Changes in prehospital emergency medical services dispatch mission patterns and their public health implications during the COVID-19 pandemic in Beijing, China

Abstract

Background

The COVID-19 pandemic placed unprecedented strain on global prehospital emergency medical services (EMS), particularly in megacities, acting as a stress test for health system resilience. However, comprehensive analyses of its impact on EMS dispatch patterns, mission types, and patient demographics in megacities like Beijing remain limited. This study aims to evaluate how the pandemic reshaped EMS operations in Beijing across multiple dimensions.

Methods

This population-based retrospective cohort study analysed 452,824 dispatch tasks from the Beijing Prehospital Emergency Centre during matched periods (20 January–20 July) in 2019 (pre-pandemic) and 2020 (pandemic). Using chi-square tests and multivariable Poisson regression models with interaction terms, we examined changes in dispatch volume, mission type distribution, patient demographics, and call locations before and during the pandemic.

Results

Among the 452,824 tasks (pre-pandemic: 225,318; pandemic: 227,506), significant changes were observed. Overall EMS dispatch volume decreased during the pandemic (IRR = 0.69, 95% CI: 0.69–0.69), though monthly variability increased significantly (χ² = 25.59, P < 0.001). Mission types underwent substantial restructuring: conventional emergency responses declined by 4.2% points, while infectious disease transfers increased over 100-fold and non-emergency transports quadrupled (1.18% to 4.40%). Significant demographic shifts occurred, with increased service demands from elderly populations (interaction IRR up to 2.24 for oldest age group) and a pronounced relocation of emergency incidents from public spaces to residential areas (50.14% to 56.42%).

Conclusions

The COVID-19 pandemic fundamentally transformed Beijing’s EMS ecosystem, characterized by reduced overall volume but increased volatility, service type restructuring, and spatial redistribution toward residential areas serving older populations.

Introduction

Infectious disease pandemics disrupt global healthcare systems, significantly straining prehospital emergency dispatch operations and critical care resources, thereby presenting a critical test for public health emergency preparedness [1]. Megacities (urban agglomerations exceeding 10 million residents as defined by the United Nations [2]) face heightened challenges during outbreaks due to population density, inequitable distribution of healthcare resources, and disruptions to medical referral systems [3, 4]. Beijing, a representative Chinese megacity and the nation’s political and cultural center, provides a critical case study for evaluating epidemic control strategies and health system resilience [5].

Prehospital emergency medical services (EMS) serve as the primary gateway to emergency care and the frontline response during public health emergencies, playing an indispensable role in health emergency and disaster risk management (Health-EDRM) [6]. During pandemics, megacity EMS systems confront dual pressures: addressing pre-existing resource allocation disparities while adapting to surges in service demand [7, 8]. Existing research focuses predominantly on in-hospital workflows, leaving the pandemic impacts on prehospital EMS relatively understudied [9]. While Handberry et al. observed decreased US EMS dispatch volumes during COVID-19, they did not analyze spatiotemporal variations in mission types [10]. Harrison et al. linked reduced dispatch activity to patient refusal rates but lacked systematic operational assessments of EMS task patterns [11].

Considering the impact of the pandemic and decisions made by health authorities (particularly political decision-makers) on patient healthcare-seeking behaviour, this study examined the effects of COVID-19 on Beijing’s emergency dispatch operations. Beijing recorded a total of 988 confirmed COVID-19 cases in 2020, with 928 of these diagnosed between 20 January and 20 July. The city experienced three distinct peaks in daily new confirmed cases [12] (Fig. 1). Concurrently, Beijing promptly initiated and adjusted its emergency response levels in accordance with the evolving epidemic situation [13]. Therefore, the period from 20 January to 20 July, 2020, provides a representative window of interest for studying the influence of COVID-19 and policy interventions on prehospital EMS dispatch mission patterns. It captures the city’s initial adaptation under the highest response level, characterized by escalating public and professional health awareness and improving diagnostic capabilities, and concludes before the transition to a sustained period of low (primarily imported) case incidence after the response level was lowered.

Fig. 1.

Daily COVID-19 cases in Beijing during the pandemic period

Methods

Study design and setting

This retrospective observational cohort study analysed prehospital EMS dispatch records from the Beijing Emergency Medical Centre. We compared two time-matched cohorts: (1) the pre-pandemic cohort (20 January to 20 July 2019) and (2) the pandemic cohort (the identical period in 2020, which spanned from Beijing’s first confirmed COVID-19 case to the containment of the Xinfadi market outbreak.

