Article Text

Download PDFPDF

Original article
Gender inequalities in cardiovascular risk factor assessment and management in primary healthcare
  1. Karice K Hyun1,
  2. Julie Redfern1,
  3. Anushka Patel1,
  4. David Peiris1,
  5. David Brieger2,
  6. David Sullivan3,
  7. Mark Harris4,
  8. Tim Usherwood1,5,
  9. Stephen MacMahon1,6,
  10. Marilyn Lyford1,7,
  11. Mark Woodward1,6
  1. 1The George Institute for Global Health, Sydney Medical School, University of Sydney, Sydney, Australia
  2. 2Department of Cardiology, Concord Hospital, University of Sydney, Sydney, Australia
  3. 3Department of Chemical Pathology, Royal Prince Alfred Hospital, Sydney, Australia
  4. 4Centre for Primary Health Care and Equity, University of New South Wales, Sydney, Australia
  5. 5Sydney Medical School Westmead, University of Sydney, Sydney, Australia
  6. 6The George Institute for Global Health, University of Oxford, Oxford, UK
  7. 7WA Centre for Rural Health, University of Western Australia, Perth, Australia
  1. Correspondence to Karice K Hyun, The George Institute for Global Health, Sydney Medical School, University of Sydney, Level 10, King George V Building, 83-117 Missenden Rd, Camperdown, NSW 2050, Australia; khyun{at}georgeinstitute.org.au

Abstract

Objectives To quantify contemporary differences in cardiovascular disease (CVD) risk factor assessment and management between women and men in Australian primary healthcare services.

Methods Records of routinely attending patients were sampled from 60 Australian primary healthcare services in 2012 for the Treatment of Cardiovascular Risk using Electronic Decision Support study. Multivariable logistic regression models were used to compare the rate of CVD risk factor assessment and recommended medication prescriptions, by gender.

Results Of 53 085 patients, 58% were female. Adjusting for demographic and clinical characteristics, women were less likely to have sufficient risk factors measured for CVD risk assessment (OR (95% CI): 0.88 (0.81 to 0.96)). Among 13 294 patients (47% women) in the CVD/high CVD risk subgroup, the adjusted odds of prescription of guideline-recommended medications were greater for women than men: 1.12 (1.01 to 1.23). However, there was heterogeneity by age (p <0.001), women in the CVD/high CVD risk subgroup aged 35–54 years were less likely to be prescribed the medications (0.63 (0.52 to 0.77)), and women in the CVD/high CVD risk subgroup aged ≥65 years were more likely to be prescribed the medications (1.34 (1.17 to 1.54)) than their male counterparts.

Conclusions Women attending primary healthcare services in Australia were less likely than men to have risk factors measured and recorded such that absolute CVD risk can be assessed. For those with, or at high risk of, CVD, the prescription of appropriate preventive medications was more frequent in older women, but less frequent in younger women, compared with their male counterparts.

Trial registration number 12611000478910, Pre-results.

Statistics from Altmetric.com

Introduction

Despite decreasing mortality rates due to cardiovascular disease (CVD) in many countries, it remains the leading cause of death worldwide for both women and men.1 ,2 Previously, CVD was assumed to be more prevalent in men and therefore women tended to be less intensively treated.3 To close the treatment gap between women and men, the improvement of cardiovascular health in women has been promoted through health initiatives and research.4 ,5 Yet, in Australia, as in the United States and the UK,6 ,7 women have a higher number of cardiovascular deaths per year than men (23 755 vs 21 867 deaths in 2012),1 largely because they live longer. Women with diabetes have over 40% greater excess risk of coronary heart disease (CHD)8 and nearly 30% higher relative risk for stroke compared with men with diabetes.9 More research is needed to uncover the reasons for these female disadvantages. One possibility is that women are less often recognised as being prone to CVD than men and are thus less likely to receive a timely diagnosis and to receive appropriate treatment after a positive diagnosis.

