Influence of Socioeconomic Status and Gender on Stroke Treatment and Diagnostics
Background and Purpose— Female patients and patients with a lower level socioeconomic status (SES) often experience less favorable health outcomes. We aimed to investigate whether gender and socioeconomic status of patients with acute ischemic stroke or transient ischemic attack are associated with different diagnostic and treatment strategies.
Methods— The risk factors of interest were gender and SES. We predefined our diagnostic and treatment end points according to current stroke guidelines and used multivariate models to adjust for age, stroke severity, and comorbidities.
Results— A total of 2606 patients were included in the analysis. Women were less likely to receive antiplatelet agents within the first 48 hours after admission (OR: 0.59, 95% CI: 0.53 to 0.89) and more likely to have their blood glucose measured on admission than men (OR: 1.52, 95% CI: 1.1 to 2.1). With higher SES patients were more likely to receive a TTE or TTE during hospital stay. Women were almost twice as likely to receive a prescription for antidepressants at discharge OR of 1.96 (95% CI: 1.48 to 2.59).
Conclusion— Socioeconomic status and gender are associated with some diagnostic and treatment differences of acute ischemic stroke. Most pronounced were a reduced chance for women to receive antiplatelet therapy on admission and a reduced chance for a TTE and TEE with a lower level of SES, whereas the rate of thrombolysis was unbiased by gender and SES.
Even in countries where each citizen is granted equal access to the healthcare system, women and people with a less favorable socioeconomic status (SES) seem to have disadvantages in acute and secondary medical care of cardiovascular diseases.1–4 Among the cardiovascular diseases stroke is one of the biggest health problems with a yearly incidence between 4% to 15% in the older age groups and 1 of 3 leading causes of death in many WHO member states.5 Studies have investigated gender differences in the acute, intermediate, and long-term treatment of stroke and came to varying results.6–8
Austria is 1 of the countries with a universal healthcare system and a low level of income inequality,9,10 implying that socioeconomic status and gender treatment differences should be small. Against this background we aimed to investigate whether SES or gender influences stroke diagnostics and treatment.
Subjects and Methods
This cohort study was based on a prospective population-based stroke registry of patients from an urban population admitted to all neurology departments in Vienna, Austria, which serve a community of 1.9 million people.11 On the basis of informed consent we prospectively documented data on all patients with ischemic stroke or transient ischemic attack (TIA) who were admitted to a neurology department within 72 hours of symptom onset between October 1998 and March 2003. We used a structured interview to assess clinical and neurological variables (such as the score on the National Institutes of Health Stroke Scale [NIHSS]), demographic factors, SES, cardiovascular risk factors, and comorbidities. The study was approved by the local ethics committees.
Inclusion and Exclusion Criteria
The diagnosis of ischemic stroke or TIA was made by one of the treating neurologists and confirmed by a subsequent CT or MRI scan. Our definition of stroke relied primarily on the clinical presentation, which complied with the definition of the World Health Organization.12 We later quantified the severity of each event by using stroke scales such as the NIHSS and the Scandinavian Stroke Scale and classified it by subtypes according to the Bamford Classification.11 We excluded patients in whom a nonischemic cerebrovascular event (eg, cerebral hemorrhage) was confirmed on admission or at a later stage.
We investigated 2 primary risk factors: (1) socioeconomic status (SES) and (2) gender. As a proxy for SES we used education as it is robust against changes over time and against gender differences caused by marriage and motherhood. It incorporates knowledge about stroke symptoms and treatment which may contribute to the results in this context. The categories for education were as follows: (1) no basic school education, (2) secondary school graduation, (3) technical training or apprenticeship, (4) higher secondary school degree, and (5) university or college degree. “Secondary school graduation” refers to patients that completed the first 9 years of compulsory schooling in Austria. To adjust for confounding we added a number of risk factors to the model.
From all available therapeutic variables of the database we prospectively chose diagnostic and treatment variables according to the guidelines by the American Stroke Association13,14 focusing on definitive recommendations. Variables that were applicable to every patient in the database (cardiac monitoring) or were congruent with our inclusion criteria (CT and /or MRI scan) could not be used. Additionally to the recommended measures we added transesophageal and transthoracic echocardiography (TEE and TEE) during hospital stay as we were interested in socioeconomic differences, and we added antidepressants at discharge where we suspected gender differences.
