The proportion of older adults in Hong Kong is expected to increase from 18% in 2019 to 26% by the year 2029.1 Overcrowding is a serious problem, such that population densities of 57 530 people/km² are present in ageing districts.2 3 Most residents of Hong Kong live in high-rise apartments that require elevators for access, but most elevators cannot accommodate an ambulance stretcher with a patient in a supine position.4 More than 50% of out-of-hospital cardiac arrest (OHCA) events occur in private homes, a location that is associated with poor survival outcomes.5 The proportion of domestic households consisting solely of people aged ≥65 years has increased by approximately 24% between 2011 and 2016, from 8.4% to 10.4%.6 Considering these demographic changes, there is a need for improved overall understanding of the prehospital management of cardiac arrests that involve older adults in homes and other locations. This study investigated patient characteristics, types of bystanders involved, and prehospital interventions that were associated with differences in survival outcomes among cardiac arrests involving older adults in homes, compared with cardiac arrests on streets and in public areas excluding streets (PAES).

This secondary analysis focused on a historical cohort from a previous study.5 The Emergency Ambulance Service of the Fire Services Department (FSD) provides most emergency medical services (EMS) in Hong Kong through a one-tiered system that serves the entire 1104 km² region. At the time of data collection, the population was around 7.1 million.7 Ambulance personnel are required to perform cardiopulmonary resuscitation (CPR) on and transfer most cases of OHCA to hospitals. A small number of patients with obvious post-mortem changes (eg, rigor mortis) may be directly transferred to the public mortuary; such patients were not included in this study. Fire Services Department ambulances will only transfer patients to emergency departments under the Hospital Authority. At the time of data collection, callers requesting for EMS for OHCA patients were not provided with post-dispatch instructions to perform CPR.

This secondary analysis included all patients with OHCA who were transferred to the Emergency Departments (EDs) by FSD ground ambulances from 1 August 2012 to 31 July 2013. Exclusion criteria were cardiac arrests caused by trauma, patients not transferred by ground ambulance, and patients directly transferred to the public mortuary. After patient selection from the primary dataset, the following additional exclusions were made: cardiac arrests that involved patients aged <65 years, occurred within residential care homes for the elderly, or occurred in the ambulance en route to hospital.

Data regarding patient characteristics and prehospital management were prospectively collected by EMS personnel who were directly involved in prehospital care for patients who experienced OHCA. The collected data included patient age and sex, location of cardiac arrest, whether the cardiac arrest was witnessed and the identity of the witness, whether bystander CPR was performed and who performed it, whether defibrillation with an automated external defibrillator (AED) was performed, what electrocardiogram rhythm was first detected, the timings of prehospital events (recognition of cardiac arrest, receipt of EMS call, initiation of bystander CPR, initiation of first defibrillation, EMS arrival at patient’s side, initiation of CPR by EMS personnel, and arrival at the ED), and return of spontaneous circulation (ROSC) before ED arrival.

Electronic medical records at the relevant ED (Accident and Emergency Information System, Hong Kong Hospital Authority) were reviewed to determine the time of defibrillation and time of ROSC at the ED, as well as whether a patient survived until admission. A patient was assumed to have received no resuscitative intervention unless specific documentation was present in the ED record. Neurological status upon discharge and survival at 30 days after cardiac arrest were determined from a territory-wide electronic medical record database (Clinical Management System, Hong Kong Hospital Authority).

Streets were defined as paved thoroughfares for pedestrians, including sidewalks. Public areas excluding streets were other areas that were accessible by the public throughout the day; these included outdoors (eg, parks and markets) and indoor facilities (eg, eateries, places of recreation, and day care facilities for older adults). Bystanders were defined in accordance with the guidelines of the Utstein Resuscitation Registry Templates for Out-of-Hospital Cardiac Arrest.8 Fire Services Department first responders dispatched to the scene were classified as EMS personnel. Older adult care workers (OACWs) are individuals who care for residents in various private and public housing arrangements for older adults. Older adult care workers accompanying patients were not dispatched as part of the organised emergency rescue team; thus, they were classified as bystanders. Public access defibrillation (PAD) was defined as a defibrillation shock delivered from an AED when a bystander performed CPR. Shocks delivered when FSD first responders performed CPR were excluded.

