Impact of COVID-19 pandemic phases on emergency medical services reaction times in Southern Poland

Abstract

This study analyses Emergency Medical Services (EMS) reaction times and call volumes in Małopolska, Poland, during and after the COVID-19 pandemic (2020–2023). The main objective was to evaluate healthcare system strain under different phases of the pandemic and to identify temporal patterns affecting EMS performance. Using statistical analyses, we investigated correlations between EMS call volumes and reaction times across various temporal scales, including hourly, daily, monthly, and yearly patterns. Results showed significant increases in EMS reaction times during pandemic peaks, especially during the vaccination rollout phase, correlating strongly with increased call volumes. The highest reaction delays occurred in December, weekends, and evening hours. Despite high call volumes, reaction times improved in later pandemic stages, indicating effective adaptation of EMS strategies. Seasonal and daily cyclicality in call patterns suggests opportunities for optimized resource allocation. The study highlights the resilience and adaptability of EMS systems under crisis conditions and recommends dynamic allocation strategies, accurate demand forecasting, and improved crisis communication. These findings offer critical insights for enhancing future EMS crisis management strategies and operational preparedness.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-14324-1.

Introduction

The SARS COVID-19 pandemic was a trial period for the efficiency of medical services working on the front lines with patients in immediate need of assistance. This trial served as a practical test of public safety plans, which included crisis management strategies for Emergency Medical Services (EMS) fleets in the event of disasters, war, or other unforeseeable events such as a pandemic. Assessing whether society successfully passed this test remains a complex and open question, given the millions of lives lost worldwide and the ongoing costs of the pandemic that we continue to bear today. The balance sheet of gains and losses is apparent, yet it is essential to recognize that the pandemic provided us with vast knowledge and data, revealing how the healthcare system functions under extreme time pressure and high demand. Therefore, through retrospective analysis of the collected information, we can draw conclusions that will better equip us for similar events in the future.

First and foremost, pandemic response strategies evolved depending on the phase of the pandemic and were tailored to local conditions¹. Each strategy, however, came with its own set of socio-economic costs and caused significant alterations in healthcare delivery, particularly in emergency services^2,3. For example,⁴ noted that healthcare policies during the pandemic focused on infection prevention and control, which directly affected service delivery in various sectors, including EMS. The pandemic's onset led to a marked decrease in emergency medical missions, as reported by who observed a significant decline in EMS calls, particularly for conditions like acute stroke, which traditionally require immediate intervention. This decline was attributed to a combination of public fear of infection and the reallocation of healthcare resources to manage COVID-19 cases, leading to longer response times and increased pressure on EMS systems⁵. The psychological impact of the pandemic on both patients and healthcare providers also played a crucial role in shaping EMS response dynamics⁶. The fear of infection led many individuals to forego necessary medical care, contributing to a rise in out-of-hospital cardiac arrests and other emergencies that could have been mitigated with timely intervention⁷. Furthermore, socioeconomic factors played a critical role in shaping healthcare access during the pandemic, with key determinants such as income inequality, level of education, and geographic location contributing to disparities in the availability and timeliness of emergency medical services^8–12. Tuczyńska et al.¹³ found that the perception of healthcare accessibility and quality varied widely among different socioeconomic groups in Poland, with many individuals reporting decreased utilization of healthcare services due to the conversion of medical facilities to accommodate COVID-19 patients. This shift in resource allocation further complicated the EMS landscape, as the availability of services became increasingly limited, leading to longer response times and potentially adverse health outcomes for patients requiring immediate assistance.

Moreover, the pandemic exacerbated existing barriers to healthcare access. Núñez et al.¹⁴ highlighted those chronic conditions, such as diabetes and hypertension, faced significant management challenges due to reduced access to facility-based care during the pandemic. This situation was compounded by the fact that many patients delayed seeking emergency care due to fears of contracting the virus in healthcare settings. However, Hwang and Heo reported a different pattern, indicating that individuals with lower education levels encountered fewer healthcare accessibility issues during the COVID-19 pandemic, possibly due to different risk perceptions¹⁵. This indicates a complex interplay between public perception and actual service availability, which ultimately affected EMS response times.

The increased workload for EMS personnel during the pandemic cannot be overlooked. The pandemic saw a significant rise in EMS calls, such as a 330% increase in Catalonia, Spain, in March 2020, with up to 40,000 calls during peak days¹⁶. To address this issue, a potential solution to mitigate increased EMS demand lies in accurate demand forecasting and strategic vehicle deployment, both essential for maintaining high standards of care and safety^15,17–21. On the other hand, Blek et al.⁵ noted that while the number of missions decreased, the complexity and duration of each mission increased due to the need for thorough COVID-19 screenings and the use of personal protective equipment (PPE). This shift not only lengthened the time required for each call but also placed additional strain on EMS resources, leading to longer wait times for patients in need of urgent care.

In addition to these challenges, the pandemic prompted innovative adaptations within EMS²². The integration of telehealth services and enhanced communication through social media platforms emerged as vital tools for maintaining patient engagement and care continuity during the crisis²³. These adaptations not only facilitated better communication between healthcare providers and patients but also served to alleviate some of the pressures faced by EMS by providing alternative avenues for care²⁴.

As the pandemic progressed, it became evident that the healthcare system's resilience was tested, leading to significant organizational changes. Glette's research²⁵ highlighted that the pandemic catalyzed previously unthinkable changes in healthcare delivery, including the reorganization of services and enhanced collaboration among healthcare providers. This adaptability was crucial for maintaining service delivery amidst the unprecedented challenges posed by COVID-19.

