ask for ambulance transportation by calling “119,” which connects them to dispatch centers of local governments, which in turn are the ones managing the EMS system. The local governments also cover the costs associated with EMS personnel and transportation, such that ambulance users are not charged any costs [8]. Additionally, the basic principle of the Fire Service Act in Japan stipulates a primary focus on lifesaving, with patients being promptly transported to a hospital after the administration of necessary first aids; only those who refuse to go to the hospital or die are not transported [9]. Although the rate of non-conveyance varies greatly worldwide (3.7–93.7%) [10], it is only 12.7% in Japan [11]. Considering this background, the need emerges for reducing prehospital time in order to meet the escalating demand for EMS among community-dwelling older adults. The expectation is that reductions in prehospital time enable ambulances to respond to a greater volume of emergency calls and the optimization of EMS operational capacity.

Prehospital time can be divided into three phases: response time, on-scene time, and transport time. Response time refers to the interval between the initial emergency call and ambulance arrival at the scene, on-scene time refers to the interval between ambulance arrival and departure from the scene, and transport time refers to the interval between ambulance departure from the scene and arrival at the destination hospital [2, 12]. While response and transport times can potentially be reduced through major organizational and logistical modifications (e.g., increasing the number of EMS units available or optimizing service area coverage), on-scene time has been identified as a more actionable target for improvements [2, 13]. The literature has generally analyzed on-scene time using the following variables: patient demographics, including age and sex [13–16]; case characteristics, including functional capacity [17] and patient acuity [13, 15, 16, 18]; medical system, such as hospital inquiries and type of destination hospital [12, 13, 15, 16, 18]; temporal factors, like emergency calls during after-hours or public holidays [15], winter months [16], and systemic disruptions caused by the COVID-19 pandemic [13]; geographical factors, inclusive of urban–rural disparities [15, 16]. However, the above factors are generally difficult to change, and there is a need to identify more modifiable determinants of on-scene time.

The role of housing structure as an environmental factor for on-scene time remains largely unexplored. While Japan’s Building Standards Act [19] and associated regulations stipulate building configurations, these can be modified in the long term if there is compelling evidence. For example, Japanese housing is frequently constructed in limited spaces, especially in high-density urban areas, leading to narrow rooms and corridors that restrict EMS operations [20]. Therefore, the location (e.g., living room, stairs, and bathroom) in which an emergency incident occurs could influence on-scene time. Furthermore, there is a global shift toward multi-story apartment living [21–23], and this trend is also evident in Japan’s older population. Over the past 10 years, the percentage of Japanese households with apartment-dwelling residents aged ≥ 65 years has increased from 25.3% to 29.9% [24], reflecting a growing preference for residences that provide easy access to essential services and enhanced safety from crime [25]. However, the reduced accessibility associated with incidents occurring on higher apartment floors can extend on-scene time [26]. Accordingly, this study aimed to examine the association between incident locations at home and on-scene EMS time among community-dwelling older adults in Japan.

Methods

Study setting and data sources

This population-based cross-sectional study was conducted using administrative data from January 2016 to December 2024 provided by the Fire Department of Higashi-Hiroshima City (Hiroshima Prefecture, Japan). In 2020, it featured 196,608 residents, among which 23.9% were aged ≥ 65 years, which is slightly lower than the national average of 28.6% [27]. The Higashi-Hiroshima Fire Department also provides EMS for neighboring municipalities such as Takehara City and Osaki-Kamijima Town.

The study database contained two types of data: ambulance dispatch records and ambulance transport records. Ambulance dispatch records are obtained from a comprehensive ledger managed by the emergency dispatch center, and contain information on all emergency response calls (including cases that did not result in hospital transport). Ambulance transport records include information documented by paramedics and other EMS personnel during or immediately following the emergency response. These data were provided under an agreement between the Higashi-Hiroshima Fire Department and Hiroshima University. Anonymized identification numbers were randomly created based on unique identifiers that can link specific individuals across datasets. These identification numbers were used to merge the two data types.

Study cohort

The study cohort comprised older adults aged ≥ 65 years who were transported by ambulance from their homes to a hospital. The database originally contained 89,536 cases, and we excluded patients aged < 65 years (n = 41,207), patients who experienced incidents outside of their homes (e.g., street or workplace) (n = 20,362), patients who were not transported to a hospital (e.g., refusal) (n = 1,215), patients who were transferred between hospitals (n = 61), and patients who had died before hospital arrival (n = 580). From the 26,111 older patients transported from their homes to hospitals, we further excluded 53 individuals due to missing data. The study cohort comprised 26,058 patients (Fig. 1).