Participants

All dispatch records during the study periods were screened. We excluded: (1) Dispatches initiated by EMS vehicles (non–caller-activated); (2) Duplicate calls (multiple records for the same incident); (3) Patient-initiated cancellations before ambulance arrival; (4) Records with > 20% missing critical variables (e.g., timestamps for response time calculation).

Variables

Data extracted from dispatch records encompass mission characteristics and patient socio-demographic information. Specifically, we collected: (1) Dispatch mission details: average daily mission volume, mission type, and response time. Mission types include: ① Emergency mission: involving acute patient onset where EMS provides initial medical contact; ② Transfer mission: The patient has already been seen at a medical institution for this episode but requires EMS transport to another hospital; ③ Non-urgent transport mission: The patient has non-urgent medical needs, with EMS undertaking medical transport duties, such as transporting mobility-impaired patients for hospital admission or follow-up appointments; ④ Infectious transport mission: Transporting confirmed or suspected infectious disease patients, or close contacts, to specialized infectious disease hospitals for treatment or to isolation facilities; ⑤ Non-Medical transport mission: Other non-medical tasks, such as collecting stretchers or transporting deceased persons. (2) Patient socio-demographic characteristics: gender, age, and location of the call.

Statistical analysis

All analyses were performed using R version 4.3.1 (R Foundation for Statistical Computing) and MATLAB R2023a (MathWorks Inc). Continuous variables were assessed for normality using Shapiro-Wilk tests. Normally distributed data are reported as mean (SD) and compared using independent t tests (equal variance assumed) or Welch t tests (unequal variance). Non-normally distributed data are reported as median (IQR) and compared using Mann-Whitney U tests. Categorical variables are reported as frequency (%) and compared using Pearson χ² tests. Poisson regression analyses were employed to examine factors associated with EMS dispatch counts, given the count nature of the outcome variable. Model results are presented as incidence rate ratios (IRR) with 95% confidence intervals. Model fit was assessed using deviance residuals and Pearson chi-square statistics. Effect sizes were calculated for significant outcomes. A two-sided P < 0.05 defined statistical significance. Missing data were excluded from analyses.

Ethical approval

This study was approved by the Peking University Third Hospital Medical Science Research Ethics Committee, which granted a waiver of informed consent as the research used deidentified retrospective data. The study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Result

Changes in EMS dispatch volume

A total of 452 824 dispatch tasks were analysed: 225 318 (49.8%) during the pre-pandemic period (2019) and 227 506 (50.2%) during the pandemic period (2020). Monthly variations in mean daily mission volume were significantly greater during the pandemic (χ² = 25.59, P < 0.001). Compared with 2019, mean daily volume increased in January 2020 (1255 vs. 1187 tasks/day), June 2020 (1443 vs. 1270 tasks/day), and July 2020 (1387 vs. 1271 missions/day), while decreasing in April 2020 (1133 vs. 1275 tasks/day) (Table 1; Fig. 2).

Table 1.

Changes in EMS dispatch tasks before and during the COVID-19 pandemic

Variable	Pre-pandemic, No. (%)	Pandemic, No. (%)	χ2	P
	N = 225,318	N = 227,506
Daily missions, n	1238	1243		0.159
Average daily missions grouped by month			25.587	<0.001
January	1187	1255
February	1130	1110
March	1214	1200
April	1275	1133
May	1290	1262
June	1270	1443
July	1271	1387
Mission type, n (%)			13,792	<0.001
Emergency mission	172,978(76.77)	165,170(72.60)
Transfer mission	33,662(14.94)	26,132(11.49)
Non-urgent transport mission	2649(1.18)	10,011(4.40)
Infectious transport mission	78(0.03)	8537(3.75)
Non-medical transport mission	15,951(7.08)	17,656(7.76)

Fig. 2.

