Article Text
Abstract
Objective It remains unknown whether patient socioeconomic factors affect interventions and survival after out-of-hospital cardiac arrest (OHCA), and whether a socioeconomic effect on bystander interventions affects survival. Therefore, this study examined patient socioeconomic disparities in prehospital factors and survival.
Methods From the Danish Cardiac Arrest Registry, patients with OHCA ≥30 years were identified, 2001–2014, and divided into quartiles of household income (highest, high, low, lowest). Associations between income and bystander cardiopulmonary resuscitation (CPR) and 30-day survival with bystander CPR as mediator were analysed by logistic regression and mediation analysis in private witnessed, public witnessed, private unwitnessed and public unwitnessed arrests, adjusted for confounders.
Results We included 21 480 patients. Highest income patients were younger, had higher education and were less comorbid relative to lowest income patients. They had higher odds for bystander CPR with the biggest difference in private unwitnessed arrests (OR 1.74, 95% CI 1.47 to 2.05). For 30-day survival, the biggest differences were in public witnessed arrests with 26.0% (95% CI 22.4% to 29.7%) higher survival in highest income compared with lowest income patients. Had bystander CPR been the same for lowest income as for highest income patients, then survival would be 25.3% (95% CI 21.5% to 29.0%) higher in highest income compared with lowest income patients, resulting in elimination of 0.79% (95% CI 0.08% to 1.50%) of the income disparity in survival. Similar trends but smaller were observed in low and high-income patients, the other three subgroups and with education instead of income. From 2002 to 2014, increases were observed in both CPR and survival in all income groups.
Conclusion Overall, lower socioeconomic status was associated with poorer prehospital factors and survival after OHCA that was not explained by patient or cardiac arrest-related factors.
- cardiac arrest
- epidemiology
- quality and outcomes of care
- health services
Statistics from Altmetric.com
Introduction
Despite improvements in out-of-hospital cardiac arrest (OHCA) management and survival over time,1 2 OHCA still remains a major health problem worldwide.1–4 A common aim has been to identify factors that could guide future healthcare investments. Especially bystander interventions with cardiopulmonary resuscitation (CPR) and defibrillation have been a main focus with observed effect on survival.1 2 However, further improvements are warranted, and in 2015 the Institute of Medicine, USA, published a report calling for studies focusing especially on socioeconomic differences in patients with OHCA to help target future interventions.5
Socioeconomic differences can be challenging to examine due to its multiple dimensions and complexity that are often affected by many different factors.6–8 Furthermore, socioeconomic differences can be examined both on area level of the OHCA location illuminating important areas for targeted interventions, and on patient level that has been found with a higher predictive value for outcomes.7 Until now, OHCA studies have primarily focused on area-level socioeconomic factors with associations noted between higher socioeconomic status and positive prognostic factors as higher rates of witnessed arrests, bystander CPR and defibrillation,9–12 whereas data on survival are more conflicting.10 12
However, it still remains unclear how patient socioeconomic factors are associated with bystander CPR and survival after OHCA overall and over time, as well as whether a potential socioeconomic difference in bystander CPR might be associated with disparities in 30-day survival. Therefore, to help improve future strategies, this nationwide study explored patient socioeconomic factors in patients with OHCA overall and over time with the hypothesis that higher socioeconomic position would be associated with higher survival, potentially mediated through increased bystander CPR.
Methods
Data sources
This nationwide study was based on the Danish Cardiac Arrest Registry2 that includes all patients with a resuscitation attempt and holds detailed information of date and location of OHCA (private/public location), witnessed status by bystander or the emergency medical services (EMS), bystander CPR or defibrillation, first registered heart rhythm (shockable/non-shockable), estimated time interval from recognition of arrest to EMS rhythm analysis (based on time of 911 call and/or interview of on-scene bystanders) and survival status on hospital arrival. Using the unique civil registration number that all Danish residents are assigned, we linked information from several national administrative registries. Information on age, sex and vital status was obtained from the Central Person Registry; information on causes of death was obtained from the National Causes of Death Registry; information from hospital admissions including admission dates, discharge dates and discharge diagnosis codes used to define comorbidities (table 1) was obtained from the Danish National Patient Registry. All diagnosis codes are in accordance with the International Classification of Diseases (ICD-8/ICD-10) system.