Hence our end points were:
Antiplatelet agents up to 36 hours after admission
Systemic thrombolysis on admission
Measurement of blood glucose on admission
Measurement of temperature on admission
Transesophageal echocardiography (TEE) during hospital stay
Transthoracic echocardiography (TTE) during hospital stay
Physiotherapy during hospital stay
Speech therapy during hospital stay
Occupational therapy during hospital stay
Antihypertensive drugs at discharge
Oral anticoagulation or antiplatelet agents at discharge
Lipid lowering drugs at discharge
Antidepressants at discharge
All end points were dichotomized (yes/no).
We used the unpaired t test, the χ2 tests, Mann–Whitney U test, ANOVA, Kruskal–Wallis test, as appropriate to perform univariate comparisons of groups. We used multivariate logistic regression models to assess the association between the end points and either gender or SES as risk factors. We adjusted for potential confounders that are known to be at least partly associated with SES and gender, and may influence the end points (age, stroke severity by NIHSS, history of hypertension, diabetes, hyperlipidemia, smoking status, history of stroke, peripheral vascular disease, and ischemic heart disease), and for either SES or gender, depending on which was the risk factor of interest. Selection of confounding variables was predefined and by clinical considerations only. To adjust for clustering on level of the different departments we used random effects models. If random effects models were not appropriate (assessed by the “quadcheck” command) we used robust standard errors. To identify possible interactions between the risk factors of interest and cofounders, we compared models with and without interaction terms using the likelihood ratio test and assumed significance of interactions at P=0.05. If interaction was present, it was further examined using stratification. All analyses were carried out using Stata (v.8; StataCorp).
Between October 1998 and December 2001, 3607 patients with suspected ischemic stroke or TIA were registered (for simplicity, referred to as ischemic stroke). Of these, we excluded 679 because the final diagnosis was a nonvascular cause of the symptoms, 252 because they had an intracerebral hemorrhage, and 30 because the underlying cause was another cerebrovascular event, such as a subarachnoid hemorrhage or a cerebral sinus thrombosis. Of the remaining patients, we excluded 40 who were admitted to the neurology department after 72 hours of symptom onset. A total of 2606 patients were available for analysis. Their baseline characteristics are presented in Table 1. Of all patients 47% were women, their mean age was 67 years (SD 14), and the median NIHSS score was 4 (IQR 2 to 8). Baseline characteristics varied across gender and levels of school education. At the time of the stroke event women were generally older (70 years [SD 14] versus 65 years [SD 13]) and had a more severe strokes than men (median NIHSS 5 [IQR 2 to 10] versus median NIHSS 4 [IQR 2 to 7]). Women less often had a history of peripheral vascular disease (9% versus 14%) and less often were current smokers than men (18% versus 32%). Patients with a higher level of SES measured by education were generally younger (secondary school graduation, mean years 69 [SD 13] versus higher secondary school degree mean years 65 [SD 1]) when they had strokes and had less severe strokes than patients with a less favorable SES (no basic education, median NIHSS 5 [IQR 2 to 7] versus university graduates, median NIHSS 3 [IQR 1 to 6]). Patients with a higher SES less often had comorbidities and risk factors like previous stroke, a history of ischemic heart disease, and peripheral vascular disease, and less often were current smokers.
Mortality after stroke according to SES has been described elsewhere15; there was no statistical significant influence of gender on 6-month mortality (data not shown).
Women were less likely to receive antiplatelet agents within the first 48 hours after admission (OR: 0.59, 95% CI: 0.53 to 0.89; Table 2).
There was no association between SES, gender, and the administration of thrombolysis (Table 2).
Women were more likely to have their blood glucose measured on admission than men (OR: 1.52, 95% CI: 1.1 to 2.1; Table 2). Body temperature was more likely to be measured in patients with a lower level of education than in patients with a university degree (patients with technical training OR: 1.54, 95% CI: 0.98 to 2.43; patients with a higher secondary school degree OR: 2.79, 95% CI: 1.41 to 5.54).
Diagnosis During Hospital Stay
With higher SES there was a significant trend toward a higher rate of echocardiographies during hospital stay (Table 3). Compared to university graduates patients with no basic education had an OR of 0.36 (95% CI: 0.24 to 0.55), secondary school graduates had an OR of 0.45 (95% CI: 0.25 to 0.79), and patients with a technical training had an OR of 0.49 (95% CI: 0.33 to 0.72) to have a TTE during hospital stay.
Compared to university graduates, patients with no basic education had an OR 0.52 (95% CI: 0.32 to 0.83), and secondary school graduates had an OR of 0.47 (95% CI: 0.32 to 0.83) to have a TEE during hospital stay.
There was no significant association between gender and the rate of echocardiographies during hospital stay.