Given the above issues, one of the key pieces of information appears to be EMS response time, which generally refers to the total time elapsed from the moment the medical dispatcher receives the emergency call to the moment the medical team physically arrives at the incident location. Within the context of EMS literature, it is crucial to differentiate between two closely related yet distinct terms: EMS reaction time and EMS response time²⁶. EMS reaction time specifically describes the interval from receiving a call to the dispatching of an available ambulance crew, reflecting the efficiency and organizational capacity of dispatch centers in managing incoming calls²⁷. Conversely, EMS response time encompasses not only this initial reaction interval but also the actual travel time required by the ambulance to reach the patient at the scene^28,29. Thus, EMS reaction time is inherently embedded within the broader measurement of EMS response time, serving as a key factor that directly influences the overall responsiveness of emergency services.

This broader understanding of EMS reaction time allows us to more comprehensively reflect the actual operational burden on healthcare systems at any given moment, capturing both organizational efficiency and external factors such as road conditions and geographic location. Moreover, total EMS response time can be significantly impacted by hospital-related delays, including ambulance queues that frequently occur while transferring patients to medical facilities ambulances caught in these queues remain unavailable to respond promptly to new calls, further highlighting the interdependent nature of healthcare resource management and emergency service performance^27,29. Therefore, the objective of this study was to measure EMS reaction time, as it serves as a valuable indicator reflecting the overload of emergency call volumes and highlighting operational challenges faced by healthcare services during the pandemic.

Despite extensive research on COVID-19’s impact on healthcare services globally, there remains limited quantitative understanding of how EMS reaction times correlated specifically with call volumes during different pandemic phases at regional scales. This study addresses this research gap by analyzing comprehensive EMS data from southern Poland (Małopolska region). Our primary research question is centered around the relationship between Emergency Medical Services (EMS) response times and call volumes during specific phases of the pandemic. We formulate a null hypothesis (H0) that states there is no significant difference in EMS response times across varying call volumes throughout the pandemic phases, implying that the response times remain constant regardless of the number of calls received. In contrast, our alternative hypothesis (Ha) posits that there is a positive correlation between EMS response times and call volumes. By investigating, our analysis contributes crucial insights into temporal patterns of healthcare system performance during crises, providing evidence-based recommendations for future emergency preparedness.

Materials and methods

This article presents a retrospective analysis of EMS reaction times in Małopolska (Poland) from 2020 to 2023, conducted across districts within the Małopolskie Voivodeship. The analysis begins with a detailed examination of EMS call volumes categorized by district, day of the week, month, and pandemic phases throughout this period. The primary aim of this study is to identify EMS operational patterns during periods of high demand, providing valuable insights that may support future crisis management planning. The lessons learned from this analysis are intended to inform healthcare strategies, thereby ensuring emergency services can effectively respond to population needs. To achieve reliable results, the study employed various statistical techniques and retrospective data analysis approaches. The subsequent sections detail the specific methodological components, including study area characteristics, data sources, processing techniques, and analytical frameworks used in this research.

Study area

The Lesser Poland Voivodeship, or Małopolska, is a province located in southern Poland. Despite its relatively compact area of 15,200 square kilometers, it is home to around 3.4 million people, ranking among the country's most densely populated and economically vital regions. Its capital, Kraków, stands as a key center for culture and commerce. This study retrospectively examines ambulance dispatch data across both the COVID-19 pandemic and the period that followed, utilizing records from the Health Department of the Małopolska Voivodeship Office, which oversees regional ambulance services. The analysis includes 14 of the voivodeship’s 22 counties, encompassing both rural areas and three city-counties to ensure a balanced representation of urban and rural demographics¹⁸. The boundaries of the Małopolska Voivodeship and the analyzed counties are depicted in Fig. 1 below.

Fig. 1.

Map of Małopolskie Voivodeship and analyzed counties. Map created by authors using QGIS software, version 3.34 (https://qgis.org/en/site/).

Poland’s healthcare system is a publicly funded model based on universal access, primarily financed through the National Health Fund. Emergency Medical Services are integrated within this system and operate under standardized national regulations. The EMS structure includes both basic and specialized teams, which respond to emergency calls coordinated by regional dispatch centers.

In 2022, there were approximately 215 hospital emergency departments and over 1600 EMS teams operating nationwide. EMS in Poland operates on a triage and dispatch model, where the medical dispatcher assigns cases based on urgency and location, often under resource constraints during peak times.

Małopolskie Voivodeship region encompasses both urban centers (e.g., Kraków) and rural areas, presenting logistical challenges in EMS deployment. In 2022, the region reported one of the higher EMS call rates per capita compared to the national average, reflecting both the urban density and the influx of tourists, especially in areas like Tatrzański and Nowotarski counties³⁰.

Furthermore, this study aims to recreate the extent of strain placed on healthcare services during the pandemic. As noted in the introduction, we intend to use EMS reaction time as an indicator of this burden. A key research hypothesis and an important factor in a comprehensive case analysis is the presumed statistically significant correlation between total EMS calls and EMS reaction time. To demonstrate this, we begin by examining EMS call patterns and variability across days of the week and months throughout the pandemic and post-pandemic periods. The total number of EMS calls for the analyzed counties from 2020 to 2023 is 785,000, with Kraków City accounting for the highest volume at 261,000 calls. Detailed statistics on the number of calls per county for each year (2020–2023) are presented in Appendix 1.