Fig. 1.

Flow diagram of study cohort selection

Outcome

The study outcome was EMS on-scene time, which was defined as the time in minutes between ambulance arrival and departure from the scene [2, 12]. This variable was manually recorded by the paramedics in the ambulance transport records.

Exposures

The study exposures were incident locations at home, which were examined using two variables: floor level and in-home location. Floor level indicated the residential floor on which an incident occurred, and was categorized into first, second, third, fourth, fifth, and sixth floor or higher. In-home location was determined using the most granular classification of incident location within a patient’s residence as recorded in the ambulance transport records. In these records, the locations are divided into nine categories: living room, corridor, stairs, toilet, bathroom, kitchen, garden, roof, and other location. We combined incidents that occurred on roofs and other locations into a single category due to the small number of cases. Therefore, our analysis used eight categories for in-home location.

Covariates

The following variables were included as covariates: patient factors (age, sex, nationality, and incident type), temporal factors (date and time of emergency call, COVID-19 pandemic period, and ambient temperature), and geographical factors (region of residence).

Patient factors included age (65–74 years, 75–84 years, or ≥ 85 years), sex (male or female), and nationality (Japanese or others). Although the database classified incident types into 12 groups [9], we consolidated these groups into the following four categories: acute illness, general accidents (unexpected accidents not classified as specific types of accidents, such as fire-related accidents, natural disasters, water-related accidents, motor vehicle accidents, occupational accidents, and sports-related accidents), self-injuries, and other incidents (assaults or the specific types of accidents described above). For the temporal factors, the dates and times of the emergency calls were identified and categorized into six groups: weekday daytime (9:00–16:59), weekday early night (17:00–00:59), weekday late night (1:00–8:59), weekend daytime (9:00–16:59), weekend early night (17:00–00:59), and weekend late night (1:00–8:59). The COVID-19 pandemic period was analyzed as three phases: pre-pandemic (January 1, 2016 to March 31, 2020), pandemic (April 1, 2020 to May 7, 2023), and post-pandemic (May 8, 2023 to December 31, 2024). These phases were set based on the Japanese government’s declaration of the first state of emergency on April 1, 2020 and the eventual downgrading of COVID-19’s infectious disease classification from Category II (high risk due to infectivity and severity of illness) to Category V (subject to national surveillance, with information disclosure to prevent occurrence and spread) on May 8, 2023 [28]. The ambient temperature for each incident was ascertained from historical weather data maintained by the Japan Meteorological Agency [29]. The data included the average temperature and wet-bulb globe temperature (WBGT) for each day in the Higashi-Hiroshima area, and these were categorized into three groups: cold (< 18 °C), hot (WBGT > 28 °C), and neither. The WBGT cut-off of 28 °C was used to indicate an increased risk of heat stroke [30]. The region of residence was analyzed using 16 sub-municipal districts (designated A to O and others) specified in the basic planning frameworks of Higashi-Hiroshima City, Takehara City, and Osaki-Kamijima Town [31–33].

Statistical analysis

First, we calculated Pearson’s correlations between on-scene time and the total prehospital time, as well as median and interquartile ranges (IQR) for EMS on-scene time according to the exposure and covariate variables. Next, we estimated the associations between incident locations at home and on-scene time using generalized linear regression models with a gamma distribution and log-link function. The coefficients and standard errors were exponentiated for interpretation as mean ratios (MRs) and 95% confidence intervals (CIs), respectively. A multivariable model that included the covariates was used to estimate the adjusted MRs (aMRs) and 95% CIs.

To examine whether the target associations were robust across severity levels, we conducted a subgroup analysis using the multivariable model after stratifying the patients according to the severity of their conditions. Japan’s Fire and Disaster Management Agency classifies patient conditions into mild, moderate, severe, and death based on the national criteria for emergency statistics [34]. These conditions were ascertained using physician-recorded data collected at the time of each patient’s initial emergency department visit. Mild conditions were defined as “injuries or diseases that do not require hospitalization or treatment”, moderate conditions were defined as “injuries or diseases that require hospitalization and treatment for less than three weeks”, and severe conditions were defined as “injuries or diseases that require hospitalization and treatment for three weeks or more”. Individuals who had died before hospital arrival were not included in the study.

