Impact of early cardiac arrest recognition by emergency medical service dispatchers on out-of-hospital cardiac arrest: a retrospective population-based observational study

Abstract

Background

Dispatch-assisted cardiopulmonary resuscitation (DACPR) improves out-of-hospital cardiac arrest (OHCA) outcomes. However, there is no established gold standard for DACPR performance. This study aimed to investigate the association between time to OHCA recognition by the dispatcher and return of spontaneous circulation (ROSC), survival, and favorable neurological outcome.

Methods

We conducted this retrospective population-based study using Utstein-style OHCA data and DACPR records between January 2016 and December 2019 in Nara, Japan. OHCA cases were included if lay rescuers performed CPR following dispatcher instructions. Patients aged < 18 years, emergency medical service-witnessed OHCAs, patients for whom CPR was already initiated at the emergency call, cases where CPR was not indicated, and cases in which callers did not follow dispatcher instructions were excluded. The primary outcome was the association between time to OHCA recognition by the dispatcher and ROSC. The secondary outcomes included 1-month survival and favorable neurological outcome (cerebral performance category 1 or 2). Multivariable logistic regression was performed to examine the association between dispatcher recognition time and outcomes. Dispatcher recognition time was modeled per 10-s increase. Stratified analyses were also performed for witnessed and non-witnessed OHCA cases.

Results

We analyzed 1,959 cases of OHCA. The median OHCA recognition time was 74 s. Longer dispatcher recognition time was associated with slightly higher odds of ROSC in the overall cohort (adjusted odds ratio [aOR] per 10-s increase, 1.01; 95% confidence interval [CI], 1.01–1.02) and lower odds of ROSC in witnessed OHCA (aOR, 0.98; 95% CI, 0.96–0.99), whereas no clear association was observed in non-witnessed OHCA (aOR, 1.02; 95% CI, 0.99–1.04). Delay in OHCA recognition by dispatchers was not significantly associated with survival or neurological outcomes.

Conclusion

The association between dispatcher recognition time and ROSC differed according to arrest context. In witnessed OHCA, longer dispatcher recognition time was associated with lower odds of ROSC. These findings suggest that minimizing dispatcher recognition delay may be particularly important in witnessed OHCA cases.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12245-026-01213-1.

Background

Bystander cardiopulmonary resuscitation (CPR) is strongly associated with better outcomes in cases of out-of-hospital cardiac arrest (OHCA) [1–5]. It can almost quadruple the chance of 1-month survival among witnessed OHCA cases [3]. However, some reports indicate that the rate of bystander CPR remains low worldwide, at approximately 40%, and thus needs to be improved [6, 7]. Emergency medical service (EMS) dispatchers substantially contribute to EMS pre-arrival resuscitation. EMS dispatchers support emergency callers in performing CPR via phone instructions. Dispatch-assisted CPR (DACPR) or telecommunicator CPR (TCPR) accounts for the majority of bystander CPR [8–11]. Previous studies have shown that DACPR or TCPR is equivalent to or even better than layperson-initiated CPR in terms of affording better outcomes in cases of OHCA [11–13]. To further improve the outcomes of OHCA in the community, it is critical to evaluate the quality of DACPR. However, there is no gold standard for reviewing DACPR performance. The proposed DACPR indicators are the proportion of OHCA cases recognized by EMS dispatchers, the dispatch of CPR instructions, and the DACPR provision [14]. Time measurements for the DACPR process, in seconds, are other key indicators: the time interval between EMS calls and OHCA recognition by dispatchers, the start of the DACPR instruction, and the first DACPR chest compression. However, the association of these time intervals with return of spontaneous circulation (ROSC), survival, and favorable neurological outcomes has not been well investigated. Therefore, in this study, we retrospectively assessed the impact of the time required for OHCA recognition by dispatchers on ROSC, survival, and neurological outcomes in patients with sudden OHCA in Nara, Japan.

