Social and clinical vulnerability in stroke and STEMI management during the COVID-19 pandemic: a registry-based study


ST-segment elevation myocardial infarction (STEMI) and stroke are life-threatening and highly time-sensitive emergencies. The time elapsed from symptom onset to treatment is a predictor of patients’ mortality and functional recovery.1 2 The standardised and timed care pathways for these two diseases depend initially on a patient’s use of the emergency medical service (EMS) system, followed by close collaboration between emergency structures and specialised technical platforms (eg, catheterisation laboratories, stroke units).1 2 The quality of care is often evaluated under the prism of the time from first medical contact (FMC) to treatment.1 2

In France, patients with acute chest pain or neurological deficit are advised to rapidly call the nationwide EMS using a unique medical dispatch number. In cases of suspected stroke or STEMI, the EMS dispatches rapid transport, including a doctor for STEMI and life-threatening situations, to transfer the patient to a specialised technical platform. If not suspected, the EMS physician may refer the patient to a general practitioner for initial evaluation or advise them to go to the emergency unit (EU).

Patients with STEMI and stroke face social and health inequality issues. Socially vulnerable (ie, disadvantaged) patients with neurocardiovascular diseases have higher morbidity and mortality rates.3 4 Four markers of social position and socioeconomic status have been associated with cardiovascular disease in high-income countries: income level, educational attainment, employment status and environmental factors.5 These inequalities are attributable to a higher prevalence of biological, behavioural and psychosocial cardiovascular risk factors in the more socially disadvantaged population but also to more difficulties in accessing healthcare and lower-quality acute care management.4 6 7 The organisation of the healthcare system, as a social health determinant, leads to health inequalities, due mainly to challenges related to communication and health literacy, implicit bias, and/or a lack of culturally competent care.8

The COVID-19 dramatically modified healthcare systems worldwide and had major consequences for patients’ access to care for stroke and STEMI.9–11 From February to March 2020, many health authorities, including those in France, implemented strict nationwide lockdowns and series of policies to curb the surge of patients requiring critical care. This crisis, and particularly the lockdown periods, induced the major reorganisation of healthcare systems and modified the use of care to accommodate the onslaught of patients with COVID-19.12 Studies of the association between the COVID-19 pandemic and the quality of stroke and STEMI management have yielded contrasting results, with most revealing longer management delays and reductions in the number of procedures performed.9 10 13

During pandemics (eg, of influenza, plague), pre-existing inequalities affecting many aspects of patients’ care pathways (eg, loss of employment and income; social isolation, especially for elderly individuals; and mental health issues, particularly for young people) are usually amplified.14–18 During the COVID-19 pandemic, COVID-19 exposure, severe disease, hospitalisation and death were more frequent among socially disadvantaged people.15 17–19 This population benefited less from the collective protective measures taken against COVID-19, had more difficulty accessing preventative healthcare and had lower rates of COVID-19 testing and vaccination.14 Some experts consider COVID-19 to be a syndemic, rather than a pandemic. These interactions between COVID-19 and pre-existing socioeconomic inequalities in non-communicable diseases are an illustration.16 Indeed, ‘syndemics are characterised by biological and social interactions between conditions and states, interactions that increase a person’s susceptibility to harm or worsen their health outcomes.’16 We hypothesised that socially vulnerable patients, defined as those with low socioeconomic status, may experience longer acute management times during the COVID-19 pandemic.

In France, to protect more vulnerable patients and adapt care, health authorities identified several risk factors of severe COVID-19 based on demographic (advanced age) and medical (especially cardiovascular comorbidities) characteristics.20 For these populations defined as ‘clinically vulnerable patients’, French authorities have stressed the importance of adhering to barrier measures, maintaining physical distancing, particularly during hospitalisation and to limit travel to high-risk areas for SARS-CoV-2 transmission. Information about these risk factors was covered widely in the media, which may have led exposed individuals with these underlying conditions to delay seeking treatment.21 Based on these recommendations, we hypothesised that additional protective measures may have been implemented for these clinically more vulnerable populations, resulting in increased management delays.

To our knowledge, only one study has evaluated whether COVID-19 modified the associations among the educational level, deprivation, hospital admission indicators and quality of hospital care, especially for patients with neurocardiovascular diseases.22 The researchers found larger declines in the hospital access of women, elderly and less-educated individuals; in contrast, the timeliness of percutaneous coronary intervention (PCI) showed no education-related or deprivation-related gradient.