Subsequently, we conducted a correspondence analysis using text data from ambulance dispatch records to generate further hypotheses on the variables characterizing each in-home location. The ambulance dispatch records included information available prior to ambulance arrival at the scene, and were based on reports provided by callers and assessments made by dispatch personnel. The analysis was performed in Japanese, and the results were translated into English and visualized in a biplot.

The descriptive analysis and regression analyses were conducted using Stata/MP version 18.0 (StataCorp LLC, College Station, Texas, USA). The correspondence analysis was conducted using KH Coder version 3 (SCREEN Advanced System Solutions Co. Ltd., Kyoto, Japan), which is a widely used Japanese software originally developed for the fields of sociology and social research [35]. The authors used Claude Sonnet 4.5 (Anthropic, California, USA) to improve the language and clarity of this manuscript. After using this service, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Ethical considerations

All data were anonymized before being received by the researchers. This study was conducted in accordance with the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of the Hiroshima University (Approval number. E2020-2381). The committee waived the need for informed consent due to the use of anonymized data collected for routine operations and the retrospective nature of the study.

Results

The relationship between on-scene time and total prehospital time is illustrated in Fig. 2. Longer on-scene time was associated with longer total prehospital time (r = 0.617). Moreover, patient characteristics and EMS on-scene times are shown in Table 1. Most emergency incidents occurred on the first floor; on-scene time generally increased with higher floors and was approximately seven minutes longer for incidents occurring on the sixth floor or higher than those occurring on the first floor. Among in-home locations, incidents occurred most frequently in living rooms, followed by corridors and gardens. Incidents occurring in gardens and other locations had the shortest on-scene time, while those occurring in toilets and bathrooms had the longest. The distribution of patient conditions according to incident location is shown in Supplementary Fig. 1.

Fig. 2.

Relationship between the duration of on-scene time and the total prehospital time

Table 1.

Patient characteristics and EMS on-scene time

		n	%	Median [IQR] on-scene time, mins
Total		26,058	100.0	16 [13–21]
Incident locations
Floor level
	1st floor	24,161	92.7	16 [12–20]
	2nd floor	1,305	5.0	18 [14–22]
	3rd floor	249	1.0	20 [16–25]
	4th floor	122	0.5	23 [17–26]
	5th floor	67	0.3	21 [16–26]
	≥ 6th floor	154	0.6	23 [18–29]
In-home location
	Living room	20,072	77.0	16 [13–21]
	Corridor	2,349	9.0	15 [12–20]
	Stairs	240	0.9	17 [13–22]
	Toilet	580	2.2	18 [14–22]
	Bathroom	564	2.2	18 [14–23]
	Kitchen	653	2.5	16 [13–21]
	Garden	1,227	4.7	15 [12–20]
	Other location	373	1.4	15 [12–20]
Patient factors
Sex
	Male	12,898	49.5	16 [12–21]
	Female	13,160	50.5	16 [13–21]
Age, years
	65–74	6,561	25.2	16 [12–20]
	75–84	10,042	38.5	16 [13–21]
	≥ 85	9,455	36.3	16 [13–21]
Nationality
	Japanese	26,026	99.9	16 [13–21]
	Others	32	0.1	18 [15–21]
Incident type
	Acute illness	20,678	79.4	16 [12–21]
	General accidents	5,232	20.1	17 [13–21]
	Self-injuries	58	0.2	19 [15–25]
	Other incidents	90	0.3	20 [15–32]
Patient condition
	Mild	8,449	32.4	16 [13–21]
	Moderate	15,013	57.6	16 [13–21]
	Severe	2,596	10.0	16 [12–20]
Temporal factors
Date/time of emergency call
	Weekday daytime	7,690	29.5	16 [12–21]
	Weekday early night	5,969	22.9	16 [13–21]
	Weekday late night	5,085	19.5	17 [13–21]
	Weekend daytime	2,975	11.4	16 [12–20]
	Weekend early night	2,500	9.6	16 [12–21]
	Weekend late night	1,839	7.1	16 [13–21]
COVID-19 pandemic period a
	Pre-pandemic	11,224	43.1	15 [12–19]
	Pandemic	9,049	34.7	17 [14–22]
	Post-pandemic	5,785	22.2	16 [13–21]
Ambient temperature b
	Cold	15,972	61.3	16 [13–21]
	Neither	6,160	23.6	16 [12–20]
	Hot	3,926	15.1	16 [12–20]
Geographical factors
Region of residence
	A	493	1.9	14 [12–19]
	B	2,964	11.4	15 [12–20]
	C	824	3.2	14 [11–18]
	D	1,279	4.9	15 [12–20]
	E	5,264	20.2	18 [14–23]
	F	2,873	11.0	16 [13–21]
	G	1,173	4.5	16 [12–20]
	H	3,288	12.6	16 [13–21]
	I	2,184	8.4	17 [13–22]
	J	2,540	9.7	15 [12–20]
	K	650	2.5	16 [13–21]
	L	312	1.2	16 [13–21]
	M	957	3.7	16 [12–20]
	N	523	2.0	17 [13–21]
	O	731	2.8	14 [11–18]
	Others	< 10	0.0	6 [6–17]