Data sources and acquisition

In this study, we analyzed over 750,000 emergency call records collected from 14 counties in the Małopolska region between January 2020 and December 2023. The primary dataset includes one entry per ambulance call and covers the period from January 1, 2020, to December 31, 2023. For analytical purposes, this data is aggregated into daily call counts (daily total calls), and grouped by weekly, monthly, or yearly frequencies. To facilitate analysis of within-day variability, the data was aggregated into hourly call counts (hourly total calls).

Our primary goal was to identify trends in EMS reaction times and observe how these trends shifted in relation to key milestones, including the rollout of COVID-19 vaccinations and the end of the pandemic.

All data used were fully anonymized and provided under a formal agreement with the Health Department of the Małopolska Voivodeship Office (Agreement No. WZ-III.63143.12.2023). According to Polish legislation, retrospective analyses of anonymized administrative data that do not involve experiments on humans or animals or the use of identifiable individual information do not require ethical approval from an Institutional Review Board (IRB) or informed consent from participants. All methods were performed in accordance with relevant guidelines and national regulations. Statistical analyses were carried out using the R programming language (version 4.2.0).

Results

The analysis of emergency medical services (EMS) call patterns provides insights into the temporal and spatial variability in response demands across the Małopolskie Voivodeship. By examining monthly, weekly, and hourly call volume (total calls) fluctuations, we can identify distinct trends and seasonal characteristics that impact EMS resource allocation. This section explores variations in total calls throughout different timeframes, highlighting both typical and unique patterns within specific counties. Through a detailed examination of monthly seasonality, weekday fluctuations, and daily cyclicality, we aim to uncover the underlying factors influencing EMS demand and illustrate how these demands vary between urban and rural regions, particularly during the COVID-19 pandemic and subsequent periods.

Monthly variability

A review of the monthly variability in total calls reveals that the lowest average daily call numbers typically occurred in May, while December had the highest. Some counties deviated from this pattern: in Brzeski, the minimum was recorded in September; in Nowy Sącz (urban), the lowest occurred in March; and in Nowotarski, the minimum was observed in both March and May. December generally had the highest daily call volume (see plots aggregated in Appendix 2).

Three counties reached peak call volumes in different months: Nowosądecki in August, Tarnowski in November, and Tatrzański in February. The difference between the maximum and minimum monthly daily call volumes varied from 10% in Tarnowski to 85% in Tatrzański. A monthly disparity between the highest and lowest daily call numbers of 10% to 19% was observed in ten counties: Bocheński (17%), Brzeski (19%), Chrzanowski (15%), the city of Kraków (17%), Krakowski (16%), Nowosądecki (17%), Olkuski (19%), Tarnowski (10%), Wadowicki (14%), and Wielicki (18%). Three counties displayed a difference of 20% to 30% between the maximum and minimum daily calls per month, averaged annually: the city of Nowy Sącz (21%), Nowotarski (24%), and the city of Tarnów (26%). The highest variability occurred in Tatrzański, with a difference of 85% (see Appendix 2).

Examining box-and-whisker plots of monthly call volumes (Appendix 2) reveals distinct county groups:

• Counties with a consistent average call volume year-round, with a slight increase in summer and at year-end: Bocheński, Brzeski, city of Nowy Sącz, city of Tarnów, Olkuski, Tarnowski, Wadowicki, and Wielicki.

• Counties with a stable average call volume year-round, showing a slight or negligible summer increase but a substantial rise at year-end: Chrzanowski, city of Kraków, and Krakowski.

• Counties with a marked increase in call volume during summer months: Nowosądecki, Nowotarski, and Tatrzański.

Weekly patterns

The weekly pattern of daily total calls displays consistent seasonality across all counties. Typically, Mondays and Saturdays see higher call volumes, while mid-weekdays and Sundays have lower call frequencies in each county. For most counties, the variability of typical call volumes remains similar throughout the week (see Appendix 3). However, some counties show unique patterns:

• The lowest average daily total calls on Tuesday were observed in six counties: Chrzanowski (23.9), Kraków (174.5), Krakowski (51.5), Tarnowski (29.8), and Tatrzański (14.3).

• The lowest on Wednesday was noted in two counties: Wadowicki (26.7) and Wielicki (24.2).

• The lowest on Thursday appeared in three counties: Bocheński (17.0), Brzeski (15.3), and Nowosądecki (30.7).

• Saturday had the lowest amount in Olkuski (18.2) and Tarnów (24.4).

• Nowy Sącz county stands out with its lowest call volume on Sunday (17.6).

The highest average daily total calls generally occur on Saturdays or Mondays. Specifically:

• Counties with a Monday peak include Bocheński (19.5), Chrzanowski (26.5), Kraków (185.6), Nowy Sącz (19.3), Olkuski (19.9), Tarnów (27.2), and Wielicki (25.9).

• Counties with a Saturday peak include Brzeski (16.8), Krakowski (55.8), Nowosądecki (34.6), Nowotarski (31.3), Tarnowski (32.8), Tatrzański (17.5), and Wadowicki (29.6).

The greatest difference in average daily total calls throughout the week was found in Tatrzański county (22.4%), while the smallest difference was observed in Kraków city (6.4%). Six counties exhibited differences below 10%: Brzeski (9.8%), Kraków (6.4%), Krakowski (8.4%), Nowy Sącz (9.7%), Olkuski (9.3%), and Wielicki (7.0%). Notable variability in average daily call frequency was seen in seven counties, with differences exceeding 10% between minimum and maximum values: Bocheński (14.7%), Chrzanowski (10.9%), Nowosądecki (12.7%), Nowotarski (10.6%), Tarnów (11.5%), Tarnowski (10.1%), and Wadowicki (10.9%).