Since 2012, the ‘CNV Registry’ of neurocardiovascular diseases evaluate the care pathways for STEMI and stroke patients in the Aquitaine region of southwestern France (3 million inhabitants). This registry provides a unique opportunity to study differences in care management and their evolution over time in a country with universal health coverage.23

COVID-19 profoundly modified access to and the use and organisation of care, against a backdrop of pre-existing inequalities in neurocardiovascular disease.12 The notion of a ‘syndemic’ and our hypothesis that management times were longer for patients at risk of severe COVID-19 during its first wave prompted our investigation of whether first COVID-19 wave resulted in the deterioration of the quality of care for socially and clinically vulnerable stroke and STEMI patients, using data from the CNV Registry.


Study design and population

We used two exhaustive retrospective cohorts of adult stroke and STEMI patients admitted to a care structure in the Aquitaine region whose data were entered into the CNV Registry between 1 January 2019 and 31 August 2020.23

The STEMI cohort comprised patients with recent (<24 hours after symptom onset) STEMI managed in one of the six health territories in Aquitaine, each centred around an EMS, comprising 30 EUs and 11 catheterisation laboratories (nearly 1800 STEMIs are seen annually).

The stroke cohort comprised patients with recent ischaemic or haemorrhagic stroke (excluding transient ischaemic attacks), diagnosed by brain imaging and validated by neurovascular physicians, which was managed in 5 health territories in Aquitaine, comprising 14 (7 with stroke units) of the 20 hospitals caring for more than 30 strokes per year in Aquitaine (nearly 5000 strokes are seen annually).

Data collection

The CNV Registry contains information on patients’ sociodemographic (age, gender and place of residence) and clinical (medical history, cardiovascular risk factors, stroke clinical severity (modified Rankin scale and National Institute of Health Stroke Score), stroke type (ischaemic/haemorrhagic)) characteristics, use of care (calls to emergency services, FMC, symptom onset care time), acute care management quality (times between key management steps, prehospital and hospital pathway types, treatment) and structural characteristics of care (care during on-call activity, calls to emergency services during care, hospital administrative status, FMC–catheterisation laboratory distance). Data are collected prospectively by physicians; consolidated retrospectively by clinical research assistants and then extracted from the hospital information system. Data from the two cohorts were integrated into one data warehouse enabling the reconstruction of the STEMI or stroke management pathway.12


The primary endpoints were acute care management times, which reflect the quality of care. For the STEMI cohort, we used the FMC procedure time (delay (in minutes) between the FMC (mobile intensive care unit arrival or EU admission) and the start of a treatment procedure). For the stroke cohort, we used the EU admission imaging time (delay (in minutes) between EU admission and the start of the first imaging). This selection of an interval that focused on the beginning of in-hospital stroke care was required due to the heterogeneity of the prehospital pathways and treatments applied.


Clinically vulnerable persons at risk of severe COVID-19 were those aged >65 years; with neurocardiovascular history including previous STEMI, stroke or transient ischaemic attack; and/or with coronary artery disease history. For the STEMI cohort, the history of a PCI, a coronary artery bypass graft was also included.

Due to the lack of individual socioeconomic data, an ecological social deprivation score was assigned to each commune (the smallest administrative unit in France) of patients’ residence using the 2015 deprivation index (Fdep15) to assess social vulnerability.24 This index is associated strongly with mortality at all geographical scales. It served as the first dimension of a principal component analysis (weighted by population size) of four socioeconomic ecological variables: the percentage of high-school graduates ≥15 years old, median household income, percentage of blue-collar workers and unemployment rate. Quintiles of the Fdep15 scores were computed for metropolitan France, whereby the first quintile (Q1) represented the least and the fifth quintile (Q5) the most disadvantaged communes. We calculated the deprivation score for each patient of our sample with reference to the quintiles of the French population.

Statistical analyses

Analyses were performed separately for each cohort and exposure variable. Associations between clinical and social vulnerabilities’ effects on care management times (introduced as continuous variables after logarithmic transformation) were analysed using multivariate linear mixed models (three each for stroke and STEMI), with random effects on hospital and health territories centred around single EMSs. Interactions in the time period were introduced. Three COVID-19 periods were defined according to the dates of first hospital reorganisation (mid-February) and the termination of national lockdown (10 May 2020): pre-wave (1 January 2019–9 February 2020), per-wave (10 February 2020–10 May 2020) and post-wave (11 May 2020–31 August 2020). Inspired by the conceptual framework developed by the Health Care Quality Indicator Project of the Organisation for Economic Co-operation and Development, we categorised determinants in four dimensions: patients, physicians, care organisations and quality of care.25 To develop the causal model, variables were classified into each of these dimensions and confounders were then identified by directed acyclic graphs (DAG; online supplemental material 1). The relationships between vulnerabilities and care management times were quantified (β) using the contrast method, with statistical significance defined as p<0.05. The exponentials of the beta values (exp(β)), associated 95% CIs, and percentage changes (1−exp(β)) were then calculated. The statistical analyses were conducted by using SAS V.9.4.