EMS: emergency medical services; IQR, interquartile range

^a “Pre-pandemic” was from January 1, 2016 to March 31, 2020; “Pandemic” was from April 1, 2020 to May 7, 2023; and “Post-pandemic” was from May 8, 2023, to December 31, 2024

^b “Cold” was defined as < 18 °C (64 °F), “Hot” was defined as wet-bulb globe temperature > 28 °C, and “Neither” referred to any temperature other than the previous two categories

The generalized linear regression analysis results are presented in Table 2. Even after adjusting for covariates, higher floors (reference: first floor) were significantly associated with longer on-scene times. However, the fifth floor exhibited a lower aMR than the fourth floor. Among in-home locations (reference: living room), corridors and gardens were associated with shorter on-scene times, while toilets, bathrooms, and kitchens were associated with longer on-scene times.

Table 2.

Associations between incident locations at home and EMS on-scene time

		Total (n = 26,058)
		Crude				Adjusted
		MR	95% CI			aMR	95% CI
Incident locations						1.00
Floor level
	1st floor (reference)	1.00				1.00
	2nd floor	1.12	1.09	1.15		1.10	1.07	1.13
	3rd floor	1.21	1.15	1.29		1.18	1.12	1.25
	4th floor	1.30	1.19	1.41		1.25	1.16	1.36
	5th floor	1.27	1.14	1.43		1.20	1.08	1.34
	≥ 6th floor	1.37	1.27	1.48		1.31	1.22	1.41
In-home location
	Living room (reference)	1.00				1.00
	Corridor	0.95	0.93	0.97		0.95	0.93	0.97
	Stairs	1.09	1.02	1.15		1.05	0.99	1.11
	Toilet	1.08	1.03	1.12		1.08	1.04	1.12
	Bathroom	1.09	1.05	1.14		1.09	1.05	1.13
	Kitchen	1.05	1.01	1.09		1.06	1.03	1.10
	Garden	0.95	0.92	0.97		0.95	0.93	0.98
	Other location	0.99	0.95	1.04		0.98	0.94	1.03
Patient factors
Sex
	Male (reference)	1.00				1.00
	Female	1.00	0.99	1.01		1.00	0.99	1.01
Age, years
	65–74 (reference)	1.00				1.00
	75–84	1.02	1.00	1.03		1.02	1.00	1.03
	≥ 85	1.03	1.01	1.04		1.03	1.01	1.04
Nationality
	Japanese (reference)	1.00				1.00
	Others	1.08	0.92	1.27		1.01	0.87	1.18
Incident type
	Acute illness (reference)	1.00				1.00
	General accidents	1.04	1.02	1.05		1.04	1.02	1.05
	Self-injuries	1.49	1.32	1.68		1.50	1.34	1.69
	Other incidents	1.36	1.24	1.50		1.45	1.32	1.59
Temporal factors
Date/time of emergency call
	Weekday daytime (reference)	1.00				1.00
	Weekday early night	1.01	1.00	1.03		1.01	1.00	1.03
	Weekday late night	1.03	1.01	1.05		1.03	1.02	1.05
	Weekend daytime	0.98	0.96	1.00		0.98	0.96	1.00
	Weekend early night	1.01	0.98	1.03		1.01	0.99	1.03
	Weekend late night	1.04	1.01	1.06		1.04	1.02	1.07
COVID-19 pandemic period a
	Pre-pandemic (reference)	1.00				1.00
	Pandemic	1.19	1.17	1.20		1.18	1.17	1.20
	Post-pandemic	1.13	1.11	1.15		1.12	1.11	1.14
Ambient temperature b
	Cold	1.03	1.02	1.04		1.03	1.02	1.05
	Neither (reference)	1.00				1.00
	Hot	1.00	0.98	1.01		1.00	0.98	1.02
Geographical factors
Region of residence
	A (reference)	1.00				1.00
	B	1.02	0.97	1.06		1.00	0.96	1.04
	C	0.91	0.86	0.95		0.91	0.87	0.96
	D	1.02	0.97	1.07		1.00	0.96	1.05
	E	1.16	1.11	1.21		1.12	1.08	1.17
	F	1.07	1.02	1.11		1.05	1.00	1.09
	G	1.04	0.99	1.10		1.04	0.99	1.09
	H	1.09	1.04	1.14		1.08	1.03	1.12
	I	1.16	1.11	1.21		1.14	1.09	1.19
	J	1.04	1.00	1.09		1.02	0.98	1.07
	K	1.11	1.05	1.17		1.11	1.05	1.17
	L	1.10	1.03	1.17		1.10	1.03	1.17
	M	1.04	0.99	1.09		1.03	0.98	1.08
	N	1.12	1.06	1.19		1.12	1.06	1.18
	O	0.91	0.87	0.96		0.92	0.87	0.97
	Others	0.59	0.35	1.00		0.38	0.22	0.63