Time intervals were rounded to the nearest minute. The decision interval was the interval between recognition of cardiac arrest and receipt of EMS call. Call-to-bystander CPR was the interval between receipt of EMS call and initiation of bystander CPR. Call-to-EMS arrival was the interval between receipt of EMS call and EMS arrival at the patient’s side. Time of first defibrillation was defined as the time of the earliest of the following three events: PAD, defibrillation by EMS, or defibrillation in the ED. Call-to-bystander CPR intervals were grouped as 0-2, 3-5, 6-8, 9-11, and 12-31 minutes, as well as no bystander CPR. Call-to-first defibrillation intervals were grouped as 0-5, 6-10, 11-15, 16-20, and 21-55 minutes, as well as no defibrillation (>55 minutes/not applicable).

Post-cardiac arrest neurological status was classified using the 5-point Glasgow-Pittsburgh Cerebral Performance Categories (CPC) scale. In the scale, CPC 1 represents patients with good cerebral performance; CPC 2 includes patients who can manage activities of daily living independently or participate in part-time work in a sheltered environment; CPC 3 to CPC 5 ranges from patients who are unable to live independently because of cerebral disability to patients who have experienced brain death. Patients with CPC 1 or CPC 2 were presumed to have a favourable neurological outcome.

Patient characteristics, interventions, and outcomes were analysed using descriptive statistics. Pearson’s χ² test was used to compare categorical variables; Fisher’s exact test was used if >20% of expected counts were <5. The Kruskal–Wallis rank sum test was used to compare non-parametric time intervals. A P value of <0.05 was considered statistically significant. Predictors of 30-day survival were analysed using univariate and multivariate logistic regression; findings were reported as odds ratios (ORs) with 95% confidence intervals. Adjusted variables included age; sex; arrest location; person witnessing the arrest (relative, OACW or other bystanders, EMS personnel, or unwitnessed); person performing bystander CPR (no bystander CPR, OACW, relative, or other); PAD (yes or no); first monitored rhythm (asystole, pulseless electrical activity, ventricular fibrillation/ventricular tachycardia); and call-to-EMS arrival, call-to-bystander CPR, and call-to-first defibrillation intervals.

Statistical analysis was performed using R software, version 3.6.1 (R Foundation for Statistical Computing, Austria). The original study was approved by the Institutional Review Board of The University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref No.: UW 15-599). No new data were collected for secondary analysis. This manuscript was prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) reporting guidelines.

Figure 1 describes patient selection from the primary dataset. The original cohort comprised 5154 patients with OHCA who were transferred to the ED by FSD ground ambulances. After the application of exclusion criteria, 2255 patients were included in the analysis. Table 1 compares the patient and bystander characteristics, interventions, and outcomes of OHCA occurring in homes, in PAES, and on streets. Patients who experienced cardiac arrest in homes were significantly older (approximately 5 years; P<0.001) than patients who experienced cardiac arrest on streets or in PAES. In all groups, there were more male patients; the sex disparity was the greatest in the streets group, followed by the PAES group.

Furthermore, most cardiac arrests (66.4% among all patients; P<0.001; Table 1) were unwitnessed. Relatives were the most common type of bystander present in witnessed arrests, whereas there were significant differences in the involvement of OACWs, EMS personnel, and other individuals at the three locations. Compared with EMS personnel, there were more OACWs as bystanders in PAES and more bystanders, represented by the ‘other’ group, in PAES and on streets. There was a significant difference in the proportion of bystanders performing CPR among the three locations (P<0.001), as illustrated in Figure 2. The bystander CPR rate was the highest in PAES and the lowest in homes. Among the nine patients who received PAD, six received it when OACWs provided CPR in PAES.

Notably, asystole was the most common initial monitored rhythm (81.2% among all patients; P<0.001; Table 1). However, cardiac arrests on streets and in PAES had significantly higher rates of shockable rhythm and PEA, compared with cardiac arrests in homes. The prevalences of shockable rhythm relative to the time from receipt of EMS call and the location of cardiac arrest are shown in Figure 3. The highest rates of shockable initial rhythm (SIR) were observed within the first 5 and 10 minutes after receipt of EMS call for cardiac arrests on streets, which were 47% (8/17) and 41% (17/42), respectively.