Methods

Study design

This retrospective population-based observational study was conducted using Utstein-style EMS-confirmed OHCA data and DACPR records of lay rescuers who performed CPR in Nara Prefecture between January 2016 and December 2019. We used data from the pre-COVID-19 pandemic period (January 2016 to December 2019) to ensure uniformity of the DACPR protocols. Data from 2020 onwards were not included because the COVID-19 pandemic led to changes in bystander behavior and dispatcher instructions, including the introduction of mouth-and-nose covering guidance in our region [15].

The exclusion criteria were as follows: (1) patients younger than 18 years; (2) patients for whom lay rescuers had already performed CPR at the time of the EMS call; (3) EMS-witnessed OHCA cases; (4) cases in which dispatcher-assisted CPR was not indicated, including patients who were responsive at the time of the EMS call and callers who were too upset to continue the call, discontinued the call, or left the victim’s side; and (5) cases in which callers did not follow dispatch instructions.

EMS in Nara prefecture

Nara Prefecture has an area of 3,690 km² and a population of 1.3 million. Three fire departments (Nara City Fire Bureau, Ikoma City Fire Department, and Nara Wide Area Fire Department) with two dispatch centers service the entire region. The number of ambulance dispatches and OHCA cases in the region is approximately 60,000 and 1,200 per year, respectively. Both EMS and emergency dispatchers follow standardized protocols established by the Nara Medical Control Council for OHCA. The Nara City Fire Bureau and Nara Wide Area Fire Department have physician-staffed ambulance systems that dispatch emergency physicians to emergency sites based on specific protocols.

Dispatch-assisted CPR protocol in Nara

The standardized DACPR protocol was established based on a “two-question protocol” [16] that EMS dispatchers use to identify OHCA and provide CPR instructions when a caller describes an individual as being (1) unconscious and (2) not breathing normally. In our region, dispatchers provided standardized hands-only CPR instructions once OHCA was recognized.

DACPR reports in Nara

Dispatchers were required to record DACPR reports when they received calls for EMS to identify OHCA at the scene. They recorded basic information (patient age, sex, witness status), OHCA identification (yes/no), provision of CPR instruction (yes/no), and barriers to OHCA identification and CPR instruction. The dispatcher who handled the call was required to report three key time intervals of the DACPR (in seconds) after reviewing the recorded telephone conversation as part of routine DACPR reporting: the time interval between the call and OHCA identification (t1), the start of the DACPR instruction (t2), and the start of the first chest compression (t3).

Primary and secondary outcomes

The primary outcome of this study was the association between the time to OHCA recognition by the dispatcher and ROSC. We focused primarily on t1 because it most directly reflects dispatcher recognition performance, whereas t2 and t3 are influenced by additional caller- and scene-related factors. The secondary outcomes were the associations between the time intervals and both 1-month survival and favorable neurological outcomes. Cerebral Performance Category (CPC) 1 or 2 at 1 month was defined as a favorable neurological outcome. Outcome data, including ROSC, 1-month survival, and neurological outcome, were obtained from the Utstein-style OHCA registry. ROSC was defined according to the Utstein-style template as the restoration of a palpable pulse, usually at a major artery such as the carotid artery, without a prespecified minimum duration [17].

Statistical analysis

Descriptive data are presented as numbers and proportions (%) or medians (interquartile range [IQR]). Continuous variables were compared using the Mann–Whitney U test, and categorical variables were compared using the chi-square test. Multivariable logistic regression analysis was performed to examine the association between dispatcher recognition time for OHCA and ROSC. The model was adjusted for biologically relevant covariates such as age, sex, and witness status. Initial shockable rhythm was not integrated into the primary models because it was assessed after the emergency call and could represent a post-exposure variable. We performed a sensitivity analysis with additional adjustment of the multivariable models for initial shockable rhythm. A variable that divided OHCA recognition time intervals by 10-s increments was included to assess the impact of recognition delays on primary and secondary outcomes. Odds ratios (ORs) with 95% confidence intervals (CIs) were estimated using this model.

Agonal respiration is often noted in witnessed OHCA cases, and the time intervals for OHCA recognition by the dispatcher are longer than those in non-witnessed cases, because the presence of agonal respiration particularly causes an increase in time required for collecting breathing information [18, 19]. Therefore, we performed separate stratified analyses for witnessed and non-witnessed OHCA cases.