EMS, emergency medical services; MR, mean ratio; CI, confidence interval; aMR, adjusted mean ratio. The results were calculated using generalized linear regression models with a gamma distribution and log-link function. Significant results are shown in bold (p < 0.05)

^a “Pre-pandemic” was from January 1, 2016 to March 31, 2020; “Pandemic” was from April 1, 2020 to May 7, 2023; and “Post-pandemic” was from May 8, 2023, to December 31, 2024

^b “Cold” was defined as < 18 °C (64 °F), “Hot” was defined as wet-bulb globe temperature > 28 °C, and “Neither” referred to any temperature other than the previous two categories

In the severity-stratified subgroup analysis, higher floor levels were generally associated with higher aMRs across the three severity levels (Fig. 3). However, aMRs for the fourth floor were higher than those for the fifth floor among patients with moderate and severe conditions, but not with mild conditions. Meanwhile, incidents occurring in toilets and bathrooms had significantly higher aMRs for patients with mild and moderate conditions, but not with severe conditions. Only kitchen had a significantly higher aMR for on-scene time among patients with severe conditions.

Fig. 3.

Associations between incident locations at home and EMS on-scene time according to patient condition. aMR, adjusted mean ratio; CI, confidence interval; NA, not applicable. Asterisks indicate statistical significance: *p < 0.05

Figure 4 presents the biplot of the correspondence analysis of text data from the ambulance dispatch records. The bathroom, toilet, and kitchen were clustered in the upper-central region, whereas the corridor, garden, and other locations were localized in the left-central region. The horizontal axis was considered to represent types of injuries or diseases, transitioning from “exogenous conditions” on the left to “endogenous conditions” on the right. The vertical axis was considered to represent the levels of consciousness, with the uppermost pole indicating “loss of consciousness” and the lowermost pole indicating “fully conscious”. The living room was positioned at the furthest point along the rightmost axis (endogenous), and was clustered with words such as “complaint”, “dizziness”, and “abdominal pain”. The garden and other locations were positioned at the opposite pole (exogenous), and were clustered with words such as “injury” or anatomical sites such as “right hand” and “back of the head”. The toilet, bathroom, and kitchen (which were associated with longer on-scene times) were positioned toward the “loss of consciousness” pole. These were clustered with words such as “cardiopulmonary arrest” and “choke”. The frequency of words according to in-home location is provided in Supplementary Table 1.

Fig. 4.

Correspondence analysis of ambulance dispatch records according to in-home location. Dimension 1: types of injuries or diseases; Dimension 2: levels of consciousness. The area around the origin has been magnified by a factor of three for better visualization of the word associations

Discussion

This cross-sectional study of EMS data examined the associations between incident locations at home and on-scene time among older adults residing in three Japanese municipalities. Longer on-scene time was significantly associated with incidents occurring on higher floors as well as in toilets, bathrooms, and kitchens. Housing structure may represent an important modifiable factor to reduce on-scene time, and our study provides novel insight that may inform future strategies to support EMS operations.