Patients with cardiac arrest in homes had significantly longer intervals in terms of receipt of EMS call, initiation of bystander CPR, and receipt of defibrillation (all P<0.001; Table 1). The median interval for EMS to reach patients was 3 minutes longer in homes than on streets. The interval between recognition of cardiac arrest and receipt of EMS call was 0 minutes in 57.5% of patients (1297/2255).

Additionally, patients with cardiac arrest in homes had significantly lower rates of ROSC, 30-day survival, and favourable neurological outcomes (all P<0.001; Table 1).

Independent predictors of 30-day survival are shown in Table 2. Older age and longer call-to-EMS arrival interval both decreased the overall likelihood of survival (ORs of 0.92 and 0.87, respectively). Pulseless electrical activity and ventricular fibrillation/ventricular tachycardia increased the likelihood of survival compared with asystole (ORs of 6.4 and 15.6, respectively). Cardiac arrest witnessed by EMS personnel and defibrillation within 15 minutes after receipt of EMS call increased the overall likelihood of survival (ORs of 6.23 and 4.07, respectively).

The relationship among the location, timing of defibrillation, and 30-day survival of cardiac arrest is shown in Figure 4. Overall, patients who received defibrillation within 5 minutes and at 6 to 10 minutes after receipt of EMS call had survival rates of 33% (2/6) and 17% (15/86), respectively. For patients who received defibrillation on streets/in PAES within 5 minutes and at 6 to 10 minutes after receipt of EMS call, the survival rates were 50% (2/4) and 22% (10/45), respectively. Two patients in the homes group received defibrillation within 5 minutes; the survival rate was 0% (0/2).

Cardiac arrest at home was a predictor of survival in univariate analysis (OR=0.076, 95% confidence interval [CI]=0.038-0.15) but not in multivariable analysis (OR=0.65, 95% CI=0.22-1.90). The effect of location on survival was mediated by the first monitored rhythm, and the call-to-EMS arrival interval.

This study investigated factors that affect the prevalences of shockable rhythm and survival outcomes among cardiac arrests involving older adults in Hong Kong. The patient characteristics, proportion of witnessed arrests, and rates of SIR and PAD for cardiac arrests involving older adults in homes were similar between the present study and a previous analysis in Japan.9 Unlike many western countries, EMS personnel in Hong Kong and Japan generally do not terminate resuscitation in the field; this similarity facilitates comparison of data between the two studies. A notable difference was that in Japanese homes, 45% of older patients received bystander CPR; this receipt of CPR was associated with rate of ROSC, 30-day survival, and favourable neurological outcomes that were threefold higher than the corresponding rates in Hong Kong.

The bystander CPR rate in Hong Kong homes was low (3.8%) [Table 1], and there was a substantial delay in its initiation. Although the type of relatives involved as bystanders was not recorded in the present study, considering the proportion of older adult households in Hong Kong,6 many of the relatives presumably were cohabiting older adults. Such individuals may not be able to follow telephone instructions to perform CPR because of physical limitations or emotional distress10; thus, the provision of post-dispatch instructions and enhancement of community-wide CPR training will not improve survival among these patients.11 Although high-rise apartments create barriers to EMS personnel, they also increase the likelihood that trained volunteers will be present in the vicinity, where they may be dispatched using mobile applications.12 13 14

In non-residential locations, most bystanders performing CPR were not relatives of the patients. Fear of legal consequences is reportedly a major cause for intervention inertia in this situation.15 A previous survey in Hong Kong, in which one-third of respondents had prior first aid training, revealed that nearly all respondents were willing to call for help but only one-fifth were willing to perform bystander CPR.16 These findings suggest that knowledge transfer is insufficient to overcome bystander inertia in Hong Kong. Training programmes should ensure that factors inhibiting intervention (eg, legal concerns, fear of disease transmission, and bystander effect) are addressed.17 18

Previous studies in Hong Kong revealed low rates of SIR in patients with OHCA, ranging from 5% to 14%, along with dismal survival rates of 0.6% to 3%.1 19 20 These low rates imply that aggressive bystander interventions (eg, defibrillation for older adults) are futile. However, the findings of the present study indicate that older adults in non-residential locations have much higher SIR rates in the initial 10 minutes after receipt of EMS call; moreover, early defibrillation is an independent predictor of survival among such patients, and high survival rates can be achieved with early defibrillation.