Daily EMS mission volume and COVID-19 case counts (January–July 2019 vs. 2020)

To further validate the impact of temporal factors on task volume, we constructed a Poisson regression model with EMS dispatch mean daily volume as the dependent variable and period (pre-pandemic/pandemic) and calendar month as primary predictors (Table 2). Model results indicate that task volume during the pandemic period was significantly lower than in the pre-pandemic period (IRR = 0.69, 95% CI: 0.69–0.69, P < 0.001), suggesting a reduction to 69% of pre-pandemic levels. The IRR for each calendar month variable exceeded 1 and was statistically significant (P < 0.001), indicating that task volumes in all months were significantly higher than the reference month (January). The most pronounced increase occurred in May (IRR = 2.35, 95% CI: 2.32–2.39), followed by March (IRR = 2.08, 95% CI: 2.05–2.11) and April (IRR = 2.07, 95% CI: 2.04–2.09). These findings demonstrate the independent influence of both the pandemic period and calendar month on EMS task volume. This pattern demonstrates that despite an overall pandemic-related decline, pronounced monthly surges persisted, likely associated with the phased nature of the pandemic and shifts in societal activity.

Table 2.

Poisson regression for EMS dispatch task volume, period and month

Independent variable	β	SE	z	p	IRR[95%CI]
Pandemic	-0.370	0.003	-123.317	< 0.001	0.69[0.69,0.69]
January	Reference
February	0.498	0.007	70.369	< 0.001	1.65[1.62,1.67]
March	0.733	0.007	106.376	< 0.001	2.08[2.05,2.11]
April	0.726	0.007	104.794	< 0.001	2.07[2.04,2.09]
May	0.856	0.007	125.232	< 0.001	2.35[2.32,2.39]
June	0.521	0.007	76.295	< 0.001	1.68[1.66,1.71]
July	0.235	0.007	32.280	< 0.001	1.27[1.25,1.28]
constant	4.413	0.006	736.485	< 0.001

Changes in mission types

The chi-square test revealed a significant shift in the distribution of pre-hospital emergency task types during the pandemic compared to the pre-pandemic period (χ² = 13,792, P < 0.001), indicating the pandemic’s profound impact on the structure of emergency services (Table 1). Specifically: (1) The proportion of routine emergency tasks decreased from 76.77% to 72.60%, a reduction of 4.17% points; (2) Infectious disease-related transport assignments experienced explosive growth, surging from an over 100-fold increase (3) Non-emergency transport assignments rose from 1.18% to 4.40%, nearly quadrupling. Among other mission types, inter-hospital transfers decreased from 14.9% to 11.5%, while non-medical missions saw a marginal increase from 7.1% to 7.8%.

To further investigate the association between task type and period (pre-pandemic/pandemic), we constructed another Poisson regression model with EMS dispatch task volume as the dependent variable, and with task types and time periods as the primary predictors (Table 3). The model results indicated a significant overall reduction in task volume during the pandemic period (IRR = 0.85, 95% CI: 0.83–0.87, P < 0.001). The main effects for each task type were statistically significant (P < 0.001): Emergency mission (IRR = 4.34, 95% CI: 4.27–4.41) and Transfer mission (IRR = 2.72, 95% CI: 2.67–2.77) were significantly higher than the Non-medical transport missions, whereas Non-urgent transport mission (IRR = 0.12, 95% CI: 0.12–0.13) and Infectious transport mission (IRR = 0.02, 95% CI: 0.02–0.03) were significantly lower.

Table 3.

Poisson regression for task types and time periods

Independent variable	β	SE	z	p	IRR[95%CI]
Pandemic	-0.161	0.011	-14.695	< 0.001	0.85[0.83,0.87]
Non-medical transfer mission	Reference
Emergency mission	1.468	0.008	177.417	< 0.001	4.34[4.27,4.41]
Transfer mission	1.001	0.010	104.175	< 0.001	2.72[2.67,2.77]
Non-urgent transport mission	-2.094	0.021	-99.815	< 0.001	0.12[0.12,0.13]
Infectious transport mission	-3.776	0.114	-33.271	< 0.001	0.02[0.02,0.03]
Pandemic × Emergency mission	-0.070	0.011	-6.111	< 0.001	0.93[0.91,0.95]
Pandemic × Transfer mission	-0.358	0.014	-26.187	< 0.001	0.7[0.68,0.72]
Pandemic × Non-urgent transport mission	1.039	0.024	42.546	< 0.001	2.83[2.7,2.97]
Pandemic ×Infectious transport mission	3.005	0.114	26.295	< 0.001	20.18[16.13,25.24]
constant	4.090	0.008	516.559	< 0.001

Crucially, the interaction term between period and task type revealed differential pandemic impacts across specific task categories. Emergency mission (IRR = 0.93, 95% CI: 0.91–0.95, P < 0.001) and Transfer mission (IRR = 0.70, 95% CI: 0.68–0.72, P < 0.001) further declined during the pandemic, indicating suppression of routine emergency services; Conversely, Non-urgent transport mission (IRR = 2.83, 95% CI: 2.70–2.97, P < 0.001) and Infectious transport mission (IRR = 20.18, 95% CI: 16.13–25.24, P < 0.001) increased significantly during the pandemic, with infectious transport mission call-outs rising over twentyfold. This aligns with observed increases in infectious disease transfers and non-emergency transport assignments.