There is evidence that risk assessments and medication adherence have a positive impact on outcomes.10 ,11 Studies from countries outside Australia have found that women with CHD are less likely than men to undergo risk factor assessments in primary healthcare.12 ,13 Some studies have also shown that women with CHD are less likely to receive recommended medications than men,13 ,14 although other studies have shown no gender differences.15

While differences in CVD assessment and management between women and men have been observed in other countries, the extent to which this may be an issue in Australian primary healthcare is unknown. We aimed to determine whether measurement of CVD risk factors and guidelines recommended medication prescriptions varied between women and men in a large Australian primary healthcare cohort.

Methods

Study design and data

Baseline data from the Treatment of Cardiovascular Risk using Electronic Decision Support (TORPEDO) study were used for analysis. TORPEDO was a cluster randomised trial to test whether a computer-guided quality improvement intervention improved CVD risk management when compared with usual care. Full details of the study and the primary results of the trial have been published.16 Although the data were extracted for a cluster randomised trial, only the baseline data were used in this observational study. Briefly, Australian general practices (GPs) and Aboriginal Community Controlled Health Services (ACCHSs) were eligible to participate if there was exclusive use of an electronic health record that was compliant with study software to record risk factor information, pathology test results and prescribe medications.

The total TORPEDO cohort included Aboriginal and Torres Strait Islander people aged ≥35 years and others aged ≥45 years (no upper age limit) from 40 GPs and 20 ACCHSs across New South Wales and Queensland. These age criteria were based on Australian guideline recommendations for conducting a CVD risk assessment.10 To be eligible for inclusion, patients had to be regular attendees of the service. The definition of a regular attendee was based on Australian GP standards and includes at least three attendances in the previous 24 months and at least one attendance in the previous 6 months. The average cluster size was 750 patients and 30% of the patients were at high CVD risk. CVD was defined as a recorded diagnosis of CHD, ischaemic stroke or peripheral vascular disease. High CVD risk was defined in a manner consistent with the definition in Australian guidelines where patients with any of the following are considered high risk of CVD: diabetes mellitus and age >60 years, diabetes mellitus and albuminuria, chronic kidney disease (stage 3B or worse) or extreme individual risk factor elevations: systolic blood pressure (SBP) ≥180 mm Hg, diastolic BP ≥110 mm Hg, total cholesterol >7.5 mmol (290 mg/dL) or a calculated 5-year CVD risk of >15% (based on the Framingham Risk equation).10 ,17

Deidentified data were extracted between September 2011 and May 2012 for all patients who met these criteria. Data extraction was performed using a validated extraction tool at randomisation.18 The study was approved by the University of Sydney Human Research Ethics Committee and the Aboriginal Health and Medical Research Council of New South Wales Human Research Ethics Committee. Individual consent waiver was granted, given that data collection was based on deidentified extracts from the electronic health record system.

Outcomes

Two outcomes were analysed. First, for patients in the total cohort, we analysed the rate of having recorded risk factor information sufficient for guideline-recommended absolute CVD risk assessment. Second, in the CVD/high CVD risk subgroup, we analysed the rate of optimal guideline-recommended preventive medication prescriptions. Based on the components of the Framingham risk score, sufficient assessment of risk factors for CVD risk assessment was defined as having recorded smoking status at least once, SBP in the previous 12 months, total cholesterol and high-density lipoprotein (HDL) cholesterol in the previous 24 months.17 Optimal prescriptions of recommended medication were defined as a current prescription for at least one BP-lowering drug and a statin for people at high risk without established CVD, or at least one BP-lowering drug, a statin and an antiplatelet agent (unless contraindicated by oral anticoagulant use) for people with a recorded diagnosis of CVD.10