Socioeconomic factors
Information of individual patient income and education was obtained from Statistics Denmark. Income was chosen as primary exposure whereas education was used as supplemental measurement to examine two different socioeconomic exposures.
Patient income was defined by household income calculated as average income over a 5-year period from the year prior to OHCA13 (to account for yearly variations and minimising potential influence of illness), corrected for inflation to year 2015 and weighted according to number of people in the household using the Organisation for Economic Co-operation and Developmen-modified scale where the first adult counts as 1 and further adults count 0.5 per person.14 The final study population was divided into four income groups (lowest, low, high, highest) using quartiles.
Patient education status was defined by the highest completed education classified according to the International Standard Classification of Education system allowing comparison internationally,15 and was divided into three groups: (1) basic education including elementary school, (2) high school and short secondary education, and (3) bachelor’s degree (Bachelor of Arts; BA), master’s degree (Master of Arts; MA) and doctoral degree or equivalent.
Study population
All patients with OHCA ≥30 years from 1 June 2001 to 31 December 2014 with presumed cardiac cause of arrest were identified from the Danish Cardiac Arrest Registry. The study period ended in 2014 due to unavailable data after 2014. Patients <30 years were excluded to minimise the inclusion of patients under education and with potential less stable income status. We excluded patients with missing data on arrest location, witnessed status and bystander CPR. Online supplemental figure 1 shows the selection process. In accordance with Utstein criteria16 and previous studies,2 OHCAs of presumed non-cardiac cause and EMS witnessed arrests were excluded.
Supplemental material
Outcomes
The main outcomes were bystander CPR and 30-day survival overall and over time. Bystander CPR was analysed as primary outcome and as mediator in relation to survival.
Statistics
Continuous variables were presented as medians with IQRs and tested with Kruskal-Wallis tests. Categorical variables were presented as frequencies with percentages and tested with Fisher’s exact test. Time trends (2002–2014) in bystander CPR and 30-day survival were tested using logistic regression adjusted for age in 10-year intervals and sex.
To examine socioeconomic disparities in 30-day survival we compared observed and counterfactual 30-day survival chances across income groups/quartiles using mediation analysis (see online supplemental file and figure 1 for details and references). We report the three outcomes: (1) the difference in observed 30-day survival across the income groups/quartiles, and based on the counterfactual analysis, (2) the remaining income disparity in 30-day survival if the chances of bystander CPR for all patients were identical to the highest income group/quartile, and at last (3) the eliminated income disparity in 30-day survival (the difference between the observed and the remaining income disparity). We performed the analysis in four subpopulations defined by all combinations of arrest location (private/public location) and witnessed status (unwitnessed/witnessed). All models were adjusted for age, sex, calendar year, education and comorbidities. We used cubic B-splines with prespecified knots at 62, 71 and 79 for age and 2005, 2009 and 2012 for calendar year. The knots were selected according to trends in the data. The 95% CIs were computed from non-parametric bootstrap procedure with 1000 bootstrap samples.
Supplemental material
Supplemental analyses were performed repeating the main analyses in subgroups defined by ages 30–65 years and >65 years, and using education instead of income. The latter adjusted for age, sex, comorbidities, income and calendar year. The level of statistical significance was set at 5%. For data management and statistical analyses, SAS V.9.4 (SAS Institute) and R V.3.6.1 were used.
Patient and public involvement
This registry-based study was done without any patient involvement.