Changes in patient and location characteristics

Significant changes in demographic characteristics were observed during the study period (Table 4). Regarding gender distribution, the proportion of male patients increased from 53.02% before the pandemic to 54.07% (χ² = 71.35, P < 0.001). Age distribution exhibited pronounced structural shifts (χ² = 315.46, P < 0.001): the proportion of infants and children (0–12 years) decreased from 2.62% to 2.17%, adolescents (12–18 years) saw a slight decline from 0.87% to 0.83%, the proportion of young adults (18–45 years) increased significantly from 24.34% to 26.24%, while the proportion of middle-aged adults (45–65 years) decreased from 25.20% to 23.89%. younger elderly (65–75 years) increased marginally from 15.24% to 15.39%, elderly (75–85 years) decreased from 17.42% to 16.61%, and very elderly (≥ 85 years) increased from 14.31% to 14.88%. Significant changes also occurred in the distribution of call locations (χ² = 2486.2, P < 0.001): calls originating from residential areas increased from 50.14% to 56.42%, while proportions from public spaces (24.82% to 21.45%) and healthcare facilities (24.16% to 21.89%) decreased, with calls from schools falling markedly (0.88% to 0.24%).

Table 4.

Patient demographics and call locations

Variable	Pre-pandemic, No. (%)	Pandemic, No. (%)	χ2	P
	N = 225,318	N = 227,506
Female, n (%)	105,848 (46.98)	105,814 (46.51)	9.888	0.002
Age, year, M (IQR)	62 (40–80)	62 (40–80)		<0.001
Age group, n (%)				<0.001
Infants and children [0,12]	5900 (2.62)	4932 (2.17)
Adolescents [12,18)	1967 (0.87)	1885 (0.83)
Young adults [18,45)	54,840 (24.34)	59,692 (26.24)
Middle-aged adults [45,65)	56,774 (25.20)	54,348 (23.89)
Younger elderly [65,75)	34,344 (15.24)	35,003 (15.39)
Elderly [75,85)	39,261 (17.42)	37,788 (16.61)
Very elderly (≥ 85)	32,232 (14.31)	33,858 (14.88)
Call location, n (%)			2486.2	<0.001
Residential	112,969(50.14)	128,355(56.42)
Public space	55,929(24.82)	48,808(21.45)
Medical facility	54,446(24.16)	49,808(21.89)
School	1974(0.88)	535(0.24)

IQR = Interquartile Range; Μ = Median

To investigate the impact of demographic factors on task volume, a Poisson regression model incorporating gender, age group, location of call, time period, and their interaction terms was constructed (Table 5). The model results indicated a significant overall reduction in task volume during the pandemic period (IRR = 0.28, 95% CI: 0.26–0.29, P < 0.001). The main effect of gender showed lower task volumes among females (IRR = 0.76, 95% CI: 0.75–0.77, P < 0.001). Age group analysis, with infants and children as the reference group, revealed the lowest task volume among adolescents (IRR = 0.50, 95% CI: 0.47–0.53) and the highest among young adults (IRR = 6.96, 95% CI: 6.71–7.22), demonstrating a pattern that rose to a peak in young adults and then declined with advancing age. Location analysis revealed significantly fewer tasks originating from healthcare facilities (IRR = 0.44, 95% CI: 0.43–0.44) and schools (IRR = 0.07, 95% CI: 0.07–0.08) compared to residential areas.

Table 5.