Compared to university graduates, patients with a higher secondary school degree had an OR of 1.29 (95% CI: 1.06 to 1.57) to undergo in-hospital speech therapy in the rehabilitation phase (Table 3). There was no other association between SES, gender, and the frequency of physiotherapy, occupational therapy, or speech therapy.
Medication on Discharge
We could not find an association between SES, gender, and the rate of prescriptions for oral anticoagulation or antiplatelet agents, antihypertensives, or lipid lowering drugs at discharge (Table 4).
Women had an OR of 1.96 (95% CI: 1.48 to 2.59) to receive a prescription for antidepressants at discharge. There was no influence of SES on the prescription of antidepressants at discharge.
All results of the main analysis were adjusted for multiple risk factors.
From our population-based data we could demonstrate that SES and gender influence some aspects of acute, intermediate, and long-term/prophylactic stroke care.
In acute stroke care we observed that women were almost half as likely to receive aspirin or other antiplatelet agents. Men received blood glucose measurements nearly half as often as women. No influences of SES and gender on the administration of thrombolysis or other acute medication could be observed.
There was a consistent trend toward a lower number of TTEs and TEEs in the groups with a lower education status compared to university graduates. Patients with a secondary school degree were more likely to receive speech therapy than university graduates.
Women were nearly twice as likely to receive antidepressants when they are discharged from the hospital but no other gender differences in the prescription on medication at discharge could be found.
The most noticeable result of our study is that in the acute phase of stroke women are only half as likely to receive antithrombotic therapy as men. To our knowledge this finding has not been described in previous stroke studies, and also studies in myocardial infarction report either no difference or only a marginally lower administration of aspirin on admission.16,17 We adjusted our model for the most obvious confounders like comorbidities and age, but the effect might be attributable to some residual confounding. Treating physicians may perceive women at lower cardiovascular risk than men, which may have resulted in a lower proportion of antithrombotic therapy. There was no gender difference in the proportion of patients who had any kind of anticoagulation therapy before stroke (data not shown).
A motivating result is that SES and gender do not influence the administration of thrombolysis. To our knowledge this is the first study reporting on the influence of SES on the administration of thrombolysis in acute stroke. For gender this finding is confirmed in the studies of Kapral (2005) et al, and in a smaller sample Müller-Nordhorn could also find no adjusted influence of gender on the rate of thrombolysis.6,18
Women were more likely to receive blood glucose measurements on presentation with ischemic stroke. This finding only partially parallels in the literature. A previous stroke study reports that women more often receive hypoglycaemic drugs,18 but there is no indication by the authors that this is attributable to a higher proportion of blood glucose measurement among women. Authors argue that physicians in their hospital were aware of diabetes being a strong risk factor for cardiovascular disease in women and therefore were more sensitive to elevated glucose levels in women. We do not know whether this was the case in our participating hospitals, as there was no difference in the diabetes prevalence between women and men (73% versus 75%, P=0.2).
The next interesting finding is the consistent trend toward a lower rate of TTEs and TEEs during hospital stay with a lower level of education. We could not find similar results in previous studies, but we found investigations showing that patients with lower SES do not receive the same diagnostics than patients with a higher SES. Jakovljevic and colleagues reported that patients with a higher income were more likely to be examined by a specialist in neurology, to receive a CT or MRI scan, whereas the low-income group was more often examined with a lumbar puncture.19 These comparisons were only age standardized. In a univariate analysis McKewitt found a weak association of occupational class with admission to hospital and stroke unit but, conversely to Jakovljevic, no socioeconomic influence on the proportion of CT/MRI scans.20 The discrepancy may be explained by the fact that the decision for or against echocardiography is rather a subjective decision than guideline-driven. Furthermore, atrial fibrillation may lie on the causal pathway between SES and subsequent echocardiography. Therefore, multivariable adjusting would in inappropriate.
Apart from a higher proportion of speech therapy for the second highest education group compared to the highest, there seems to be no major influence of SES on rehabilitation. Other studies presented differing results. Kapral and colleagues found that patients in the lowest income quintiles were less likely than those in the highest income quintile to receive physiotherapy (58% versus 61%, P<0.001), occupational therapy (36% versus 47%, P=0.001), and speech–language therapy (21% versus 28%, P=0.001).21 This comparison was without adjustment for confounders. In our univariate analysis there was no association between occupational therapy and SES and only a borderline significant association between a higher number of in-hospital physiotherapy in the lower educations groups (data not shown). The weak association between physiotherapy and SES disappeared after adjustment for confounders, whereas the association for speech therapy remained after adjustment. It seems that the decision to prescribe in-hospital physiotherapy in our study at least partly depends on other factors like age, stroke severity, and some comorbidities than on gender and SES. The results of McKewitt et al support our conclusion20: after adjustment for stroke severity there was no influence of SES and gender on physiotherapy/occupational therapy, but patients with manual professions more often received speech therapy than those with nonmanual professions (OR: 0.43, 95% CI: 0.21 to 0.88). In an exploratory attempt we looked at the rates of aphasias and dysarthrias (NIHSS item) on admission with groups of education and could not find a statistical significant difference (data not shown).