Secondary outcomes were 1-month survival and favorable neurological outcome. To improve clinical interpretability, dispatcher recognition time was dichotomized at 60 s in additional exploratory analyses. We also conducted a supplementary analysis to assess potential selection bias, descriptively comparing the analyzed cohort with the major excluded adult OHCA groups. Restricted cubic spline regression was performed in the overall cohort to examine the shape of the association between dispatcher recognition time and ROSC, using the same covariates as in the primary multivariable model. Non-linearity was assessed using Wald tests for the non-linear spline components. In addition, a supplementary spline analysis was performed for witnessed OHCA cases using the same covariates as in the corresponding subgroup model. Exploratory causal mediation analysis was also performed to evaluate whether prehospital adrenaline administration mediated the association between dispatcher recognition delay and ROSC. Prehospital adrenaline administration was considered a potential mediator rather than a confounder because it may occur after dispatcher recognition and may be influenced by recognition timing. Mediation analysis was conducted using the “mediation” package in R with logistic regression models, estimating the average causal mediation effect (ACME), average direct effect (ADE), total effect, and the proportion mediated using 1,000 bootstrap simulations to obtain 95% CIs. All data were analyzed using R version 4.3.1. Statistical significance was set at P < 0.05.

Ethical review

The study was approved by the ethics committee of Nara Medical University Hospital (approval number: 3777). The requirement for informed consent was waived for all enrolled patients because the study used anonymized public-domain data.

Results

During the study period, 4,991 OHCA cases were recorded. Among these, 44 cases involved individuals younger than 18 years old. The subjects were OHCA witnessed by EMS in 158 cases, and CPR had already started in 366 cases. Regarding the remaining 4,423 OHCA cases, CPR was not indicated in 1,331 cases (26.7%), callers did not follow dispatch instructions in 1,133 cases (22.7%), and 1,959 cases were further analyzed (Fig. 1).

Fig. 1.

Flow diagram of patient inclusion and exclusion in the study

Table 1 shows patient characteristics. The median patient age was 80 (IQR: 70–87) years, and 1,082 patients were male (55.2%). Cardiogenic OHCA accounted for 1,554 cases (79.3%), and an initial shockable rhythm was recorded in 180 cases (9.2%). ROSC was achieved in 291 cases. ROSC was achieved in the prehospital phase in 259 cases (13.2%). Regarding the time intervals pertaining to the DACPR process, the median time intervals for OHCA recognition (t1), start of dispatch instruction (t2), and first DACPR chest compression (t3) were 74 (IQR: 48–130) s, 155 (IQR: 98.5–240) s, and 212 (IQR: 141–300) s, respectively. A descriptive comparison of the analyzed cohort and major excluded adult OHCA groups is provided in Supplementary Table 1.

Table 1.

Characteristics of the study population

	Total	Witnessed	Non-witnessed
	N = 1959	n = 708	n = 1251
Age, years (median)	80 (70–87)	80 (70–87)	81 (71–87)
Male sex, n (%)	1082 (55.2)	420 (59.3)	662 (52.9)
Initial shockable rhythm, n (%)	180 (9.2)	116 (16.4)	64 (5.1)
Cardiogenic OHCA, n (%)	1,554 (79.3)	500 (70.6)	1,054 (84.3)
Physician at the scene, n (%)	278 (14.2)	161 (22.7)	117 (9.4)
Prehospital adrenaline administration by EMS, n (%)	913 (46.6)	342 (48.3)	571 (45.6)
Advanced airway by EMS, n (%)	1224 (62.5)	456 (64.4)	768 (61.4)
Prehospital ROSC, n (%)	259 (13.2)	159 (22.5)	100 (8.0)
Total cases of ROSC, n (%)	291 (14.9)	175 (24.7)	116 (9.3)
t1: Time interval between the EMS call and OHCA recognition by dispatcher, s (median, IQR)	74 (48–130)	88 (55–166.25)	69 (45–113.5)
t2: Time interval between the EMS call and dispatch instructions, s (median, IQR)	155 (98.5–240)	156 (97–257.25)	155 (99.5–235)
t3: Time interval between the EMS call and DACPR, s (median, IQR)	212 (141–300)	209.5 (126.75–312)	216 (147.5–299)