Hourly cyclicality

The analysis of daily periodicity using hourly-aggregated data reveals a clear cyclical pattern. Typically, average hourly total calls are low during nighttime hours (2–5 a.m.), followed by a strong increase in the morning (6–9 a.m.). Local peaks occur around 9 or 10 a.m., usually the daily maximum, with another peak between 8 and 9 p.m.

Appendix 4 highlights differences in calls cyclicality across individual days within the week. In most counties, variability is pronounced, with significant declines during nighttime hours and sharp increases during the day.

Tatrzański County displays a weaker pattern, with a slower decline at night and less distinct peaks during the day. Mondays are characterized by higher average call volumes, with a slower decline from the morning peak in counties such as Bocheński, Chrzanowski, city Kraków, Olkuski, city Tarnów, Tarnowski, and Wadowicki.

On Saturdays, there is an increase in call numbers during the evening hours. In Chrzanowski, city Kraków, and Wielicki counties, call volumes rise after 9 p.m. Nowosądecki County sees elevated call numbers from 9 a.m., while Krakowski, Nowotarski, and Wadowicki counties experience increases at 8 p.m., 5 p.m., and 7 p.m., respectively. Tatrzański County shows the most significant rise after 10 p.m.

Sundays typically show higher average call numbers during nighttime and early morning hours, with counties like Bocheński, city Kraków, Krakowski, Nowosądecki, Nowotarski, city Tarnów, Tarnowski, Wadowicki, and Wielicki recording significantly more calls on average.

Two distinct groups of counties emerge from the analysis (Appendix 4). The first group experiences high call volumes primarily in the morning, including Bocheński, city Kraków, city Nowy Sącz, Olkuski, city Tarnów, Tarnowski, Wadowicki, and Wielicki. The second group, consisting of Brzeski, Chrzanowski, Krakowski, Nowosądecki, Nowotarski, Tarnowski, and Tatrzański, displays high call volumes in both the morning and early evening hours.

Pandemic phases

The analyzed period was divided into five phases due to the COVID-19 pandemic, as outlined in Table 1, along with each phase’s duration and basic statistics for unsegmented data.

Table 1.

Pandemic stages with total and average number of emergency calls.

Stage/ phase	Date	Statistics
Pre-Pandemic Period (PPP)	2020-01-01; 2020-03-03 62 days	total calls = 31,534 mean_calls_per_day = 508.3 min_calls_per_day = 435 max_calls_per_day = 664 std_dev_calls_per_day = 39.4
Initial Pandemic Onset (IPO)	2020-03-04; 2021-01-14 316 days	total calls = 138,569 mean_calls_per_day = 438.5 min_calls_per_day = 298 max_calls_per_day = 644 std_dev_calls_per_day = 59.8
Vaccination Rollout Phase (VRP)	2021-01-15; 2022-05-15 485 days	total calls = 253,695 mean_calls_per_day = 523.1 min_calls_per_day = 359 max_calls_per_day = 770 std_dev_calls_per_day = 61.1
Epidemic Threat Declaration (ETD)	2022-05-16; 2023-07-01 411 days	total calls = 226,417 mean_calls_per_day = 550.9 min_calls_per_day = 433 max_calls_per_day = 823 std_dev_calls_per_day = 51.8
Post-Pandemic Recovery (PPR)	2023-07-02; 2023-12-31 182 days	total calls = 102,772 mean_calls_per_day = 564.7 min_calls_per_day = 435 max_calls_per_day = 718 std_dev_calls_per_day = 45.5

The first phase, the Pre-Pandemic Period (PPP), represents the global onset of the pandemic before cases appeared in the country, with a mean value (500.7) and notably low standard deviation of 39.4. The second phase, the Initial Pandemic Onset (IPO), corresponds to the first occurrence of COVID-19 within the country. During this period, the number of calls initially dropped, as shown by the mean value (438.5). This phase is marked by a high standard deviation, indicating significant variability in call volume. The third phase, the Vaccination Rollout Phase (VRP), saw a slight increase in the mean number of calls (527.5) compared to the PPP. The standard deviation indicates similar data dispersion as in the IPO. The fourth phase, Epidemic Threat Declaration (ETD), saw a rise in the mean call volume to 550.9. Notably, the standard deviation dropped substantially to 51.8, as shown in Table 1. The final phase, the Post-Pandemic Recovery (PPR), is characterized by a slight increase in the mean call volume to 564.7 and standard deviation to 45.5 (Table 1).

EMS reaction time

A useful metric to illustrate the level of strain on medical services is the EMS Reaction Time, defined as the time elapsed from when an emergency call is made to the moment the call is assigned to an available EMS unit that can immediately respond. During the COVID-19 pandemic, EMS reaction time reflects two key aspects: the demand volume and the capacity of hospitals to receive patients transported to Emergency Departments.

The following subsections present a detailed analysis of EMS Reaction Times across various timeframes and conditions, examining yearly, monthly, weekly, and hourly patterns. Each subsection explores correlations between call volumes and reaction times to identify trends in EMS performance, both during and beyond the pandemic.

Yearly data

Spearman’s rank correlation was computed to assess the relationship between yearly average total calls and yearly average reaction time. There was a negative correlation between two variables, r(2) = −0.4, p = 0.75. Figure 2 shows a substantial increase in average reaction time in 2021. Despite an increase in yearly total calls, the average reaction time decreased in 2022 and 2023, reaching a minimum of 3.34 in 2023. The minimum reaction time for the years 2020, 2021, and 2022 is the same at 0.02 min. For 2023, it is 0.1 min. The maximum reaction time is significantly higher, peaking in 2021. The values for the years 2020–2023 are as follows: 666.5 min, 1451.7 min, 690.8 min, 569.8 min.