Poisson regression of demographic factors on task volume

Independent variable	β	SE	z	p	IRR[95%CI]
Gender (Female)	-0.279	0.006	-46.110	< 0.001	0.76[0.75,0.77]
Infants and children [0,12]	Reference
Adolescents [12,18)	-0.689	0.031	-21.940	< 0.001	0.5[0.47,0.53]
Young adults [18,45)	1.940	0.019	102.291	< 0.001	6.96[6.71,7.22]
Middle-aged adults [45,65)	1.732	0.019	90.882	< 0.001	5.65[5.45,5.87]
Younger elderly [65,75)	1.025	0.020	51.099	< 0.001	2.79[2.68,2.9]
Elderly [75,85)	1.111	0.020	55.649	< 0.001	3.04[2.92,3.16]
Very elderly (≥ 85)	0.869	0.021	42.309	< 0.001	2.39[2.29,2.48]
Medical facility mission	-0.832	0.006	-137.957	< 0.001	0.44[0.43,0.44]
School mission	-2.596	0.023	-113.242	< 0.001	0.07[0.07,0.08]
Pandemic	-1.286	0.031	-41.237	< 0.001	0.28[0.26,0.29]
Gender × Pandemic	-0.043	0.009	-4.892	< 0.001	0.96[0.94,0.97]
Pandemic × adolescents	0.309	0.048	6.493	< 0.001	1.36[1.24,1.5]
Pandemic × young adults	0.461	0.029	15.720	< 0.001	1.59[1.5,1.68]
Pandemic × middle-aged adults	0.404	0.030	13.663	< 0.001	1.5[1.41,1.59]
Pandemic × younger elderly	0.553	0.031	17.877	< 0.001	1.74[1.64,1.85]
Pandemic × elderly	0.635	0.031	20.562	< 0.001	1.89[1.78,2]
Pandemic × very elderly	0.806	0.031	25.622	< 0.001	2.24[2.1,2.38]
Pandemic × Medical facility mission	0.409	0.009	46.444	< 0.001	1.51[1.48,1.53]
Pandemic×School mission	-0.574	0.049	-11.679	< 0.001	0.56[0.51,0.62]
Constant	4.392	0.020	216.327	< 0.001

Interaction term analysis indicated distinct patterns of change across factors during the pandemic. The interaction term between gender and period indicated a further increase in the proportion of male callers during the pandemic (IRR = 0.96, 95% CI: 0.94–0.97, P < 0.001). Both age group and period interaction terms exhibited positive effects, with magnitude increasing with age: from an IRR of 1.36 (95% CI: 1.24–1.50) among adolescents to 2.24 (95% CI: 2.10–2.38) among the very elderly, indicating a marked relative increase in call demand among older adults during the pandemic. The interaction term for location revealed a relative increase in calls from medical facility during the pandemic (IRR = 1.51, 95% CI: 1.48–1.53), while calls from schools decreased further (IRR = 0.56, 95% CI: 0.51–0.62).

Discussion

Temporal dynamics and pandemic linkages

Consistent with global reports of reduced EMS demand [14, 15], Beijing experienced overall lower mission volumes during most pandemic months. This phenomenon may be attributed to social activity restrictions, altered public healthcare-seeking behavior, and reduced non-urgent medical demand during the pandemic. However, this decline was not uniformly distributed, with markedly heightened monthly volatility (χ² = 25.59, P < 0.001), indicating that the pandemic’s impact on emergency medical services exhibited significant spatiotemporal heterogeneity. Contextualized with Beijing’s pandemic transmission patterns and control policies [13], significant surges occurred at pandemic inflection points: a 5.7% increase following the first confirmed case (January 2020), and 13.6% and 9.1% rises during the Xinfadi market outbreak (June-July 2020). This pattern aligns with conceptualizing COVID-19 as a sustained surge event [16], challenging conventional EMS response paradigms.

The Poisson regression model further validated the independent effects of period and month on task volume: despite an overall reduction in task volume during the pandemic, as shown in Fig. 2, the average daily EMS call volume during the pandemic period in specific months (June and July) was significantly higher than pre-pandemic levels. This fluctuation pattern may reflect the phased characteristics of the pandemic’s progression.

Research findings reveal the complex impact of the pandemic on pre-hospital emergency services: despite an overall decline in mission volume, pronounced monthly fluctuations suggest that emergency resource allocation must be dynamically adjusted according to the temporal and spatial characteristics of the pandemic’s progression to address shifting service demands across different phases. Our findings validate the application of the emergency management cycle framework (mitigation, preparedness, response, recovery) to dynamically reallocate EMS resources during future pandemics, consistent with the WHO Health-EDRM approach to health emergency management.