DACPR: dispatch-assisted cardiopulmonary resuscitation; EMS: emergency medical service; IQR: interquartile range; OHCA: out-of-hospital cardiac arrest; ROSC: return of spontaneous circulation

Restricted cubic spline modelling of dispatcher recognition time (adjusted for age, sex, and witness status) in the overall cohort did not suggest a non-linear relationship with ROSC (p for non-linearity = 0.19) (Fig. 2). The results of the supplementary spline analysis for witnessed OHCA cases are provided in Supplementary Fig. 1. Table 2 shows the results of the multivariable logistic regression analysis for ROSC in the overall cohort. In the adjusted model, older age was associated with lower odds of ROSC, whereas witnessed OHCA was associated with higher odds of ROSC. Dispatcher recognition time was also associated with ROSC, with an adjusted OR (aOR) of 1.01 (95% CI, 1.01–1.02) for every 10-s increase (Table 2).

Fig. 2.

Restricted cubic spline curve for the association between dispatcher recognition time and ROSC in the overall cohort

Table 2.

Crude and adjusted odds ratios for ROSC in all cases

Variable	Crude			Adjusted
	Odds ratio	95% CI	p-value	Odds ratio	95% CI	p-value
Age (per 1-year increase)	0.99	0.98–0.99	< 0.05	0.99	0.98–0.99	< 0.05
Male sex	1.09	0.85–1.40	0.50	0.96	0.74–1.25	0.76
Witnessed OHCA	3.21	2.49–4.15	< 0.001	3.05	2.35–3.96	< 0.001
Delayed OHCA recognition by dispatcher by every 10-s interval	1.02	1.01–1.03	< 0.001	1.01	1.01–1.02	< 0.05

CI: confidence interval; OHCA: out-of-hospital cardiac arrest; ROSC: return of spontaneous circulation

Dispatcher recognition time was significantly longer in witnessed than in non-witnessed OHCA cases (88 s vs. 69 s, p < 0.001; Table 1). Therefore, stratified analyses were performed for witnessed and non-witnessed OHCA cases. In witnessed OHCA, longer dispatcher recognition time was associated with lower odds of ROSC, with an aOR of 0.98 (95% CI, 0.96–0.99) for every 10-s increase (Table 3), whereas no clear association was observed in non-witnessed OHCA cases (aOR, 1.02; 95% CI, 0.99–1.04) (Table 4).

Table 3.

Crude and adjusted odds ratios for ROSC in witnessed OHCA cases

Variable	Crude			Adjusted
	Odds ratio	95% CI	p-value	Odds ratio	95% CI	p-value
Age (per 1-year increase)	0.99	0.98–0.99	< 0.05	0.98	0.97–0.99	< 0.05
Male sex	0.92	0.65–1.29	0.62	0.83	0.58–1.18	0.30
Delayed OHCA recognition by dispatcher by every 10-s interval	0.98	0.96–0.99	< 0.05	0.98	0.96–0.99	< 0.05

CI: confidence interval; OHCA: out-of-hospital cardiac arrest; ROSC: return of spontaneous circulation

Table 4.

Crude and adjusted odds ratios for ROSC in non-witnessed OHCA cases

Variable	Crude			Adjusted
	Odds ratio	95% CI	p-value	Odds ratio	95% CI	p-value
Age (per 1-year increase)	0.98	0.97–0.99	< 0.05	0.98	0.97–0.99	< 0.05
Male sex	1.04	0.71–1.52	0.85	0.96	0.65–1.41	0.81
Delayed OHCA recognition by dispatcher by every 10-s interval	1.02	0.99–1.04	0.06	1.02	0.99–1.04	0.08

CI: confidence interval; OHCA: out-of-hospital cardiac arrest; ROSC: return of spontaneous circulation