Statistical analyses

Missing data were not imputed for these analyses. Assessment of individual risk factors and prescription of individual medications were defined as having a record; if the value was missing, this was considered as having had no assessment or prescription. Baseline characteristics and prescription of medications to women and men were compared using χ2 tests for categorical variables and t-tests for continuous variables. Assessment of individual risk factors was compared between gender and age groups using χ2 tests. The definition for sufficient assessment of CVD risk factors was as follows: having recorded smoking status at least once, SBP in the previous 12 months, total cholesterol and HDL cholesterol in the previous 24 months.17 Although body mass index (BMI) and fasting glucose are not included in the Framingham risk score, and therefore not included in the definition for sufficient assessment of CVD risk factors, since they are also important risk factors for CVD, we analysed these variables also. Multiple-adjusted generalised estimating equation (to account for the clustering by healthcare provider) logistic regression models with an exchangeable working correlation matrix were used for each of the outcomes to estimate the ORs and the corresponding 95% CIs for women versus men. For both outcomes, independent predictors included in the models were as follows: gender (women vs men), age groups (35–54 years, 55–64 years, ≥65 years), Aboriginal/Torres Strait Islander status, diabetes status, overweight/obese (BMI ≥25 kg/m2 vs not), high BP (SBP ≥140 mmHg/diastolic BP ≥90 mm Hg vs not), high total cholesterol (total cholesterol ≥5.5 mmol vs not), low HDL cholesterol (HDL cholesterol ≤1 mmol vs not) and smoking status (current smokers vs ex/never smokers). Moreover, the ORs and the corresponding 95% CIs for women versus men were derived for individual CVD risk factor assessments separately using the same model, but excluding covariates that are related to the outcomes. For example, if the odds of smoking assessment were being analysed, smoking status was excluded from the list of covariates. Similarly, the ORs and the corresponding 95% CIs for women versus men were derived for individual recommended medication prescriptions. The model predicting the measurement and recording of CVD risk factors was performed for the total cohort, while the analysis predicting the receipt of appropriate CVD medications was restricted to those in the CVD/high CVD risk subgroup. For each outcome, predefined subgroup analyses were performed by adding an interaction term between gender and the CVD risk factors (age groups, indigenous status, overweight/obese, high BP, high total cholesterol, low HDL cholesterol and current smoker status) to the multiple-adjusted logistic regression model. The covariates and the restrictions for the analyses of each outcome remained the same as above. The subgroup analyses for Aboriginal/Torres Strait Islander status were performed for those who were 45 or older only to keep the age groups consistent. Furthermore, the interaction between gender and age groups for the two outcomes was also explored by Aboriginal/Torres Strait Islander status, using the same model. To explore the interaction between gender and age groups on prescription of each class of medication (BP medication, statin and antiplatelet), multiple-adjusted generalised estimating equation logistic regression models were used. The covariates for these analyses were as listed above. Interaction was addressed by fitting age group as a continuous variable to produce a test for trend.

Data were analysed using SAS V.9.4 (SAS Institute, Cary, North Carolina, USA).

Results

Data on 53 085 patients were extracted from 60 services at baseline; 30 601 (58%) patients were women. Of the total cohort, 13 294 (25%) were in the CVD/high CVD risk subgroup (6621 patients with prevalent CVD and 6673 at high CVD risk), 47% (n=6202) of whom were women.

Gender differences in the overall cohort

The mean age was 61 years for women and 60 years for men. Women had lower rates of recorded diabetes and of being a current smoker. Women also had a lower mean SBP and a higher mean total cholesterol (table 1).

Table 1

Baseline comparison by gender

Overall, only 43.3% of patients had all necessary CVD risk factors recorded for absolute risk assessment. Compared with men, women were less likely to have the CVD risk factors recorded (41.4% vs 45.9%, p<0.001). Women were significantly less likely to have smoking status, SBP, total cholesterol, HDL cholesterol and BMI recorded compared with men (table 2). Also, when the rates of risk factor assessment were compared within age groups, where there were significant differences between women and men, women were consistently less likely to have their risk measured than men (table 2).