Results
Overall characteristics
In total, 21 480 patients with OHCA were included (online supplemental figure 1). Higher income patients were overall younger (median age 61 years in highest (Q4) vs 77 years in lowest (Q1)), less often female (20.7% in highest (Q4) vs 37.0% in lowest (Q1)), had higher education (30.5% with BA, MA or doctoral degree in highest (Q4) vs 3.4% in lowest (Q1)) and had fewer comorbidities (table 1). They had higher chance of public located arrests, witnessed arrests, bystander CPR (57.6% in highest (Q4) vs 34.7% in lowest (Q1)), bystander defibrillation and shockable rhythm, as well as higher 30-day survival (19.4% in highest (Q4) vs 4.2% in lowest (Q1)) (table 1), also when stratifying after bystander interventions and shockable rhythm (online supplemental table 1).
Supplemental material
Temporal trends according to income
From 2002 to 2014, significant increases were observed for all groups in bystander CPR and 30-day survival (figure 2); p<0.001 for trends over time for all groups. No significant interactions were found between the groups and year (CPR: p=0.165; survival: p=0.344).
Adjusted odds for bystander CPR according to income
Figure 3 shows significant higher odds for bystander CPR in highest income (Q4) patients in reference to lowest income (Q1) patients in all subpopulations: private located witnessed: OR 1.54 (95% CI 1.30 to 1.82), private located unwitnessed: OR 1.74 (95% CI 1.47 to 2.05), public located witnessed: OR 1.64 (95% CI 1.29 to 2.10), public located unwitnessed: OR 1.66 (95% CI 1.18 to 2.34). A similar pattern but with lower odds was observed for low-income (Q2) and high-income (Q3) patients, except in public located unwitnessed arrests for high-income patients (OR 0.98, 95% CI 0.71 to 1.34).
Income disparities in 30-day survival
The highest income disparity in 30-day survival was observed in public located witnessed arrests resulting in an income disparity of 26.0% (95% CI 22.4% to 29.7%) between the highest income (Q4) and lowest income (Q1) patients (figure 4). The related remaining income disparity under the stochastic intervention (lowest income patients (Q1) having the same chance for bystander CPR as highest income patients (Q4)) was 25.3% (95% CI 21.5% to 29%). This eliminates 0.79% (95% CI 0.08% to 1.50%) of the income disparity in 30-day survival between the lowest income (Q1) and highest income (Q4) patients. Similar trends but less pronounced were found for low-income (Q2) and high-income (Q3) patients compared with the highest (Q4) as well as for the other three subpopulations (figure 4). Overall, the eliminated income disparity by the stochastic intervention of bystander CPR was found to be <1% across all income groups in all subpopulations (figure 4). Online supplemental figure 2 shows the estimated counterfactual 30-day survival probabilities for each income group in each subpopulation.
Supplemental material
Income disparities in 30-day survival in age groups
In ages 30–65, the income disparities in 30-day survival were smallest between lowest income (Q1) and highest income (Q4) patients, except in private located unwitnessed arrests (online supplemental figure 3). The highest income disparities were still found in public located witnessed arrests where the largest disparity in 30-day survival that could be eliminated by the stochastic intervention was in the lowest income group (Q1) (2.5%, 95% CI 0.02% to 4.9%) (online supplemental figure 3). Online supplemental figure 4 shows the related counterfactual 30-day survival probabilities. For patients aged >65, the same trends as for the total population were observed, but less pronounced (online supplemental figure 5). Online supplemental figure 6 shows the related counterfactual 30-day survival probabilities.
Supplemental material
Supplemental material
Supplemental material
Supplemental material
Educational disparities
Overall, the same trend in bystander CPR and 30-day survival was observed for groups of patient education as for income (online supplemental table 2). Educational disparities in 30-day survival showed also comparable results as for income (online supplemental figure 7), with <1% eliminated by the stochastic intervention of bystander CPR (patients with basic education or high school having the same chance for bystander CPR as patients with BA/MA/PhD).
Supplemental material
Supplemental material
Discussion
This nationwide study aimed to examine the potential effects of patient socioeconomic factors on bystander CPR and 30-day survival after OHCA overall and over time. Overall, we found a positive association of higher income patients having higher odds for bystander CPR and surviving 30 days after OHCA compared with lowest income patients. However, when we tried to account for the observed income disparity in bystander CPR in the survival analyses, no major changes in survival were observed. This, overall, indicates that, even though bystander CPR is very important for survival, socioeconomic differences in survival did not seem to be affected by socioeconomic differences in bystander CPR. Additionally, we found significant increases over time for all income groups in both bystander CPR and 30-day survival, indicating that the improvements in OHCA management over time have benefited all patients regardless of socioeconomic position. This study highlights the importance of understanding socioeconomic differences to help improve future interventions in patients with OHCA.