Fig. 2.

Yearly average total ambulance calls and average EMS reaction times, showing a peak in reaction time in 2021 and a gradual decrease in subsequent years, reaching the lowest average in 2023.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean number of calls across different years. The results indicated a significant difference between years F(3, 1457) = 280.9, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which year exhibited significant differences from one another. Tukey's HSD test revealed that significant years differences were observed between all years excluding difference between 2022 and 2023 where mean difference was 3.71 and p.adj = 0.825.

A one-way Analysis of Variance (ANOVA) was also performed to determine if there is a statistically significant difference in the mean reaction time across different years. The results indicated a significant difference between years F(3, 1457) = 16.48, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which year exhibited significant differences from one another. Tukey's HSD test revealed that significant years differences were observed between all years excluding difference between 2020 and 2022 where mean difference was 0.35 and p.adj = 0.898 and between 2020 and 2023 where mean difference was −1.03 and p.adj = 0.172.

Monthly data

A statistically significant correlation exists between total monthly calls and average EMS reaction time, showing positive relationship. Spearman’s rank correlation is r(10) = 0.294, p = 0.354. The average reaction time is significantly lower during warmer months, even with higher call volumes in July and August (Fig. 3). The average reaction time for these months does not exceed 3.5 min. December has the highest monthly total calls and the longest reaction time, reaching 8.15 min. Reaction times increase during colder months, ranging from 4.7 min in January to 5.6 min in April, with an interesting local dip in monthly total calls in February, associated with a lower average reaction time of 3.7 min.

Fig. 3.

Monthly distribution of total average ambulance calls and average EMS reaction times, highlighting lower reaction times in warmer months and peak call volumes with the longest reaction times in December.

The minimum reaction time for each month is 0.02 min. The maximum reaction time shows considerable variability. The lowest maximum was observed in July (222.8 min). In the colder months—January, October, November, and December the maximum reaction times exceed 630 min. In the warmer months, the maximum reaction time ranges between 330 min (June) and 455 min (August). February is unusual, with a maximum reaction time deviating from other winter months at 362.5 min, and March has the highest maximum reaction time, reaching 1451.7 min.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean number of calls across different months. The results indicated a significant difference between months F(11, 1449) = 11.34, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which month exhibited significant differences from one another. Tukey's HSD test revealed that significant months differences were observed between pairs of months: January and December, February and December, March and December, April and July, April and November, April and December, May and June, May and July, May and August, May and October, May and November, May and December, June and December, July and December, August and December, September and December, October and December, November and December.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean reaction time across different months. The results indicated a significant difference between months F(11, 1449) = 5.46, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which month exhibited significant differences from one another. Tukey's HSD test revealed that significant months differences were observed between pairs of months: May and December, June and December, July and December, August and November, August and December, September and December.

Weekly data

There is a statistically significant correlation between weekday total calls and the average EMS reaction time, with a nonlinear relationship. The Spearman rank correlation is r(5) = 0.43, p = 0.354. A visualization of ambulance total calls across the days of the week reveals a peak on Mondays and higher total calls over the weekend (Fig. 4). The highest average reaction time occurs on Mondays (5.5 min), possibly due to increased traffic congestion. The lowest average reaction times occur on the weekend, with the minimum on Saturday (4.17 min). The relationship between average reaction time and day of the week is also noticeable on weekdays, with a local minimum on Wednesday (4.5 min), followed by an increase towards Friday (4.9 min). The minimum average reaction time for each day of the week is 0.02 min. In the case of maximum reaction times, there is substantial variability. The highest maximum reaction time was observed on Thursday (1451.7 min). The lowest maximum reaction time occurred on Sunday, at 524.8 min. For the remaining days of the week, the maximum values ranged between 632 min (Wednesday) and 829 min (Tuesday).

Fig. 4.

Weekday distribution of daily total average ambulance calls and average EMS reaction times, showing peak call volumes on Mondays and weekends, with the highest average reaction time observed on Fridays.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean number of calls across different weekdays. The results indicated a significant difference between weekdays F(6, 1454) = 7.32, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which weekday exhibited significant differences from one another. Tukey's HSD test revealed that significant weekdays differences were observed between pairs of weekdays: Monday and Sunday, Monday and Thursday, Monday and Tuesday, Monday and Wednesday, Tuesday and Saturday, Wednesday and Saturday.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean reaction time across different weekdays. The results indicated a non-significant difference between weekdays F(6, 1454) = 1.13, p = 0.343.

Hourly data without county division

There are statistically significant correlations between the time of ambulance calls, hourly total calls, and the average ambulance reaction time. The correlation between the time of the ambulance call and hourly total calls is r(22) = 0.48, |p = 0.019. The correlation between the call time and average reaction time is r(22) = 0.66, p < 0.001. Moreover, the correlation between hourly total calls and average reaction time is r(22) = 0.68, p < 0.001. These relationships are shown in the line plot (Fig. 5), which reveals a clear daily periodicity and significant similarity between the variables. The highest average reaction time occurs at 18:00, reaching 7.1 min, while the lowest average reaction time is at 07:00, at 2.6 min. A distinct division appears in the day, with average reaction times between 3 and 4 min from 00:00 to 09:00 and again from 22:00 to 23:00. During the day, variability is greater, ranging from 4.4 min (at 22:00) to 7.1 min. Minimum reaction times range from 0.02 to 0.15 min, with the longest minimum times recorded between 00:00 and 02:00. Maximum reaction times display a daily cycle similar to average times. At night, maximum reaction times are lower, between 169 min (at 04:00) and 314 min (at 00:00). At 05:00 and 06:00, maximum reaction times increase to 756 min and 691 min, respectively, then decrease again at 07:00 and 08:00 to 213 min and 299 min. Between 09:00 and 20:00, maximum reaction times range from 480 min (at 17:00) to 829 min (at 16:00). The highest maximum reaction time is 1451.7 min (at 22:00).