The present study revealed a 2% per-minute decrease in the rate of SIR. This is similar to the findings in a large multinational study from northern Europe.21 Differences in SIR rates between residential and non-residential locations may be partly related to patient factors (eg, age and presence of co-morbidities); they could also be related to differences in the decision interval (ie, time elapsed between recognition of cardiac arrest [as reported by a bystander] and receipt of EMS call). A previous study in Hong Kong showed that efforts to seek advice from relatives often contributed to delayed receipt of EMS call.4 Longer decision intervals and consequential delays in EMS arrival lead to interactions with later parts of the shockable rhythm downslope and lower SIR rates. In practice, the recall of decision intervals by bystanders is unreliable. This is consistent with the decision interval of 0 minutes reported by most bystanders in the present study. Despite this confounding factor, the findings in this study indicate that bystanders should not hesitate to provide aggressive resuscitation and early defibrillation for older patients.

Notably, very few patients received PAD in this study, and most instances of PAD administration were performed by OACWs in PAES. According to a nationwide study in Japan, 16.5% of patients received PAD during witnessed ventricular fibrillation cardiac arrest.22 Estimation of the AED coverage rate in Hong Kong using a horizontal level walking route distance model revealed that only 11% of patients with OHCA would have an AED within 100 m.23 Considering the large number of OHCA events occurring within high-rise buildings, the actual coverage rate is presumably lower. Furthermore, there is evidence that most people in Hong Kong do not know the location of the AED nearest to their home or workplace.16 Unless AEDs are easy to locate and readily accessible at all times, PAD rates will remain low.24

In a previous study in Hong Kong, the proportions of patients with OHCA who could be accessed by elevator or stairs and by stairs alone were 74% and 14%, respectively.4 In the present study, the median interval for EMS to reach patients was 3 minutes longer in homes than on streets. This represents the ‘vertical response time’ component of the call-to-EMS arrival interval.25 In a previous study, survival was lower among patients who experienced cardiac arrest at higher levels within buildings.26 Nearly 70% of lifts in Hong Kong do not have sufficient area to accommodate the ambulance stretcher.4 Therefore, the vertical response time leads to a delay in EMS interventions and deterioration in CPR quality, both of which may contribute to the poor outcomes of cardiac arrests that occur in homes. The use of circulatory adjuncts to enhance cerebral perfusion during head-up position CPR within lifts should be considered.27

Importantly, only patients transported to hospital by FSD ground ambulances were included in this study; a small number of patients with OHCA may have been transported to hospital by other means.

Furthermore, data regarding the timings of recognition of cardiac arrest, bystander CPR, and PAD obtained from bystanders may have been subject to response bias. The lack of blinding of emergency department personnel towards patient factors (eg, absence of shockable rhythm and prehospital defibrillation, longer time to ROSC, co-morbidities, and advanced age) may have led to selection bias regarding treatment decisions, including the termination of resuscitation, arrangement of intensive care unit resources, and coronary angiography; such bias has been reported to negatively influence the survival rate.28 Data regarding pre-arrest co-morbidity and functional status were not available, which may have resulted in a confounding effect on survival outcomes. Additionally, a small number of patients received defibrillation within 5 minutes. All of the factors listed here may have affected the accuracy of conclusions drawn from this subset.

This study was based on territory-wide data collected in 2012 to 2013. Thus, it may not reflect the current situation because of changes in patient demography, prevalence of shockable rhythm, and survival enhancement interventions introduced in the past several years. A large multinational study in northern Europe investigated the rate of SIR among OHCA events occurring in residential and public locations from 2006 to 2015. The rate of SIR in public locations remained stable during that period. A decrease was observed in residential locations between 2006 and 2010, but the proportion has remained stable since 2011.21 Therefore, despite these limitations, the findings of the present study add to the broader understanding of OHCA involving older adults.