Table 2

Assessment of individual risk factors by gender and by age groups

After adjustment, the odds of women having all necessary risk factors recorded for an absolute risk assessment was 12% lower compared with men (OR (95% CI): 0.88 (0.81 to 0.96)) (figure 1). Taking the individual risk factors separately, the odds of having smoking status recorded was 22% lower (0.78 (0.66–0.91)), SBP recorded was 12% lower (0.88 (0.81–0.96)) and cholesterol (total and/or HDL cholesterol) recorded was 8% lower (0.92 (0.86–0.98)) in women than men. The only significant subgroup heterogeneity observed was in relation to total cholesterol (pinteraction=0.02); however, no gender difference was observed among individuals with high total cholesterol (figure 1). As well as a lack of evidence for heterogeneity in the total cohort, no heterogeneity was found between age groups for both Aboriginal/Torres Strait Islanders and others (pinteraction=0.49 and 0.69, respectively). For age groups 35–54, 55–64 and 65 and older, the ORs (95% CIs) for all Aboriginal/Torres Strait Island people were 0.75 (0.63 to 0.89), 0.72 (0.51 to 1.02) and 1.05 (0.64 to 1.72), and the corresponding results for others were 0.92 (0.81 to 1.05), 0.87 (0.76 to 0.99) and 0.94 (0.83 to 1.05).

Figure 1

Multiple-adjusted female to male ORs and 95% CIs for the assessment of cardiovascular risk factors. HDL, high-density lipoprotein.Indigenous: Aboriginal and/or Torres Strait Islander; overweight/obese: body mass index ≥25 kg/m2; high blood pressure: systolic blood pressure ≥140 mm Hg and diastolic blood pressure ≥90 mm Hg; high total cholesterol: total cholesterol ≥5.5 mmol and low HDL cholesterol: HDL cholesterol ≤1 mmol. Each variable in the figure is adjusted for all the others.

Gender differences in patients in the CVD/high CVD risk subgroup

Of the 13 294 patients found to be in the CVD/high CVD risk subgroup, the mean age of women was 70 years and of men was 68 years (table 1). The proportion of women and men who were prescribed recommended medications were similar (46.9% vs 48.0%, p=0.20) but low overall, where only 47.5% of the patients in the CVD/high CVD risk subgroup received the prescriptions. When prescription of each class of medication was compared, women were less likely to be prescribed statin and antiplatelet, but slightly more likely to be prescribed BP-lowering medication compared with men (table 3). Women in the youngest (35–54 years) age group were substantially less likely to be prescribed BP-lowering medication, statin and antiplatelet than men in the same age group (table 3).

Table 3

Prescription of individual classes of medications to patients in the CVD/high CVD risk subgroup by gender and age group

After adjusting for covariates, the odds of women receiving recommended medication were greater by 12% than for men (OR (95% CI): 1.12 (1.01 to 1.23)) (figure 2). Significant heterogeneity was found between age groups (p for interaction <0.001), where younger women were substantially less likely to be prescribed recommended medications by 37% (0.63 (0.52 to 0.77)) and older women were 34% more likely to be prescribed medications (1.34 (1.17 to 1.54)) compared with men (figure 2). Moreover, both Aboriginal/Torres Strait Island people and others showed lower odds of being prescribed recommended medication in younger women than younger men. For age groups 35–54, 55–64 and 65 and older, the ORs (95% CIs) for all Aboriginal/Torres Strait Island people were 0.65 (0.53 to 0.80), 1.30 (1.00 to 1.69) and 1.23 (0.93 to 1.62) (pinteraction=0.01), and the corresponding results for others were 0.60 (0.37 to 0.96), 0.91 (0.76 to 1.09) and 1.36 (1.18 to 1.57) (pinteraction=0.004), respectively.

Figure 2

Multiple-adjusted female to male ORs and 95% CIs for guideline-recommended medication prescription among the patients in the cardiovascular disease/high cardiovascular risk subgroup. HDL, high-density lipoprotein.Indigenous: Aboriginal and/or Torres Strait Islander; overweight/obese: body mass index ≥25 kg/m2; high blood pressure: systolic blood pressure ≥140 mm Hg and diastolic blood pressure ≥90 mm Hg; high total cholesterol: total cholesterol ≥5.5 mmol; and low HDL cholesterol: HDL cholesterol ≤1 mmol. Each variable in the figure is adjusted for all the others.