Fig. 5.

Plot illustrating the correlation between the hour of ambulance calls, the hourly average number of total calls, and the average EMS reaction time, showing clear daily periodicity and significant similarity among the variables.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean number of calls across different hours. The results indicated a significant difference between hours F(23, 35,035) = 2498, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which hour exhibited significant differences from one another. Tukey's HSD test revealed that significant hours differences were observed between pairs of hours excluding: 6 and 0, 3 and 2, 4 and 2, 5 and 2, 4 and 3, 5 and 3, 10 and 9, 11 and 9, 12 and 9, 18 and 9, 19 and 9, 11 and 10, 12 and 11, 19 and 11, 13 and 12, 18 and 12, 19 and 12, 15:19 and 13, 17 and 14, 20 and 14, 17 and 15, 20 and 15, 17 and 16, 20 and 17, 19 and 18.

EMS reaction times and pandemic phases

There is a statistically significant, nonlinear correlation between total calls during pandemic phases and average EMS reaction time. Spearman’s rank correlation is r(3) = −0.2, p = 0.78. Figure 6 shows a relationship between reaction time and total average call volume. Interestingly, ambulance call volumes are higher during the vaccination and threat phases than in the initial phase of the pandemic. However, the average reaction time is lower in the threat phase (3.8) compared to the initial pandemic phase (4.69), despite a significantly higher number of calls. The highest average reaction time occurs during the vaccination rollout phase, at 6.5 min. The minimum reaction time for all phases except the post-pandemic recovery phase is 0.02 min; for the PPR phase, the minimum time is 0.1 min. There is considerable variability in maximum reaction times, ranging from 254 min for the pre-pandemic period. The highest maximum reaction time is 1451.7 min during the vaccination rollout phase. For the remaining phases, maximum reaction times range between 570 and 691 min.

Fig. 6.

Relationship between total average ambulance calls and average EMS reaction times across pandemic phases, illustrating increased call volumes and varying reaction times, with lower reaction times in the threat phase despite higher call numbers.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean number of calls across different pandemic phases. The results indicated a significant difference between pandemic phase F(4, 1456) = 264.6, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which pandemic phase exhibited significant differences from one another. Tukey's HSD test revealed that significant pandemic phase differences were observed between all phases excluding PPR and ETD where mean difference was 11.79 and p.adj = 0.12.

A one-way Analysis of Variance (ANOVA) was performed to determine if there is a statistically significant difference in the mean reaction time across different pandemic phases. The results indicated a significant difference between pandemic phase F(4, 1456) = 15.22, p < 0.001. A post-hoc Tukey's Honest Significant Difference (HSD) test was conducted to identify which pandemic phase exhibited significant differences from one another. Tukey's HSD test revealed that significant pandemic phase differences were observed between VRP and ETD where mean difference was 2.83 and p.adj < 0.001, VRP and IPO where mean difference was 2.01 and p.adj < 0.001, VRP and PPP where mean difference was 3.86 and p.adj < 0.001, VRP and PPR where mean difference was 3.48 and p.adj < 0.001.

Pandemic phases by hour

The analysis examined relationships between the time of day, average hourly total calls by pandemic phase, and average hourly reaction time for each phase (Fig. 7). The Spearman correlation between average hourly total calls and hours is r(118) = 0.52, p < 0.001, while the correlation between hours and average reaction time is r(118) = 0.36, p < 0.001. The relationship exists between average hourly total calls and average reaction times—r(118) = 0.35, p < 0.001.

Fig. 7.

Hourly average EMS reaction times by pandemic phase, showing strong daily periodicity, with reaction times peaking at 6 p.m. across all phases and decreasing rapidly thereafter.

The lowest average reaction times occurred during the first pandemic phase, with a high similarity between average reaction times and average hourly total calls throughout the day (see Appendix 4). The average reaction time for this phase ranges from 2.2 min at 07:00 to 3.6 min at 18:00, with most average reaction times below 3 min, except for the peak. Minimum reaction times range from 0.02 min (at 14:00 and 16:00) to 0.27 min (at 05:00). Higher minimum reaction times are observed in the early morning hours from 04:00 to 09:00. Maximum reaction times during this phase range from 10.5 min at 01:00 to 253.4 min at 15:00, with most maximum reaction times below 100 min.

Reaction times increase in the second and third phases, peaking in phase 3 (VRP). In the second pandemic phase, average reaction times range from 3.1 min at 07:00 to 6.48 min at 18:00. Minimum reaction times of 0.02 min occur at 07:00, 11:00, 14:00–18:00, and 22:00, with the highest minimum values observed at night, reaching 0.17 min at 00:00 and 01:00. Maximum reaction times are higher than in the previous phase, ranging from 72.9 min at 01:00 to 667 min at 20:00, with the highest values occurring between 09:00 and 20:00.