This study revealed significant differences in the patient and bystander characteristics and prehospital interventions among cardiac arrests involving older adults that occurred in homes, on streets, and in other public locations. Many older adults who experienced cardiac arrest in non-residential locations had a shockable rhythm in the early period after receipt of EMS call. Early defibrillation, an independent predictor of survival, was associated with favourable survival outcomes in older adults. These findings suggest that bystanders should provide aggressive resuscitation, including early defibrillation. Additionally, low rates of shockable rhythm and significant delays in bystander and EMS processes were observed within homes. New interventions (eg, volunteer dispatch via mobile applications) are needed to overcome unfavourable factors that affect cardiac arrests occurring within older adult households. Finally, the overall bystander CPR rate was low, indicating that additional measures are needed to overcome bystander inertia. The insights from this study will help to improve survival outcomes in OHCAs involving older adults.

The author contributed to the concept or design, analysis or interpretation of data, drafting of the manuscript, and critical revision of the manuscript for important intellectual content. The author had full access to the data, contributed to the study, approved the final version for publication, and takes responsibility for its accuracy and integrity.

The author has no conflicts of interest to disclose.

The author thanks Dr Ling-pong Leung, Emergency Medicine Unit of The University of Hong Kong, for providing the original dataset and permitting its use for secondary analysis in the study. The author also thanks Mr Min Fan and Ms Lujie Chen, both from Emergency Medicine Unit of The University of Hong Kong, for their technical support in this study.

This research received no specific grant from any funding in the public, commercial, or not-for-profit sectors.

This study is a secondary analysis of a historical cohort study which was approved by the Institutional Review Board of The University of Hong Kong/Hospital Authority Hong Kong West Cluster (Ref No.: UW 15-599). The requirement for informed patient consent was waived because of the retrospective study design. All patient data in the dataset were anonymous.

1. Census and Statistics Department, Hong Kong SAR Government. Hong Kong population projections 2020-2069. Available from: https://www.censtatd.gov.hk/hkstat/sub/sp190.jsp?productCode=B1120015. Accessed 7 Nov 2020.

2. Census and Statistics Department, Hong Kong SAR Government. The profile of Hong Kong population analysed by District Council district, 2018. Table 2. Proportion of land-based non-institutional population by District Council district and age group, 2017. Available from: https://www.censtatd.gov.hk/en/data/stat_report/ product/FA100096/att/B71807FB2018XXXXB0100.pdf. Accessed 24 Mar 2020.

3. Census and Statistics Department, Hong Kong SAR Government. 2016 Population By-census Office. Main tables (demographic). Population density by District Council district and year. 2017. Available from: https://www.bycensus2016.gov.hk/en/bc-mt.html?search=A202. Accessed 24 Mar 2020.

4. Leung LP, Wong TW, Tong HK, Lo CB, Kan PG. Out-of-hospital cardiac arrest in Hong Kong. Prehosp Emerg Care 2001;5:308-11. Crossref

5. Fan KL, Leung LP, Siu YC. Out-of-hospital cardiac arrest in Hong Kong: a territory-wide study. Hong Kong Med J 2017;23:48-53. Crossref

6. Census and Statistics Department, Hong Kong SAR Government. 2016 Population By-census. Domestic households in Hong Kong. Available from: https://www.bycensus2016.gov.hk/en/Snapshot-04.html. Accessed 24 Mar 2020.

7. Census and Statistics Department, Hong Kong SAR Government. 2011 Hong Kong Population Census. Census results. Available from: https://www.censtatd.gov.hk/en/scode170.html. Accessed 17 Mar 2023.

8. Perkins GD, Jacobs IG, Nadkarni VM, et al. Cardiac arrest and cardiopulmonary resuscitation outcome reports: update of the Utstein Resuscitation Registry Templates for Out-of-Hospital Cardiac Arrest: a statement for healthcare professionals from a task force of the International Liaison Committee on Resuscitation (American Heart Association, European Resuscitation Council, Australian and New Zealand Council on Resuscitation, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Southern Africa, Resuscitation Council of Asia); and the American Heart Association Emergency Cardiovascular Care Committee and the Council on Cardiopulmonary, Critical Care, Perioperative and Resuscitation. Circulation 2015;132:1286-300. Crossref