Over time, several improvements have been made in OHCA management and survival has subsequently increased.1 2 Our study found significant increases in both bystander CPR and survival in all income groups from 2002 to 2014, indicating that the improvements benefited all patients regardless of socioeconomic position.2
However, overall, we found associations of higher odds for bystander CPR and survival in patients of higher socioeconomic status (income and education). In terms of bystander defibrillation, we chose to focus on bystander CPR because only a small number of patients were defibrillated and nearly all these patients received bystander CPR. Though we acknowledge that bystander defibrillation is important for survival, we believe that a direct impact of bystander defibrillation in this study is likely to be small. Therefore, focusing on bystander CPR, we assessed the pathway from patient socioeconomic factors via bystander CPR to 30-day survival, including both important patient-related and cardiac arrest-related factors. Higher income patients had more positive prognostic characteristics as younger age17 and subsequent lower comorbidity burden,18 and higher chance of public arrests, witnessed arrests, bystander CPR and defibrillation.9 10 12 The younger age in higher income patients could be a result of still active-working patients when suffering from their OHCA that again could improve the chances for a public located witnessed arrest. However, when including these patient factors through adjustments and examining survival in subgroups of private/public location and witnessed/unwitnessed status, and including the observed socioeconomic disparity in bystander CPR, our analyses showed overall evidence for an almost non-existing effect of the observed socioeconomic disparity in bystander CPR on survival. This underlines that if at all, only a small amount of the difference in survival according to patient income (or education) is driven by differences in bystander interventions, whereas the majority of the socioeconomic differences in survival must be driven by other factors. Interestingly, the biggest effect of socioeconomic disparity in bystander CPR on survival was observed in the public located witnessed arrests in ages 30–65.
This study therefore indicates that patient socioeconomic factors can influence outcomes after OHCA. Yet, the mechanisms remain unknown and are probably multifactorial, as a direct causal relationship between a patient’s socioeconomic position and survival from OHCA is doubtful in a strictly biological perspective. Socioeconomic position may then reflect something else, and potentially be a result of accumulated risk factors throughout life.19 Patients of higher socioeconomic position could exercise more, have lower rates of active smoking, obesity and alcohol consumption,20 21 and have better knowledge of own health status,22 which again could reflect differences in healthcare contacts.23 The observed effect could also be related to disparities in cardiac arrest-related factors as this study found more arrests in private homes and of unwitnessed status in low-income patients, which again could affect time for arrest recognition and bystander interventions. However, our analyses tried to account for this by examining the outcomes in subgroups of arrest location and witnessed status. However, both the recognition, time of 911 call, dispatch assistance and effectiveness of CPR could be issues that could affect the findings since the quality of the bystander intervention naturally depends on the person who intervenes which might differ according to socioeconomic position.21 This supports our finding of a greater effect of bystander CPR in ages 30–65 in public located witnessed arrests. Other factors as for example differences in EMS response and in-hospital factors could also affect the outcomes. However, data were not available for the current study.
Our observed socioeconomic differences are surprising as Denmark overall is thought of as a very homogenous country with only 0.7% of the Danish population termed ‘economically poor’24 compared with 13.5% in the USA.25 Furthermore, Denmark has a high rate of basic life support-trained individuals.26 Therefore, the observed differences could be greater in other parts of the world. Even though socioeconomic differences in bystander CPR did not explain the socioeconomic differences in survival in Denmark, it is still well known that these more easily modifiable bystander interventions (CPR and defibrillation) are important for survival in all patients regardless of socioeconomic status. Understanding socioeconomic differences can potentially help improve future strategies and more studies on this are warranted.