In the third pandemic phase, average reaction times range from 2.9 min at 07:00 to 10 min at 18:00. Minimum reaction times for this phase vary from 0.02 min at 04:00, 08:00–11:00, 14:00, 17:00–18:00, and 23:00 to 0.15 min at 00:00. Nighttime minimum reaction times (00:00–03:00) are generally higher. Maximum reaction times range from 72.1 min at 02:00 to 1451.7 min at 22:00; however, apart from peak times at 05:00 (756.5 min), 16:00 (828.7 min), 20:00 (632.4 min), and 22:00 (1451.7 min), they remain below 500 min. Notably, while a peak in total calls occurs around 10 a.m., it does not correspond with a higher average reaction time, though a local peak in calls around 20:00 does.

In phase 4 (ETD), the average reaction time decreases despite high hourly total calls. In this phase, average reaction times range from 2.3 min at 07:00 to 6.18 min at 18:00. Minimum reaction times range from 0.02 min at 09:00, 12:00, and 14:00 to 0.28 min at 03:00, with higher minimum reaction times observed during the night and early morning (00:00–07:00). Maximum reaction times are lower than in the previous phase, ranging from 78 min at 05:00 to 690.8 min at 06:00, with all but the highest value below 420 min.

In the final phase, total calls approach pre-pandemic levels, exhibiting pronounced daily cycles. Reaction times during this phase remain higher than in phase 1 and show clearer daily periodicity. The average reaction time ranges from 2 min at 07:00 to 4.17 min at 18:00. Minimum reaction times range from 0.1 min at 14:00 to 0.32 min at 04:00, with higher minimum times during the night and early morning (00:00-06:00). Maximum reaction times range from 75.1 min at 05:00 to 569.8 min at 06:00. Lower values, not exceeding 105 min, occur during nighttime (00:00-05:00). Apart from the highest maximum reaction time and a high value of 455.2 min at 23:00, maximum reaction times remain below 270 min.

Discussion

The analysis of EMS reaction times and call volumes during the COVID-19 pandemic reveals critical insights into how healthcare systems adapted to unprecedented demands and operational challenges. The data underscore that EMS performance was profoundly affected by different pandemic phases, marked by sharp increases in call volumes and notable fluctuations in reaction times. Statistically significant correlations observed between call volumes and reaction times at different times of day, days of the week, and months reflect the strain on EMS resources and suggest that reaction times were sensitive to demand surges shaped by specific temporal patterns.

Comparable observations have been reported in other countries, supporting the global relevance of our results. For instance, the study by Prezant et al.³¹ reported significant increases in daily EMS call volumes in New York City during the peak of COVID-19, coinciding with a rise in respiratory and cardiovascular incidents echoing our findings on longer reaction times associated with higher call volumes. Research from multiple countries highlights significant EMS delays during the COVID-19 pandemic, often driven by patient hesitation and system strain. In Switzerland, STEMI patients delayed EMS calls nearly threefold due to fear of infection, worsening outcomes³². In Daegu, Korea, AIS patients experienced longer EMS response times, correlating with worse NIHSS scores, as EMS teams faced added safety protocols³³. Even minimally affected areas, like Okayama, Japan, reported increased prehospital times in 2020, pointing to the universal impact of operational strain³⁴. In Victoria, Australia, reduced call volumes contrasted with longer response and transfer times, reflecting resource limitations³⁵. Similarly, in the Netherlands, fear and access issues led especially older adults to delay EMS calls, worsening outcomes for cardiac patients³⁶.

Moreover, the analysis of emergency medical services (EMS) call volumes during the COVID-19 pandemic has revealed significant temporal patterns that encompass both seasonal and weekly variations. In the context of the analyzed region (Małopolska, Poland) it was observed that EMS call volumes exhibited strong temporal regularities between 2020 and 2023, with a consistent year-over-year increase. These patterns were also reflected in fluctuations in EMS response times. Furthermore, monthly peaks were particularly visible in December, while additional increases were noted in July and August in tourist-heavy counties such as Tatrzański, Nowotarski, and Nowosądecki. Although these summer months showed local spikes in demand, overall call volumes tended to decline during warmer periods, indicating regional variation in seasonal patterns. Similar seasonal effects have been identified in studies from other countries, including Spain, where Rudilosso et al. (2020) noted that EMS patterns are closely influenced by climate and holiday periods. Comparable seasonal peaks were also reported by Ferron et al.³⁷ and O’Connor et al.³⁸, reinforcing the broader applicability of these observations.

Moreover, the analysis of weekly rhythms in EMS call volumes highlighted a pronounced cyclical pattern. Mondays, Fridays, and Saturdays tended to experience higher call volumes, whereas Tuesdays consistently recorded the lowest. These fluctuations differed across counties, suggesting that local social dynamics and service access influenced EMS demand. These findings mirror those of Prezant et al.³¹ and O’Connor et al.³⁸, who noted similar weekday-based differences in EMS activity.

Hourly distribution patterns further elucidated the cyclical nature of EMS demand. Call volumes were lowest during the early morning hours, with notable peaks around 10:00 and 18:00. It was also noted that minimum EMS reaction times occurred during the night and early morning (00:00-07:00/09:00), while maximum values were typically observed in the late afternoon, particularly at 18:00. These findings are consistent with those reported by Ferron et al.³⁷ and Xie et al.³⁹, who found comparable intra-day peaks in EMS demand, and by Slavova et al.⁴⁰, who linked response time efficiencies to specific time intervals. Furthermore, O’Connor et al.³⁸ raised concerns about potential delays in patient outcomes during nocturnal phases. Therefore, this cyclicity suggests opportunities for optimal resource allocation during known high-demand periods, contrasting with findings from other regions that may not have distinctly examined such variability over seasonal cycles¹⁶.