A notable finding of our study was that higher floors were associated with longer on-scene time. This was consistent with expectations as the increased vertical distance and logistical challenges of higher floors would result in longer durations needed to reach, stabilize, and transport a patient to the ambulance. This finding aligns with a previous interview-based study in which paramedics reported longer on-scene times for high-rise apartments [26]. However, the regression analysis revealed a slightly higher aMR for the fourth floor (vs. fifth floor). A potential explanation for this finding is the presence/absence of elevators. In Japan, approximately 10% of buildings with four floors have elevators, whereas this figure rises to around 40% for buildings with five floors and 100% for buildings with six or more floors [19]. Logically, for incidents occurring on the fourth floor in buildings without elevators, paramedics must transport patients using stairs. Future studies are invited to probe into this matter using actual data on the presence of elevator access to clarify the reasoning behind slightly higher aMRs for the fourth floor. As multi-story apartments are an increasingly common form of housing [21–23], the operational challenges involved in transporting patients from higher-level residences place a heavier burden on EMS.

Next, we found that on-scene times differed among the in-home locations, with corridors and gardens associated with shorter on-scene times than living rooms. These associations remained consistent for patients with mild and moderate conditions, suggesting that the shorter on-scene times were not attributable to less severe conditions for incidents in corridors or gardens. A potential explanation is that there may be more ambulatory patients at these locations who do not require a stretcher and can cooperate with paramedics. In the correspondence analysis, locations with shorter on-scene times tended to be positioned in the “exogenous” axis, away from the “loss of consciousness” pole; more studies are hence warranted to investigate the effects of patient consciousness on on-scene time. Furthermore, gardens tend to offer better accessibility than indoor locations, potentially streamlining the transition from the incident site to the ambulance and perhaps explaining the association with shorter on-scene time.

Although toilets and bathrooms were associated with longer on-scene times overall, these associations were attenuated for patients with severe conditions. This attenuation may be explained by a “load and go” approach (also referred to as “scoop and run”) [1]: when EMS personnel identify a patient as critically ill, they may prioritize rapid transport irrespective of in-home location, performing necessary interventions during transport rather than at the scene. Under a “load and go” approach, necessary interventions such as intravenous access and drug administration would be performed during transport to the hospital rather than at the scene, thereby reducing the influence of location-specific physical barriers on on-scene time.

On the other hand, incidents occurring in toilets or bathrooms were associated with longer on-scene times for patients with mild and moderate conditions. Although overall severity is an important indicator of patient condition, it does not necessarily reflect consciousness levels. For example, fully conscious patients with femoral neck fractures would still be diagnosed with a severe condition, for obvious reasons. In our correspondence analysis, toilets and bathrooms were located at the “loss of consciousness” pole, allowing for the hypothesis that lower levels of consciousness, regardless of severity, hinder patient cooperation and reduce the paramedics’ ability to gather information. Importantly, we could not examine the associations of consciousness levels in this secondary analysis owing to the lack of data, and thus we suggest, once more, that researchers explore the impact of consciousness on on-scene time in the near future.

In addition, bathrooms and toilets are designed to ensure user privacy, resulting in structural configurations that are inherently difficult for paramedics to access from the outside. Drowning in bathtubs is the leading cause of indoor accidents among older adults in Japan [36], and approximately half of all fatal drowning accidents are attributable to acute medical events experienced by older adults during baths [37]. Such incidents may require paramedics to physically remove patients from the bathtub and dry their bodies before initiating life-saving interventions, thereby prolonging on-scene time. Next, the structural configuration of toilets in Japan may also affect on-scene times. Japanese homes typically separate toilets, dressing rooms, and bathrooms [20]. While these characteristics are preferred by Japanese people from the perspectives of hygiene and privacy, the subdivision of these rooms can impede paramedic access. Moreover, some residential units are equipped with inward-opening doors, and a patient falling unconscious directly behind such a door could prevent paramedics from gaining entry.

The associations between kitchen-based incidents and prolonged on-scene times were constant at all severity levels. In Japan, kitchens frequently include dining areas. As a result, incidents occurring in the kitchen are heterogenous and encompass a broad spectrum of cases ranging from burns during cooking to choking during meals. Further research should employ a more detailed classification of emergency cases originating in the kitchen.