Supplemental material

Patient and public involvement

As members of the CNV registry scientific boards, patient representatives were involved in study conception, implementation and dissemination; they validated data collection and analysis, and results diffusion. Dissemination of results was conducted on the CNV registry website, to the scientific boards and to care-structure physicians.

This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology guidelines and is registered with (NCT04979208).


Main results

This analysis of the healthcare pathways for STEMI and stroke patients included in the CNV Registry showed that care management times for socially or clinically vulnerable patients did not worsen during the first wave of the COVID-19 pandemic, despite changes in the access to and use and organisation of care. Nonetheless, regardless of the COVID-19 period, acute care management times were longer for elderly and the most disadvantaged STEMI patients.

Social and clinical vulnerability in stroke and STEMI management during the COVID-19 pandemic

Our results are concordant with those of a study conducted in Italy, which revealed no educational or deprivation gradient for cardiovascular acute care management times.22 Several factors can explain the resilience of stroke and STEMI care pathways for vulnerable populations.

First, STEMI and stroke networks in France are structured as well-defined, organised and dedicated pathways. Highly structured patient-centred clinical pathways improve the quality of care for chronic and acute conditions with predictable trajectories.26 27 Guidelines and national stroke and STEMI improvement programmes recommend the implementation of such structured pathways, which include close collaborations between healthcare professionals and patient orientation to the EMS system and specialised technical platforms. A study of the impacts of changes in the use of care and implementation of hospital reorganisation spurred by the COVID-19 pandemic on acute management times for stroke and STEMI revealed no deep alteration of the emergency pathway construct.12 Socially and/or clinically vulnerable populations have also benefited from the resilience of the STEMI and stroke pathways.

Second, in the particular context of the first COVID-19 wave, the mass media widely relayed information from health institutions. The whole population was worried and very concerned about its health. Lockdown measures made people more available, and routinely exposed, to mainstream media that were highly focused on the pandemic and health messages. These factors are associated with a greater likelihood of the adoption of recommended prevention practices.28 Thus, broad health-related media coverage may have had a positive influence on health literacy for the whole population, which may have positively influenced the use of the healthcare system.29

Third, the EMS was identified as the first contact to limit exposure and regulate urgent calls during the first COVID-19 wave in France. The media relayed this information. French hospitals increased regulation capacities to face the rise in EMS calls, in an attempt to preserve access to care and the capacity to handle vital emergencies for the entire population.30

Fourth, France adopted a specific strategy in March 2020 to support the economy, companies and jobs through measures that include financial support for disadvantaged populations, salary preservation, the prohibition of layoffs and housing assistance.31 Associated with universal healthcare coverage, these actions may have contributed to mitigate the social consequences of the pandemic.

Social and clinical vulnerability in stroke and STEMI management regardless of the COVID-19 pandemic

Several studies, including the present work, have shown that acute care management times are longer for elderly and socially vulnerable STEMI patients.32–34 Concerning stroke, we found no alteration in the acute care management time for elderly and socially vulnerable stroke patients. The results pertaining to stroke patients may be explained by our examination of the EU admission imaging time focused on the beginning of in-hospital care. Unlike the STEMI pathway, this time involves such a small portion of stroke patients’ pathways that it could have been difficult to detect an effect.


Regarding specifically age for STEMI patients, greater initial clinical severity, atypical symptoms and a longer delay in admission may explain these findings.33 Half of the STEMI patients in our sample were aged >65 years. The proportion of stroke patients >65 years was 81%, which made it difficult to demonstrate an effect. To our knowledge, only one study, conducted in England, has revealed an association between older age and a longer admission CT time for stroke patients.35

Socioeconomic status

Findings with respect to socioeconomic status do not converge for STEMI. Biswas et al32 found that the median time to reperfusion in Australia, a country with universal healthcare, between 2005 and 2015 was 4 min longer for lower socioeconomic quintiles than for the highest quintile. Vasaiwala and Vidovich34 found a direct correlation between income levels in the USA and the proportion of patients meeting the guideline-recommended door–balloon time. In contrast, Heo et al36 found no association between the educational level and door–balloon time in Korea. None of these studies involved control for the confounders. Additional dedicated analyses of the relationship between socioeconomic status and acute care management time are needed, especially for elderly patients with accumulated comorbid factors due to their disadvantaged status.

Few studies have involved the exploration of acute stroke management times according to socioeconomic status, with contrasting results explained by the specificities of healthcare systems.3 37 In a study conducted in England, socially vulnerable patients were less likely to undergo high-quality recommended processes and more likely to undergo early supported discharge.3 A study conducted in Sweden showed that university-educated patients were more likely to be treated than were less-educated patients.37

Neurocardiovascular history

Regardless of the COVID-19 period, we found no significant influence of patients’ neurocardiovascular history on acute care management times, consistent with reported findings for STEMI patients.38 To our knowledge, no other study has evaluated this relationship for stroke care.