When examining individual classes of medication, women aged 35–54 years were less likely to be prescribed BP medication, statin and antiplatelet than their male peers (0.69 (0.54 to 0.87), 0.65 (0.49 to 0.87) and 0.73 (0.56 to 0.96) respectively) (figure 3). Furthermore, a positive trend of the odds of women receiving each medication compared with men by age group was observed for three of the four medication classes. Most especially, women in the youngest age group were less likely to be prescribed BP-lowering medication and statin, whereas women in the oldest age group were more likely to be prescribed BP-lowering medication and statin compared with their male counterparts. Overall, prescription of recommended medications to overweight/obese women and men were comparable; however, of those who were not overweight/obese, women had 36% higher odds of being prescribed recommended medications than men (pinteraction=0.02) (figure 2).

Figure 3

Multiple-adjusted female to male ORs and 95% CIs for the prescription of blood pressure-lowering medication, statin, antiplatelet and anticoagulant therapy among the patients in the cardiovascular disease/high cardiovascular risk subgroup, by age group. Each variable in the figure is adjusted for age groups, Aboriginal/Torres Strait Islander status, diabetes status, overweight/obese, high blood pressure, high total cholesterol, low high-density lipoprotein cholesterol and smoking status.

Discussion

This study provides contemporary gender comparisons of CVD risk assessment and management provided by a sample of primary healthcare services in Australia. Our results show that overall proportion of patients who were assessed and treated were notably low, and women were less likely to have CVD risk factors measured than men. Although there was no evidence of heterogeneity by age in risk factor measurement between the genders, there was such heterogeneity in treatment for patients in the CVD/high CVD risk subgroup. Our finding suggested that older women in the CVD/high CVD risk subgroup overall had greater odds of being prescribed recommended medications than men, while the converse was true for their younger counterparts.

Despite guidelines, including those of The Royal Australian College of General Practitioners and the National Vascular Disease Prevention Alliance, regarding screening for CVD risk factors,10 ,19 ,20 inequality in CVD prevention through assessing risk factors remains between women and men. Our results are similar to those from a study from the UK where the likelihood of assessing smoking status, cholesterol, BP and BMI was 35% higher in men than in women.13 Another UK study has found that men tended to have higher rates of comprehensive risk factor assessment than women (22% vs 20%), although this difference was not statistically significant.12 This gap may occur due to barriers at individual, social and system levels.21 At individual and social levels, physicians may not be aware or familiar with the updated guidelines, or physicians and women may have the old misconception of CVD being a man's disease.22 Furthermore, the misconception of senior physicians may have been passed on to the younger generation of physicians; therefore, where there are financial disincentives and time and resource constraints, women have been disadvantaged in receiving appropriate CVD risk factor assessment.

There are a number of study limitations that are noteworthy. First, the services recruited for the TORPEDO study did not represent a random sample of primary healthcare services in Australia. All of the GPs were located in urban areas of New South Wales and Queensland, whereas about 30% of Australian GPs are in rural or remote areas. Furthermore, several services were teaching GPs and this may reflect a stronger commitment to improving healthcare quality compared with non-teaching practices. It is possible, therefore, that the gaps in care practices encountered in this study may potentially be even larger in the broader primary healthcare sector. Second, the study population for the TORPEDO study was restricted to regular attendees of the primary healthcare service. Regular attendees are commonly used in the denominator for quality improvement indicators and for the purposes of the randomised trial it was important to define a relatively stable population to test the effect of the intervention. However, this means that we are unable to extrapolate the findings to infrequent attenders who may exhibit different demographic, health and healthcare seeking characteristics. Third, we were unable to capture contraindications to recommended medications (eg, pregnancy, high bleeding risk and allergies or medication intolerances). This could explain a small proportion of the treatment gaps encountered.