What is more, during the pandemic, substantial differences in maximum EMS reaction times were evident. The lowest maximum times were observed in the Pre-Pandemic Period (PPP), while the highest were recorded during the Vaccination Rollout Phase (VRP), slightly exceeding those in the Initial Pandemic Onset (IPO) phase. No significant differences were noted in minimum times. The temporal pattern for maximum reaction times mirrored that of average times in terms of distribution.

A particularly notable observation was the marked increase in EMS reaction times during 2021, which aligned with the VRP, a phase characterized by the highest call volumes and concurrent increases in operational delays. This finding is consistent with the work of Baldi et al.³², who reported prolonged EMS reaction times during COVID-19 peaks in Italy, particularly in out-of-hospital cardiac arrest cases, as well as with Ahn et al.⁴¹, who documented similar delays during surging case numbers in South Korea. Despite the call volume spike in July and August, reaction times were observed to decrease during these warmer months, only to rise again as colder weather approached. Interestingly, the Epidemic Threat Declaration (ETD) phase showed relatively improved reaction times, even under elevated call pressure. This observation is consistent with the conclusions drawn in the study of Chanrak & Kraisawat⁴², who highlighted how repeated exposure to pandemic conditions led to strategies that optimized operational readiness, thus reducing initial delays observed during the pandemic's onset. This adaptive capacity indicates that EMS systems began to learn from their experiences, refining their approaches to handle peak demands more effectively.

The increased periodicity in EMS reaction times, with predictable peaks during evening hours, highlights a potentially valuable area for strategic resource planning. This pattern suggests that EMS systems could allocate additional resources during peak times to mitigate the impact of heightened demand. Additionally, as demonstrated by the improved response times in later pandemic phases, it appears that EMS systems learned to adapt, integrating lessons from initial waves to optimize performance.

The COVID-19 pandemic has underscored the critical need for emergency medical systems to have both the flexibility and resilience to respond to fluctuating demand and unforeseen challenges⁴³. Future pandemic preparedness strategies should consider both the logistical adjustments and public education initiatives needed to reduce delays, ensuring that patients seek timely emergency care regardless of pandemic fears or misconceptions.

To summarize, although the study offers valuable evidence on temporal patterns in EMS operations, several limitations should be acknowledged. We did not examine qualitative factors behind delays, such as specific clinical or procedural factors causing prolonged EMS reactions. Furthermore, data were geographically limited to southern Poland, which might restrict generalizability to other regions. Future research should include qualitative analyses and multi-region comparisons for broader applicability.

Conclusions

The COVID-19 pandemic has had a profound impact on healthcare systems globally, placing significant strain on Emergency Medical Services (EMS) and exposing vulnerabilities in emergency response infrastructure. A retrospective analysis of EMS reaction times and call volumes in the Małopolskie Voivodeship (Poland) has provided important insights into how EMS systems responded to unprecedented operational challenges. The results indicate a significant increase in EMS reaction times, particularly during periods of heightened call volumes. The Vaccination Rollout Phase (VRP) was associated with the greatest operational strain and the highest reaction times.

Initial phases of the pandemic were characterized by system-wide difficulty in meeting the surge in demand. However, a progressive improvement in reaction times was observed in subsequent phases, despite the persistence of high call volumes. This trend likely reflects adaptive strategies such as optimized resource distribution, revised operational protocols, and improved forecasting of emergency demand. Seasonal trends further emphasized the importance of flexible operational planning. Shorter reaction times recorded during the summer months, despite elevated call volumes, suggest effective adjustments in staffing or deployment. In contrast, increased delays in colder months may be attributed to environmental constraints or seasonal illness peaks.

The analysis also identified clear patterns in daily and hourly EMS call volumes, with predictable peaks offering opportunities for more targeted resource allocation. The findings underscore the role of public perception and health-seeking behavior, particularly during a pandemic. Fear of infection and changes in service utilization patterns likely contributed to additional delays and operational complexity.

To enhance future EMS preparedness, the implementation of dynamic ambulance allocation strategies, integration of data-driven demand forecasting, and improvement of public risk communication are recommended. Although limited in geographical scope and lacking qualitative data, the study provides key benchmarks for evaluating EMS resilience and supports evidence-based decision-making for policymakers and healthcare system planners. Future research should explore the integration of EMS operational data with patient outcome metrics, as well as the influence of environmental and staffing variables on system responsiveness. Moreover, comparative studies across different regions both urban and rural would provide deeper insights into how structural and demographic factors affect emergency medical logistics during crises.

In conclusion, this study contributes novel, data-driven evidence on EMS system performance under pandemic stress. It offers important lessons for emergency preparedness and supports the case for dynamic, temporally informed resource allocation models that can adapt to both predictable patterns (e.g., seasonal peaks) and unpredictable surges (e.g., pandemic waves). These findings are relevant not only for managing future health emergencies but also for informing structural reforms to improve EMS efficiency in standard operating conditions.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

M.L. conceived the study design and experimental framework. M.L. and M.C. conducted the data collection and performed the analyses. M.C. visualized the results and contributed to data interpretation. M.L. and M.C. jointly drafted the manuscript and reviewed it for accuracy and clarity. All authors reviewed and approved the final version of the manuscript.

Data availability

The datasets generated and/or analyzed during the current study are included in this published article and its Supplementary Information files. Additional data that support the findings of this study are available from the corresponding author, Michał Lupa, upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.