Implications for clinical practice and health system performance

While the COVID-19 pandemic crisis is nearly resolved, our findings remain valuable for health institutions and professionals to prepare for future health crises. The structured emergency pathway for strokes and STEMI patients and hospital reorganisations ensured sustained care quality.12 In our study, the COVID-19 crisis did not have any differential impact on social health inequalities, suggesting a good resilience of the French healthcare network. Organisational strategies employed, such as a dedicated life-threatening emergency pathway, transversal reorganisations aiming at concentrating resources on emergency care,12 targeted communication and increased regulation capacities, could be replicated in new crises and extended to other conditions. Pre-existing STEMI management inequalities partly result from the healthcare system organisation. In a study about disparities in cardiovascular disease, these inequalities are linked to language challenges, health literacy, implicit bias and the absence of culturally competent care.8 This may lead to less accurate medical interviews and suboptimal medical decisions. Further research is essential to investigate these hypotheses and evaluate potential corrective measures.

Strengths and weaknesses

Our study, one of the first to examine the effects of the COVID-19 crisis on the quality of care for STEMI and stroke patients in Europe with consideration of health and social inequalities, involved the parallel analysis of two high-quality databases containing data on large numbers of stroke and STEMI patients managed in a large panel of care structures in the Aquitaine region.

Our study has some limitations, particularly with regard to the population. The study area was limited to the Aquitaine region, one of the regions least affected by the first wave of the COVID-19 pandemic.39 This situation could have led to the exertion of less pressure on health services (especially the EMS, STEMI and stroke network). Arguments support the sample’s representativeness for stroke and STEMI patients in hospitals during this period, making our results likely applicable to all of France. First, a stroke study showed that the use of care was similar regardless of pandemic intensity.40 Second, a previous study with the same database highlighted results consistent with other French studies on the evolution of stroke and STEMI patient admissions.12 Third, characteristics and acute management times for stroke and STEMI patients in the ‘CNV registry’ align with those in other French regions. It would be interesting to repeat the study in another region, or in another country more affected by the pandemic, to test the external validity of the results.

Moreover, patients who did not enter the healthcare system because they had died or did not benefit from hospital care, as well as STEMI patients with symptoms for >24 hours, were not included. The exclusion of these patients may have generated selection bias and prevented us from quantifying the phenomenon of healthcare system avoidance that could be supposed to be more frequent among socially and/or clinically vulnerable patients during the COVID-19 crisis, as stated in a Danish study41; it also entails the risk that increases in the delay to use of care were underestimated for some patient subgroups. A French study revealed a 24% decrease in emergency consultations for STEMI and an 18% decrease in stroke.42 However, a national survey analysed the characteristics associated with not seeking care, in 2017 and 2020, revealing factors such as younger age, foreign nationality, living alone and lack of general practitioner care.43 The proportion of patients not seeking care increased during COVID-19 pandemic, but the population was not significantly different from the one before, suggesting a limited selection bias.

Our explanatory analyses yield robust results, with the inclusion of appropriate confounding variables identified by the DAG method. The large panel of data collected enabled the integration of a wide variety of confounders, including clinical characteristics and sociogeographical factors.

Given the lack of individual-level socioeconomic data in the CNV Registry, which prevented the assignment of social determinants for each patient, we used a residence area-based measure, which is a major limitation of our study. However, we determined deprivation indices using a validated tool that has been used in many studies conducted in France.24 Moreover, the socioecological measure of deprivation tends to underestimate social inequalities observed using individual data; thus, caution is advised when attributing group-level estimates to individuals.6 Additional limitations of this study include our inability to include all clinical risk factors of severe COVID-19 and information about patients’ educational levels, individual resources and social support to further explore their precariousness and health literacy. The COVID-19 pandemic may have had a greater impact on the access to and quality of care for the most precarious individuals.

A major methodological issue of this study is that we defined the per-wave period according to the implementation of healthcare reorganisation and transformation of societal functioning to fight the COVID-19 pandemic.12 We began the per-wave period at the time of the first hospital reorganisation implementation and ended it at the time of lockdown lifting. Although data for the CNV Registry are collected continuously, we terminated the follow-up period at the end of August 2020 to enable the timely reporting of results.

Finally, we did not explore gender as a distinct vulnerable group9 and short-term or long-term outcomes such as morbidity, mortality, disability or rehospitalisation after initial hospitalisation for STEMI or stroke, for which a wide range of socioeconomic disparities exist.3 41 Separate studies on gender inequalities and inequalities following acute care are currently underway, with a focus on the COVID period.

This post was originally published on