This study has several limitations. First, the cross-sectional design cannot establish causality. However, as emergency incidents requiring EMS would precede on-scene time, the potential for reverse causation is negligible. Second, although the results for in-home locations reached statistical significance, the aMRs were modest (longer on-scene time: 1.05 to 1.09; shorter on-scene time: 0.95), such that the magnitude for individual prehospital time would be limited. However, it remains that the annual number of ambulances is increasing [7], making this limited magnitude still an important difference from a public health perspective, even if related interpretations should be made with caution. Third, this study could not account for several potential confounders, such as patient characteristics (e.g., functional status and cognitive impairment) and detailed building environments (e.g., the presence of elevators and building layout characteristics). Still, patients with functional decline, such as those with frailty, may prefer living on the lower floor rather than higher floors. Regarding cognitive function, patients with cognitive decline tend to experience suffocation [38] and can call paramedics after experiencing an incident in the kitchen. Their cognitive decline can also hinder effective communication, require family communication, and create difficulties associated with the agreement with the transfer process, thereby prolonging on-scene time [17]. Furthermore, older adults may tend to select housing with elevators as they become frailer; however, we could not obtain data on the presence of elevator access. In particular, elevator availability is associated with the daily accessibility of older residents on middle floors [39] while building layout may affect daily accessibility and emergency accessibility. More research considering these factors is necessary. Fourth, incident locations may have been misclassified in cases where patients were transferred by family members prior to EMS arrival, although such cases are unlikely to be common. Finally, this study used data from three municipalities in Hiroshima Prefecture, which limits the generalizability of our findings. Because these municipalities are located close to Hiroshima City, which is the capital of Hiroshima prefecture, our results may specifically reflect conditions in regional cities rather than major metropolitan areas such as Tokyo.

Despite these limitations, our study indicates that EMS strategies should also consider environmental aspects to minimize on-scene time in order to improve clinical outcomes and operational efficiency. Further research is needed to elucidate the mechanisms by which incident locations can influence on-scene time, thus clarifying potential targets for interventions. Moreover, although housing accessibility is important for enabling older adults to maintain their independence, current Japanese legislation does not mandate accessibility standards for private residences [40]. The Japanese government enacted the “Act on the Promotion of Smooth Transportation of Older People and Persons with Disabilities” (commonly known as the “Barrier-Free Act”) in 2006, but its statutes are restricted to buildings used by the general public or facilities designated for older adults or persons with disabilities [41]. From the perspective of building accessibility for paramedics, the “Guidelines for Ensuring Fire Department Access” specify requirements for large-scale complexes and high-rise residential apartments. However, these guidelines primarily focus on access for fire engines, and target communal areas and entrances/exits without specifications for individual homes [42]. Legislative frameworks and building guidelines must ensure EMS accessibility for patient transportation and evacuation during emergencies that can occur at any time [43]. We propose that existing building-related laws and guidelines should be updated to include comprehensive provisions for residential structures to facilitate the rapid transport of patients during emergencies.

Conclusions

Higher floors and specific in-home locations (i.e., toilets, bathrooms, and kitchens) were associated with longer on-scene time and delays in patient transportation to the hospital. To mitigate the operational burden on EMS personnel and improve patient outcomes, further studies focusing on patients’ consciousness levels are needed. These efforts may contribute to building codes that incorporate universal accessibility standards to ensure that paramedics can rapidly reach and transport emergency patients regardless of location.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

EMS

Emergency Medical Services

WBGT

Wet-Bulb Globe Temperature

IQR

Interquartile Range

MRs

Mean Ratios

aMRs

Adjusted Mean Ratios

CI

Confidence Intervals

Author contributions

Conceptualization: RTI, KU, CT, and HI; Data curation: RTI, CT; Formal analysis: RTI; Funding acquisition: CT; Investigation: n. a.; Methodology: RTI, KU, and CT; Resources: RTI and CT; Software: n. a.; Supervision: CT, TS and KT; Validation: n. a.; Visualization: RTI; Writing-original draft: RTI; Writing-review and editing: RTI, KU, CT, HI, TS, and KT.

Funding

This work was supported by JSPS KAKENHI Grants-in-Aid for Scientific Research [Grant numbers JP23H03228 and JP23K27918].

Data availability

The datasets generated and analyzed in this study are not publicly available due to licensing restrictions governing their use, but are available from the corresponding author upon reasonable request and with permission from the participating municipal fire department.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Ethics Committee of the Hiroshima University (Approval number E2020-2381).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.