At discharge there seemed to be no major influence of gender or SES on the prescription of prophylactic medication. This finding contrasts the results of Simpson et al who found that women received slightly less often oral anticoagulation or antiplatelets (OR: 0.82, 95% CI: 0.74 to 0.90) and statins at discharge (OR: 0.84, 95% CI: 0.75 to 0.94), as well as the results by Glader et al who found that women less often received antithrombotic agents for the primary and secondary prevention of stroke (OR: 0.75, 95% CI: 0.69 to 0.80; OR: 0.86, 95% CI: 0.76 to 0.98).22
In our database women received antidepressants almost twice as often as men. Previous studies have reported that the prevalence of poststroke depression was around 12% for men and 16% for women.23,24 Assuming a similar percentage of poststroke depression for our patients there could be the possibility that men are under treated. This is clinically important as antidepressants seem to be effective for the majority of patients suffering from poststroke depression.25 A limiting factor is that we do not know how many patients were on antidepressant therapy before stroke. This number is probably small: a recent study has reported that about 6% of patients take antidepressants.26
Information of education was lacking in 14% of our patients. Patients with missing data on education were more often male (54% versus 46%, P=0.003), were older (mean 66 years, STD 14, versus mean 70 years, STD 13, P<0.01), they had more severe strokes according to the NIHSS (median 4 (IQR 2 to 7) versus median 7 (IQR 3 to 16), P<0.001). It may well be that information on education is lacking because patients were too severely ill. We repeated all analyses treating patients with missing information on education as separate educational group and compared them with the primary analyses. For the end point “Platelet aggregation inhibitors” we found a change of the odds ratio of for the variables of interest of a maximum of 6% (eg, from OR 1.6 to OR 1.5). For all other variables the point estimate did not change for more than 2%. This we conclude that there is no relevant influence from missing values.
Looking at the influence of included comorbidities on the end points it seems that there is a high level of data robustness (for example strong influence of stroke severity on rehabilitation or history of hypertension on the prescription of antihypertensives). In contrast to other previous publications of the Vienna Stroke Registry15 we have chosen education as the sole measure of SES for the reasons stated above (robustness against changes over time, against gender differences through marriage and motherhood, incorporates knowledge about stroke). On the other hand it may not be comparable of the structures of the labor market in other countries. We have refrained from investigating other measures of SES as this would have resulted in a multiplication of the existing 13 multivariate models, a volume which would have gone beyond the limits of this article. We also avoided introducing composite measures of SES as they are often hard to interpret. The possibility remains that other measures of SES show a stronger association with the diagnostic and therapeutic measures of stroke. As mentioned above, education to some degree incorporates knowledge. In our analysis the only end point that could be influenced by knowledge about stroke symptoms was thrombolysis. However, the effect on thrombolysis and the time interval between symptom onset and admission to the hospital was not influenced by education or gender (data not shown).
One of the major strengths of this study is the detailed individual patient data which allows a comparison of how different properties of SES influence the decision to stroke diagnostics and therapy. However, we cannot guarantee that all stroke patients in Vienna could be documented without gaps as the documentation is confined to municipal hospitals. This number might be small as municipal beds account for 80% of all beds, and patients with acute vascular events are preferably brought to municipal hospitals by the ambulance service. However, external validity refers rather to stroke centers than the stroke patient.
Although our end points were predefined and based on the guidelines of the American Heart association, we ended up with 13 models. Within each model multiplicity is handled as they were multivariate models, but between the models multiplicity may matter. However, the main conclusion is based on a “nondifference,” so multiple testing should not matter to a considerable extent. For the end point “speech therapy” we saw the result not being robust, so a type I error may be an issue.
We found several interactions in our regression models. To investigate the nature of the interactions we stratified the analysis to the variable interacting with the risk factors of interest and repeated the multivariate analyses. In the different strata most of the point estimates for gender and SES lost statistical significance, most likely because of small cell sizes. For variables that showed an interaction with “gender” there was no major change in the point estimates in the stratified models. For variables that were interacting with SES the effect remained positive in all stratified models. Only in the model stratified to “hypertension,” in the group of patients without hypertension, patients with technical training had a trend toward an OR of 0.62 (95% CI: 0.32 to 1.2) which suggests that the effect in this subgroup was reversed, however the result was not statistically significant.