Resource competition and system implications

This study found that the COVID-19 pandemic not only altered the volume of pre-hospital EMS task volume but also profoundly reshaped the task composition. Significant shifts in task distribution (χ² = 13,792, P < 0.001) indicate the pandemic exerted structural impacts on the EMS system. This finding transcends mere quantitative changes, revealing patterns of healthcare service reconfiguration during public health crises.

The decline in routine Emergency mission from 76.77% to 72.60% may be attributed to multiple factors: on the one hand, social distancing measures and lockdown policies reduced traditional emergency demands such as traffic accidents and workplace injuries; on the other hand, public concerns over infection risks may have deterred non-urgent calls for assistance. Concurrently, the reduction in Transfer mission (14.9% to 11.5%) reflects hospitals’ resource consolidation during the pandemic to safeguard critical care capacity, alongside the suspension of non-essential medical services.

The most pronounced change was the explosive growth in infectious disease-related transfers (0.03% to 3.75%), exhibiting an incidence rate ratio of 20.18 (95% CI: 16.13–25.24). This phenomenon directly reflects the new demands placed on the emergency medical system by pandemic control measures—the transport of large numbers of suspected or confirmed patients became the new normal for EMS. Concurrently, the nearly fourfold increase in Non-urgent transport mission (1.18% to 4.40%, IRR = 2.83) suggests a shift towards ‘non-urgent’ medical services during the pandemic. This increase can be attributed to factors including the disruption of routine community-based healthcare and tiered medical services, fears of in-hospital infection and limited transportation options during the pandemic period. This shift underscores how pandemics can transform EMS from a purely emergency service into a critical backbone for maintaining essential patient flow when conventional systems are compromised.

The interaction effect between mission type and period in the Poisson regression model further demonstrates the pandemic’s differential impact. The suppression of routine Emergency mission (Emergency mission IRR = 0.93; Transfer mission IRR = 0.70) stands in stark contrast to the surge in public health-related missions. This reflects the functional restructuring of the emergency medical system during the pandemic—shifting from traditional emergency medical response towards a diversified service system that also addresses public health emergencies.

The diversion of finite EMS capacity away from core emergency functions risks compromising conventional care delivery, a phenomenon similarly observed in other megacities during respiratory disease outbreaks. This resource competition exemplifies the critical need for coordinated response systems that can maintain essential health services during emergencies, as emphasized in the Sendai Framework for Disaster Risk Reduction. Establishing dedicated infectious disease transport systems, analogous to Germany’s Central COVID-19 Coordination Centers [17], could preserve routine EMS operational integrity during outbreaks and enhance overall health system preparedness.

Sociodemographic shifts and public health implications

This study reveals the profound impact of the COVID-19 pandemic on the demographic profile and geographical distribution of pre-hospital emergency service users. Shifts in demographic characteristics reflect the pandemic’s differential effects across population groups, while changes in call locations vividly illustrate the restructuring of social activity patterns during a public health crisis.

Structural shifts in age distribution warrant particular attention: the rise in young adults (24.34% to 26.24%) may relate to this cohort’s status as the primary workforce maintaining high activity levels during the pandemic, or be driven by pandemic-induced mental health crises [18, 19] and health anxieties [20]. Meanwhile, the pattern of change within the elderly cohort—a stable proportion of younger seniors and an increase in the proportion of older seniors (14.31% to 14.88%)—reveals differential vulnerability across age strata. As the cohort most prone to chronic diseases and pronounced physical decline, the elderly face heightened vulnerability due to isolation-related morbidity [21] and the exacerbation of multimorbidity [22, 23]. Consequently, their demand for healthcare services became particularly pronounced under pandemic pressures.

The dramatic shift in the distribution of call locations (χ² = 2486.2, P < 0.001) directly reflects the impact of epidemic control measures on the use of social spaces. The substantial increase in residential area called proportions (50.14% to 56.42%) closely aligned with restrictive measures such as stay-at-home orders [24], indicating homes became the primary setting for healthcare needs during the pandemic. Correspondingly, reduced calls from public spaces reflected diminished social activity levels, while the sharp decline in school-related calls (0.88% to 0.24%) tangibly demonstrated the policy impact of educational institution closures. Although the absolute proportion of calls to healthcare facilities decreased, the interaction effect in the Poisson regression (IRR = 1.51) indicates a relative increase in their importance. This may stem from healthcare settings’ unique status as high-risk infection venues during the pandemic.