Associations were not observed between a 10-s delay and survival at 1 month or a favorable neurological outcome in the overall, witnessed, or non-witnessed OHCA cohorts (Supplementary Table 2). In sensitivity analyses with additional adjustment for initial shockable rhythm, the results were materially unchanged (Supplementary Table 3). The results of exploratory analyses using a 60-s cutoff for dispatcher recognition time are shown in Supplementary Table 4. In the exploratory causal mediation analysis, the direct effect (ADE) of dispatcher recognition delay on ROSC remained statistically significant, whereas the indirect effect via prehospital adrenaline administration (ACME) was not significant (Supplementary Table 5).

Discussion

This study aimed to investigate the association between DACPR intervals and ROSC, survival, and favorable neurological outcome. The multivariable analysis showed that a 10-s delay in OHCA recognition by the dispatcher was associated with slightly higher odds of ROSC in the overall cohort but with lower odds in witnessed OHCA cases, whereas no clear association was observed in non-witnessed OHCA cases. These findings suggest that the clinical meaning of dispatcher recognition time differs according to arrest context and that pooled analyses of mixed OHCA populations may yield paradoxical associations.

In general, witnessed OHCAs have a better prognosis than non-witnessed OHCAs because the no-flow period is shorter. However, dispatchers must process a large amount of information in witnessed cases, including confusion at the scene, upset callers, and the presence of agonal respiration, which can be confusing for lay rescuers. Therefore, the interval for OHCA recognition by the dispatcher is generally longer for witnessed cases than for non-witnessed cases [18, 19]. This trend was also confirmed in this study. Accordingly, the paradoxical positive association observed in the overall cohort likely reflects case-mix heterogeneity rather than a beneficial effect of delayed recognition itself. In this context, the finding that longer dispatcher recognition time was associated with lower odds of ROSC in witnessed OHCA appears clinically plausible. Additionally, our causal mediation analysis demonstrated that the impact of dispatcher recognition delay on ROSC was primarily direct, with no significant mediation via prehospital adrenaline administration. This exploratory finding suggests that the observed association was not clearly explained by prehospital adrenaline administration alone.

Regarding the time interval for dispatchers to recognize an OHCA, previous studies have reported that it generally takes between 70 and 120 s [20–24]. The American Heart Association has emphasized prompt OHCA recognition and initiation of chest compressions by dispatchers. It recommends OHCA recognition within 90 s and the first chest compression within 150 s of the emergency call acceptance [14]. However, this recommended time window has not yet been fully validated. Byrsell et al. first supported this time window by showing that OHCA cases with a dispatch recognition time of > 90 s had poor outcomes at 1 month [25]. Our findings support the importance of minimizing recognition delay, particularly in witnessed OHCA cases. At the same time, restricted cubic spline modelling in the overall cohort did not suggest marked non-linearity, supporting the use of dispatcher recognition time as a continuous variable in 10-s increments in the main analyses. Exploratory analyses using a 60-s cutoff were also performed, but these were intended to complement rather than replace the primary continuous-time model.

Prompt bystander CPR and immediate OHCA recognition following DACPR can shorten no-flow time, leading to better outcomes. However, this is not always the case when it comes to early OHCA recognition by dispatchers. In this study, the median time to OHCA recognition was 88 s in witnessed OHCA cases. However, in some cases, 150 s or more was required. Dispatchers frequently encounter various barriers to early OHCA recognition [18, 24, 26], which might explain why a substantial number of callers did not follow dispatcher instructions despite attempted DACPR. Rather than relying solely on individual dispatcher efforts, incorporating time from call to OHCA recognition into continuous quality improvement programs within EMS systems and routinely reviewing this metric may be essential for enhancing system performance. Although this study focused on the time from call receipt to cardiac arrest recognition, we acknowledge that the delay from cardiac arrest onset to EMS call is also a critical factor affecting the outcomes. Public education and community interventions to encourage prompt EMS activation may help reduce pre-call delay, thereby improving survival in OHCA, as highlighted in recent studies [27, 28].