Socioeconomic status and gender influence some diagnostic and therapeutic measures of stroke care. A more detailed analysis, focusing on the end point variables of concern, could help to confirm these findings and elucidate reasons.
- Received October 3, 2007.
- Revision received November 15, 2007.
- Accepted November 21, 2007.
Carrabba N, Santoro GM, Balzi D, Barchielli A, Marchionni N, Fabiani P, Landini C, Scarti L, Santoro G, Valente S, Verdiani V, Buiatti E. In-hospital management and outcome in women with acute myocardial infarction (data from the AMI-Florence Registry). Am J Cardiol. 2004; 94: 1118–1123.
World Report on Violence and Health. Geneva: World Health Organization; 2002.
Kapral MK, Fang J, Hill MD, Silver F, Richards J, Jaigobin C, Cheung AM. Sex differences in stroke care and outcomes: results from the Registry of the Canadian Stroke Network. Stroke. 2005; 36: 809–814.
Smith MA, Lisabeth LD, Brown DL, Morgenstern LB. Gender comparisons of diagnostic evaluation for ischemic stroke patients. Neurology. 2005; 65: 855–858.
World_Bank. Global Economic Prospects and the Developing Countries 2001. Washington D.C.: World Bank; 2001.
World_Health_Organization. World Health Report 2000. Geneva: WHO; 2000.
Adams HP Jr, del Zoppo G, Alberts MJ, Bhatt DL, Brass L, Furlan A, Grubb RL, Higashida RT, Jauch EC, Kidwell C, Lyden PD, Morgenstern LB, Qureshi AI, Rosenwasser RH, Scott PA, Wijdicks EF. Guidelines for the early management of adults with ischemic stroke: a guideline from the American Heart Association/American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: the American Academy of Neurology affirms the value of this guideline as an educational tool for neurologists. Stroke. 2007; 38: 1655–1711.
Sacco RL, Adams R, Albers G, Alberts MJ, Benavente O, Furie K, Goldstein LB, Gorelick P, Halperin J, Harbaugh R, Johnston SC, Katzan I, Kelly-Hayes M, Kenton EJ, Marks M, Schwamm LH, Tomsick T. Guidelines for prevention of stroke in patients with ischemic stroke or transient ischemic attack: a statement for healthcare professionals from the American Heart Association/American Stroke Association Council on Stroke: co-sponsored by the Council on Cardiovascular Radiology and Intervention: the American Academy of Neurology affirms the value of this guideline. Stroke. 2006; 37: 577–617.
Arrich J, Lalouschek W, Müllner M. Influence of socioeconomic status on mortality after stroke: retrospective cohort study. Stroke. 2005; 36: 310–314.
Jakovljevic D, Sarti C, Sivenius J, Torppa J, Mahonen M, Immonen-Raiha P, Kaarsalo E, Alhainen K, Kuulasmaa K, Tuomilehto J, Puska P, Salomaa V. Socioeconomic status and ischemic stroke: The FINMONICA Stroke Register. Stroke. 2001; 32: 1492–1498.
McKevitt C, Coshall C, Tilling K, Wolfe C. Are there inequalities in the provision of stroke care? Analysis of an inner-city stroke register. Stroke. 2005; 36: 315–320.
Kapral MK, Wang H, Mamdani M, Tu JV. Effect of socioeconomic status on treatment and mortality after stroke. Stroke. 2002; 33: 268–273.
Glader EL, Stegmayr B, Norrving B, Terent A, Hulter-Asberg K, Wester PO, Asplund K. Sex differences in management and outcome after stroke: a Swedish national perspective. Stroke. 2003; 34: 1970–1975.
Eriksson M, Asplund K, Glader EL, Norrving B, Stegmayr B, Terent A, Asberg KH, Wester PO. Self-reported depression and use of antidepressants after stroke: a national survey. Stroke. 2004; 35: 936–941.
Paul SL, Dewey HM, Sturm JW, Macdonell RA, Thrift AG. Prevalence of depression and use of antidepressant medication at 5-years poststroke in the North East Melbourne Stroke Incidence Study. Stroke. 2006; 37: 2854–2855.
Salaycik KJ, Kelly-Hayes M, Beiser A, Nguyen AH, Brady SM, Kase CS, Wolf PA. Depressive symptoms and risk of stroke: the Framingham Study. Stroke. 2007; 38: 16–21.