The Poisson regression model reveals complex mechanisms underpinning the impact of various factors on task volume. The inverted U-shaped distribution of age effects (younger age group IRR = 6.96 to older age group IRR = 2.39) aligns with life-course health risk patterns. However, the age gradient in interaction effects among the elderly during the pandemic (IRR rising from 1.36 to 2.24) strongly suggests that the relative impact of the pandemic on demand for emergency services becomes more pronounced with increasing age. This finding supports the theory of ‘age-stratified vulnerability,’ indicating that the elderly face cumulative effects from multiple health threats during emergencies [25].

Policy and system-level recommendations

These findings collectively illustrate how public health emergencies can fundamentally alter emergency service ecosystems. EMS systems must develop adaptive strategies that maintain core emergency response capabilities while building surge capacity for specialized pandemic-related functions. Future preparedness plans should incorporate flexible resource allocation models that account for temporal fluctuations, evolving service demands, and changing population vulnerability patterns during public health crises. The lessons learned from Beijing’s experience provide valuable insights for emergency medical systems worldwide in building resilience against future pandemics.

Based on our research findings, we propose the following three key strategies to enhance public health emergency preparedness: (1) A reduction in overall workload during outbreaks does not imply diminished service pressure; rather, heightened volatility necessitates more flexible resource allocation strategies; (2) The system must rapidly establish large-scale dedicated infectious disease transport channels while maintaining routine emergency capabilities, and adapt to increased non-emergency transport demands. Therefore, establishing parallel transport pathways for infectious diseases and routine emergencies ensures system functionality during demand surges; (3) Emergency resource allocation must account for demographic shifts in emergency patterns, particularly enhancing focus on elderly populations.

Integrating mental health support and community care into emergency response frameworks is essential to address evolving healthcare demand structures. Future emergency service development must enhance resilience in responding to specific populations and scenarios while maintaining traditional service capabilities, thereby addressing structural challenges posed by major public health events. These recommendations align with the principles of multisectoral collaboration and community engagement emphasised in the World Health Organisation’s Health Emergency Response Management Framework (Health-EDRM), offering concrete strategies for building more resilient health systems in global megacities.

Limitations

This study has several limitations. First, the analysis focused on a period of significant pandemic fluctuation in Beijing. While it included most newly confirmed cases, it did not encompass the entire year of 2020. Second, unmeasured confounders (e.g., socioeconomic status, comorbidities, health-seeking behaviours) may influence EMS utilisation patterns. This study relied solely on emergency response levels to represent local health policies, failing to comprehensively elucidate policy impacts on patient healthcare-seeking behaviors. Third, the lack of staff-level data precludes assessment of pandemic impacts on workforce capacity, provider well-being, or operational adaptations at the provider level. Despite these constraints, our findings provide valuable insights for optimising prehospital EMS resilience and inform public health strategies for health emergency preparedness in megacities.

Conclusion

This study demonstrates that the COVID-19 pandemic significantly reshaped emergency medical services in Beijing across three critical dimensions. First, while overall dispatch volume decreased, monthly variability increased substantially, indicating that pandemic impacts were temporally heterogeneous rather than uniformly distributed. Second, the service mix underwent fundamental restructuring, with conventional emergency responses declining while infectious disease transfers surged over 100-fold and non-emergency transports increased nearly fourfold. Third, significant demographic and geographic shifts occurred, characterized by increased service demands from elderly populations and a pronounced relocation of emergency incidents from public spaces to residential areas.

Author contributions

Dr ZR, Dr SHD, and Dr HZ contributed equally as co–first authors. ZR, SHD and HZ conceived and designed the study. SL supervised the study conduct. YZJ provided statistical expertise. JJZ and QBM managed data quality. ZR drafted the manuscript. All authors contributed to data interpretation, critically revised the manuscript for important intellectual content, and approved the final version.

Funding

This study was supported by the National Natural Science Foundation of China (Grant No. 7217040327) and The Special Fund of the National Clinical Key Specialty Construction Program, P. R. China (2022) 301–2305. The funders had no role in the design of the study; collection, analysis, and interpretation of data; or in writing the manuscript.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The Institutional Review Board of Peking University Third Hospital approved this study (Approval No: S2020297). The requirement for informed consent was waived due to the use of de-identified retrospective data.

Consent for publication

Not applicable.

Guarantor

HC and QBM are the guarantors.

Competing interests

The authors declare no competing interests.