Limitations
Because of the observational study design, the findings represent associations and do not prove causation. An important limitation is the exclusion of patients with missing information on arrest location, witnessed status or bystander CPR in order to perform complete case analyses. However, no major differences were observed in patient characteristics between the included and excluded patients in sensitivity analyses. Another limitation is due to the estimates of the counterfactual survival probabilities that can be interpreted causally under rather strong identifiability assumptions: consistency, positivity and exchangeability.27 When consistency and positivity are mostly technical, exchangeability is violated if there are unmeasured variables that confound the relationship between bystander CPR and 30-day survival. Such factors could be smoking, obesity, physical activity, and so on. Unfortunately, we did not have data on these factors but used comorbidities as proxies. Also, our data were limited regarding specific cause of arrest; as most OHCA studies, we only had data on presumed cause of arrest.
Conclusion
Despite nationwide improvements in bystander CPR and 30-day survival from 2002 to 2014, we found significant socioeconomic disparities with higher income patients receiving more bystander CPR and having higher 30-day survival compared with low-income patients. These results persisted when including age, sex and comorbidities, important cardiac arrest-related factors and when accounting for the socioeconomic disparity in bystander CPR. Socioeconomic differences in survival may therefore also be driven by other factors than disparities in interventions. More research is required to improve OHCA management for all patients regardless of socioeconomic status.
Key messages
What is already known on this subject?
Over time, improved out-of-hospital cardiac arrest (OHCA) prehospital care and survival has been observed; however, survival remains low. Socioeconomic factors have previously been examined in relation to patients with OHCA, but have primarily focused on area-level socioeconomic factors. These studies have noted associations between higher socioeconomic status and positive prognostic factors as higher rates of witnessed arrests, bystander cardiopulmonary resuscitation (CPR) and defibrillation whereas data on survival have been more conflicting.
What might this study add?
Instead of examining area-level socioeconomic factors in relation to prehospital care and survival in patients with OHCA, this nationwide study examines the association of the individual patient socioeconomic factors, bystander CPR and survival after OHCA overall and over time. Additionally, this study examines whether a potential socioeconomic difference in bystander CPR is also associated with disparities in 30-day survival.
How might this impact on clinical practice?
Understanding socioeconomic differences and the potential influence of socioeconomic factors in relation to care and survival after OHCA can potentially help improve future strategies for all patients regardless of socioeconomic position.
Acknowledgments
Thanks to the EMS personnel who supply the Danish Cardiac Arrest Registry with OHCA data.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors SM contributed to the conception and design of the study, the data acquisition, the data analysis, the data interpretation, the manuscript drafting and the critical revision of the manuscript. SM has full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. MW, LS, CTP and TAG contributed to the conception and design of the study, the supervision, the data acquisition, the data analysis, the data interpretation, the manuscript drafting and the critical revision of the manuscript. KK contributed to the conception and design of the study, the data interpretation, the manuscript drafting and the critical revision of the manuscript. SMH, KBR, FF, JA, CMH, FL, LK and GHG contributed to the data interpretation, the manuscript drafting and the critical revision of the manuscript.
Funding SM has received grants from Karen Elise Jensens Fond, Laerdal Foundation and Helsefonden. FF has received grants from Novo Nordisk Foundation and Laerdal Foundation. CMH has received grants from TrygFonden, Helsefonden and Laerdal Foundation. KK has received grants from Laerdal Foundation. GHG has received grants from Novo Nordisk. CTP has received grants from Bayer and Novo Nordisk.
Disclaimer The Danish Cardiac Arrest Registry is supported by TrygFonden but has no commercial interests in the OHCA area, and has no influence in the management, design, data collection, analyses, interpretation of data, preparation, review, reporting, manuscript approval or submission decision for publication of this study.
Competing interests LK has received lecture fees from Sanofi and Novartis.
Patient consent for publication Not required.
Ethics approval The Danish Data Protection Agency (reference number 2007-58-0015, local reference number GEH-2014-017, I-Suite number 02735) has approved this study. In Denmark, registry-based studies do not require ethical approval.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. The data in the manuscript are completely deidentifiable data.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.