9. Okabayashi S, Matsuyama T, Kitamura T, et al. Outcomes of patients 65 years or older after out-of-hospital cardiac arrest based on location of cardiac arrest in Japan. JAMA Netw Open 2019;2:e191011. Crossref

10. Dami F, Carron PN, Praz L, Fuchs V, Yersin B. Why bystanders decline telephone cardiac resuscitation advice. Acad Emerg Med 2010;17:1012-5. Crossref

11. Kiyohara K, Nishiyama C, Matsuyama T, et al. Out-of-hospital cardiac arrest at home in Japan. Am J Cardiol 2019;123:1060-8. Crossref

12. Ringh M, Rosenqvist M, Hollenberg J, et al. Mobile-phone dispatch of laypersons for CPR in out-of-hospital cardiac arrest. N Engl J Med 2015;372:2316-25. Crossref

13. Smith CM, Wilson MH, Ghorbangholi A, et al. The use of trained volunteers in the response to out-of-hospital cardiac arrest—the GoodSAM experience. Resuscitation 2017;121:123-6. Crossref

14. Mao DR, Ong ME. High-rise residential resuscitation: scaling the challenge. CMAJ 2016;188:399-400. Crossref

15. Coons SJ, Guy MC. Performing bystander CPR for sudden cardiac arrest: behavioral intentions among the general adult population in Arizona. Resuscitation 2009;80:334-40. Crossref

16. Fan KL, Leung LP, Poon HT, Chiu HY, Liu HL, Tang WY. Public knowledge of how to use an automatic external defibrillator in out-of-hospital cardiac arrest in Hong Kong. Hong Kong Med J 2016;22:582-8. Crossref

17. Resuscitation Council UK. Cardiopulmonary resuscitation, automated defibrillators and the law. 2018. Available from: https://www.resus.org.uk/sites/default/files/2020-05/CPR%20AEDs%20and%20the%20law%20%285%29.pdf. Accessed 24 Mar 2020. Crossref

18. Sayre MR, Barnard LM, Counts CR, et al. Prevalence of COVID-19 in out-of-hospital cardiac arrest: implications for bystander cardiopulmonary resuscitation. Circulation 2020;142:507-9. Crossref

19. Wong TW, Yeung KC. Out-of-hospital cardiac arrest: two and a half years experience of an accident and emergency department in Hong Kong. J Accid Emerg Med 1995;12:34-9. Crossref

20. Lau CL, Lai JC, Hung CY, Kam CW. Outcome of out-of-hospital cardiac arrest in a regional hospital in Hong Kong. Hong Kong J Emerg Med 2005;12:224-7. Crossref

21. Oving I, de Graaf C, Karlsson L, et al. Occurrence of shockable rhythm in out-of-hospital cardiac arrest over time: a report from the COSTA group. Resuscitation 2020;151:67-74. Crossref

22. Kitamura T, Kiyohara K, Sakai T, et al. Public-access defibrillation and out-of-hospital cardiac arrest in Japan. N Engl J Med 2016;375:1649-59. Crossref

23. Fan M, Fan KL, Leung LP. Walking route–based calculation is recommended for optimizing deployment of publicly accessible defibrillators in urban cities. J Am Heart Assoc 2020;9:e014398 Crossref

24. Agerskov M, Nielsen AM, Hansen CM, et al. Public access defibrillation: great benefit and potential but infrequently used. Resuscitation 2015;96:53-8. Crossref

25. Silverman RA, Galea S, Blaney S, et al. The “vertical response time”: barriers to ambulance response in an urban area. Acad Emerg Med 2007;14:772-8. Crossref

26. Drennan IR, Strum RP, Byers A, et al. Out-of-hospital cardiac arrest in high-rise buildings: delays to patient care and effect on survival. CMAJ 2016;188:413-9. Crossref

27. Moore JC, Segal N, Debaty G, Lurie KG. The “do’s and don’ts” of head up CPR: lessons learned from the animal laboratory. Resuscitation 2018;129:e6-e7. Crossref

28. Winther-Jensen M, Kjaergaard J, Hassager C, et al. Resuscitation and post resuscitation care of the very old after out-of-hospital cardiac arrest is worthwhile. Int J Cardiol 2015;201:616-23. Crossref