Limitations

This study had several limitations. First, to investigate the impact of OHCA recognition time, we excluded cases in which CPR was not indicated or callers did not follow dispatcher instructions, as recognition time data were unavailable for these cases. Although we added a descriptive comparison of the analyzed cohort and the major excluded adult OHCA groups, selection bias may still have been present. Second, there were few survivors and patients with favorable neurological outcomes, which may have limited the statistical power to detect significant differences in secondary outcomes. Third, the time variables were documented by the dispatcher who handled the call after reviewing the recorded telephone conversation and therefore might have been affected by documentation errors or variations in recording practices. Fourth, this study was conducted in a single region in Japan. The relatively low proportion of CPR already in progress before dispatcher-assisted CPR in our cohort may reflect setting-specific differences in community response and reliance on dispatch-assisted CPR, which should be considered when comparing these findings with those from other EMS systems. In addition, the analyzed cohort represented a selected subgroup of adult OHCA cases in which dispatcher-assisted CPR was pursued and documented. Therefore, caution is required when generalizing the results. In addition, our study was based on the pre-COVID-19 pandemic period (2016–2019). The COVID-19 pandemic is reportedly associated with changes in OHCA characteristics, bystander behavior, and outcomes, and because dispatch protocols were modified during the pandemic period in our region (including the introduction of mouth-and-nose covering guidance), the comparability of DACPR processes across time periods may be limited [15, 29–31]. We focused on a pre-pandemic cohort to ensure greater uniformity in dispatch protocols and data comparability. Nevertheless, the generalizability of our findings to contemporary post-pandemic practice requires confirmation in more recent datasets. Furthermore, this study did not examine post-ROSC treatments in the different hospitals to which the patients were transported, and other unknown factors may have influenced the outcomes. In addition, detailed data on CPR quality and other arrival interventions were unavailable and not included in the primary models because these variables may represent post-exposure processes. Therefore, unmeasured intra-arrest and post-arrival factors may still have influenced the observed association between dispatcher recognition time and ROSC. Finally, we categorized dispatcher recognition delay using 10-second intervals for analysis, which may have limited statistical power given the sample size. We acknowledge this as a limitation, and future studies with larger datasets may explore alternative categorizations or advanced modelling approaches to refine these findings.

Conclusions

In this study, the association between dispatcher recognition time and ROSC differed according to arrest context. In witnessed OHCA, longer dispatcher recognition time was associated with lower odds of ROSC, suggesting that minimizing dispatcher recognition delay may be particularly important in witnessed OHCA cases.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CPR

Cardiopulmonary resuscitation

DACPR

Dispatch-assisted cardiopulmonary resuscitation

OHCA

Out-of-hospital cardiac arrest

ROSC

Return of spontaneous circulation

IQR

Interquartile range

OR

Odds ratio

CI

Confidence interval

EMS

Emergency medical service

TCPR

Telecommunicator cardiopulmonary resuscitation

CPC

Cerebral performance category

Author contributions

Koji Yamamoto: Conceptualization, Methodology, Formal analysis, Investigation, Data Curation, Writing-Original Draft, Writing-Review & Editing, Visualization, Project administration. Hirozumi Okuda: Writing-Review & Editing. Arisa Kinoshita: Writing-Review & Editing. Toru Osaki: Writing-Review & Editing. Keisuke Tsuruta: Writing-Review & Editing. Keita Miyazaki: Writing-Review & Editing. Hideki Asai: Investigation, Writing-Review & Editing. Yasuyuki Kawai: Writing-Review & Editing. Hidetada Fukushima: Conceptualization, Methodology, Formal analysis, Writing-Review & Editing.

Funding

None.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to restrictions related to the protection of personal information and ethical approval conditions but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the ethics committee of Nara Medical University Hospital (Approval Number: 3777). The requirement for informed consent was waived for all enrolled patients because the study used anonymized public-domain data. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable. This study used anonymized retrospective data, and no individual person’s data in any form (including individual details, images, or videos) is included. The requirement for informed consent was waived by the ethics committee.

Competing interests

The authors declare no competing interests.