To overcome the misconceptions and barriers that may prevent gender equality in assessment and management of CVD, some doctors, researchers and organisations from across the world have made significant efforts to promote women's CVD health. Studies regarding CVD management for women in the past decade have informed educational campaigns for greater attention to women's cardiovascular health.23 Programmes such as Go Red for Women and The Women's Room are examples of these initiatives.4 ,5 Despite such initiatives, our study found that younger women with CVD/high risk of CVD between the ages of 35 and 54 were significantly less often prescribed recommended medications than younger men with CVD/high risk of CVD. This is consistent with previous studies reported from other countries.13 ,24 In contrast, older women with CVD/high risk of CVD were more likely to be prescribed than older men with CVD/high risk of CVD, suggesting possible presence of age stereotypes in medical management of CVD. This indicates that strategies and incentives are needed at the system level, together with education of physicians and public about this inequality at the individual and social level, to minimise the treatment gap between women and men. For example, academic programmes such as those of NPS MedicineWise can promote better understanding and improved prescribing of medications;25 health quality identification system such as gender-specific quality indicators may be incorporated into quality improvement programmes to support better assessment and treatment of patients and targeted funding model such as specific Medicare Benefits Schedule reimbursement for patients with CVD/high risk of CVD can provide financial incentives for physicians.

Conclusion

Despite guidelines recommending that risk factors should be assessed and managed in both sexes, inequalities remain. Women are not receiving equitable testing in Australian primary healthcare and, while women in general receive appropriate medications more often than do men, younger women with CVD/high risk of CVD are disadvantaged in terms of receipt of essential treatments to prevent CVD events. System-wide solutions to these inequalities and the increase of the overall rates of assessment and management of CVD are needed.

Key messages

What is already known on this subject?

  • Risk assessment and medication adherence have a positive impact on preventing and managing cardiovascular disease (CVD); however, differences in CVD assessment and management between women and men have been observed in countries such as the UK.

What might this study add?

  • In Australian primary healthcare services, women were less likely to be assessed for CVD risk factors at primary healthcare services. Of those at high risk of CVD, younger women (35–54 years) were less likely to be prescribed guideline-recommended medications than younger men (35–54 years), whereas older women (≥65 years) were more likely to be prescribed guideline-recommended medications than their counterparts.

How might this impact on clinical practice?

  • System-level strategies are needed to improve the provision of CVD assessment and treatment to minimise the gap between women and men.

Acknowledgments

The authors acknowledge the support of the general practices and Aboriginal Community Controlled Health Services involved in this study. They also acknowledge the Queensland Aboriginal and Islander Health Council, NSW Aboriginal Health and Medical Research Council, WentWest and Central and Eastern Sydney Primary Health Networks for supporting the TORPEDO study.

References

View Abstract

Footnotes

  • Contributors KKH, MW, JR and DP participated in the design of the study. KKH performed the statistical analysis. KKH, MW, JR, DP and AP participated in the interpretation of the results. KKH drafted and all authors reviewed the manuscript. All authors read and approved the final manuscript.

  • Funding The National Health and Medical Research Council of Australia and the New South Wales Department of Health funded TORPEDO. KKH is funded by the University of Sydney Postgraduate Award. DP is supported by a National Health and Medical Research Council (NHMRC) postdoctoral fellowship (1054754). AP is supported by an NHMRC Senior Research Fellowship (632938). JR is funded by a NHMRC Career Development Fellowship (1061793) cofunded with a National Heart Foundation Future Fellowship (G160523) and MW is supported by a NHMRC Principal Research Fellowship.

  • Disclaimer Neither institutions had any role in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review or approval of the article or decision to submit the article for publication.

  • Competing interests None declared.

  • Ethics approval The study was approved by the University of Sydney Human Research Ethics Committee and the Aboriginal Health and Medical Research Council of New South Wales Human Research Ethics Committee. Individual consent waiver was granted, given that data collection was based on deidentified extracts from the electronic health record system.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles