Article Text
PDF
Abstract
Objectives To determine whether sex differences exist in the management of patients with a history of coronary heart disease (CHD) in primary care.
Methods General practice records of patients aged ≥18 years with a history of CHD in a large general practice dataset in Australia, MedicineInsight, were analysed. Sex-specific, age-standardised proportions of patients prescribed with recommended medications; assessed for cardiovascular risk factors; and achieved treatment targets according to the General Practice Management Plan were reported.
Results Records of 130 926 patients (47% women) from 438 sites were available from 2014 to 2018. Women were less likely to be prescribed with recommended medications (prescribed ≥3 medications: women 44%, men 61%; p<0.001). Younger patients, especially women aged <45 years, were substantially underprescribed (aged <45 years prescribed ≥3 medications: women 2%, men 8%; p<0.001). Lower proportions of women were assessed for cardiovascular risk factors (blood test for lipids: women 70%–76%, men 77%–81%; p<0.001). Body size was not commonly assessed (body mass index: women 59%, men 62%; p<0.001; waist: women 23%, men 25%; p<0.001). Higher proportions of women than men achieved targets for most risk factors (achieved ≥4 targets in patients assessed for all risk factors: women 82%, men 76%).
Conclusion Gaps in preventative management including prescription of indicated medications and risk factor monitoring have been reported from the late 1990s and this large-scale general practice data analysis indicate they still persist. Moreover, the gap is larger in women compared to men. We need new ways to address these gaps and the sex inequity.
- primary care
- coronary artery disease
Statistics from Altmetric.com
Introduction
Coronary heart disease (CHD) remains one of the leading causes of death and disability worldwide despite significant advances in its identification and treatment over the past several decades.1 In Australia, 156 968 adults were hospitalised with CHD in 2015–2016.2 Latest figures from the two most populated Australian states suggested that of the patients hospitalised with CHD, 34% are readmitted for cardiovascular disease (CVD) within 2 years and 8% die within 2 years.3 Primary care has been shown to be an important component for the secondary prevention of CVD. For example, after first hospitalisation with CHD, those who visited a general practitioner (GP) compared to those who did not during the 2-year follow-up period had 11% lower risk of a CVD emergency readmission. Similarly, patients who had a chronic disease management plan had 5% lower risk of readmission compared with those who did not.3 Moreover, patients on combination drug therapy following an acute myocardial infarction were at lower mortality risk than those who were not dispensed with CVD drugs within 28 days after hospital discharge.4
A survey on the use of secondary prevention drugs for CVD reported differences in the rates of drug use in patients with CHD with the highest rates in North America and Europe and the lowest rates in Africa.5 In Europe, cross-sectional surveys evaluating guideline implementations in coronary patients have been conducted since 1995–1996. Despite the increase in antihypertensive and lipid-lowering medications use over time, the latest survey with 7998 patients from 24 European countries reported only a third of the patients met the blood pressure (BP) target and a fifth of the patients met the low-density lipoprotein (LDL) cholesterol target.6 Previous primary care studies in Australia identified significant treatment gaps in patients at high CVD risk. Only half of the patients with established CVD were prescribed with a combination of major cardiovascular medications.7 Importantly, 53% prescribed with a statin and 59% prescribed an antihypertensive medication were not meeting the respective targets for LDL-cholesterol and BP. A recent study reported 43% of patients with a high absolute CVD risk were not treated with statins, while 36% on statins had low CVD risk.8 Another study reported women were less likely to have cardiovascular risk factors assessed and younger women were less likely to be prescribed with cardiovascular medications than men of the same age.9
The General Practice Management Plan (GPMP) for CHD is a tool for the management of patients with CHD in primary care.10 The recommendations are similar to the guidelines for secondary prevention of CVD in the UK and Europe.11–13 The aim of this study was to determine in a large contemporary Australian primary care dataset, whether sex disparities exist in the management of patients with a history of CHD according to GPMP.
Methods
MedicineInsight is a national general practice data programme established by NPS MedicineWise in 2011 (online supplementary methods). By October 2018, the database has included records of 2.97 million active patients from 671 general practices and 4834 GPs across Australia, wherein active patients were defined as individuals with ≥3 encounters recorded in 2 years.14 In this study, deidentified records of active patients aged ≥18 years at baseline with a history of CHD were provided by NPS MedicineWise. Patients with CHD were identified based on an algorithm developed by NPS MedicineWise, which incorporated information from related code recorded by GPs and/or free-text terms in designated data fields (online supplementary methods).14 A complete list of variable codes is provided in the MedicineInsight data book.14 Patients with missing data on sex (n=14) or had <1 year of follow-up (n=7648) were excluded. The study period included the most recent 2 years of records with the baseline date assigned as the encounter date closest to 2 years prior to the last encounter date. Therefore, the respective study period for the 82.7% of patients with last encounter date in 2018, 15.4% with last encounter date in 2017 and 2.0% with last encounter date in 2016 were 2016–2018, 2015–2017 and 2014–2016.
Supplemental material
Recommendations in GPMP that could be evaluated with MedicineInsight are listed in table 1. The full list of recommendations is available on the Heart Foundation website.10 Of the five medications recommended for use in all patients, antiplatelet agents, angiotensin-converting enzyme (ACE) inhibitors or angiotensin II receptor antagonists (ARA), beta-blockers and statins were recommended for daily use. Medications were considered as current if they were prescribed during the study period. Except for body height, measurements of risk factors recorded during the study period were considered as current. Therefore, missing data indicated non-assessment. For patients who were not flagged as having diabetes according to MedicineInsight’s algorithm,15 treatment target for diabetes was achieved if a diabetes diagnostic test (ie, oral glucose tolerance test or blood test for fasting glucose or glycated haemoglobin (HbA1c)) was performed during the study period.
Recommendations in the General Practice Management Plan for coronary heart disease that could be evaluated* with MedicineInsight
Statistical analysis
The proportions of patients prescribed with recommended medications, assessed for risk factors and achieved treatment targets were reported by sex and by sex and age groups. Sex-specific crude estimates were age standardised to the Australian CHD standard population as reported in the National Health Survey (NHS) 2014–20152 using the direct standardisation method, which calculates a weighted average of the group’s age-specific rates where the weights represent the age-specific sizes of the standard population. The standardisation was separately done by sex. Sex differences in the age-adjusted figures were evaluated using the Mantel-Haenszel χ2 test, which tested the null hypothesis that the individual age stratum odds ratios were equal to one versus the alternative hypothesis that at least one odds ratio was different from unity. For patients with risk factors assessed more than once during the study period, treatment targets were based on the last measurements and were compared with the first measurements in the 2-year period. Furthermore, the proportions achieving targets for lipids and BP were reported by relevant prescriptions of recommended medications. All analyses were performed using Stata/SE V.15.0.
Results
Records of 130 926 patients with CHD (46.7% women) from 438 sites across Australia were analysed. The proportion of women with CHD ranged from 9.7% in the age 45–54 years group to 15.3% in the <45 years group and 36.1% in the ≥75 years group (table 2). In men, the proportion with CHD ranged from 7.1% in the <45 years group to 34.5% in the ≥75 years group. Compared to men, women were less likely to be: prescribed with any antihypertensive medications or any lipid-lowering medications; current smokers; overweight or obese; or to have diabetes (table 2).
Demographics of MedicineInsight patients with a history of coronary heart disease by sex
The reported proportions that follow are either age standardised or age stratified.
CHD management according to pharmacotherapy recommendations
Women were less likely to be prescribed with any of the recommended medications than men (figure 1). Of the four medications recommended for daily use, 21.2% of women and 10.0% of men were not prescribed with any of these medications, and 21.9% of women and 33.6% of men were prescribed with all medications.
Age-standardised proportion of patients currently prescribed with recommended medications according to the General Practice Management Plan for coronary heart disease. P<0.001 unless stated in the figure. ACE-I, ACE inhibitor; ARA, angiotensin II receptor antagonists.
While statin is the recommended lipid-lowering medication, 10.6% of women and 14.5% of men prescribed with statin were also prescribed with ezetimibe. When all antihypertensive drug classes were considered, 73.1% of women and 85.9% of men were prescribed with ≥1 antihypertensive medication and 37.9% of women and 45.9% of men were prescribed with ≥3 antihypertensive medications.
In general, younger patients were less likely to be prescribed with the recommended medications and women were less likely to be prescribed with the recommended medications compared with men in the same age group (table 3). Women in younger age groups, especially those <45 years, were underprescribed.
Proportion of patients currently prescribed with recommended medications according to the General Practice Management Plan for coronary heart disease by sex and age group
Achievement of recommended treatment targets
Assessment of BP and smoking status at least once during the 2-year study occurred in >90% of patients (figure 2). Risk factors that required blood testing were assessed in ≥70% of patients. Apart from diabetes testing, risk factor testing was consistently less frequent among women (p<0.001). Similarly, among patients with diabetes, only 50.5% of women and 57.0% of men had an HbA1c test. Likewise, body size was not commonly assessed, less than two-thirds of patients had their weight, height or BMI recorded and less than a quarter of patients had their waist recorded.
Age-standardised proportion of patients assessed for risk factors included in the General Practice Management Plan for coronary heart disease in the last 2 years. P<0.001 unless stated in the figure. BMI, body mass index; HDL cholesterol, high-density lipoprotein cholesterol; LDL cholesterol, low-density lipoprotein cholesterol.
Of those who had risk factors assessed at least once during the study period, higher proportions of women than men achieved targets for HDL-cholesterol, triglycerides, HbA1c in those with diabetes, smoking and BMI (p<0.001; figure 3A). Nevertheless, less than a third of the patients who were assessed achieved targets for LDL-cholesterol, BMI or waist circumference. For patients who had all seven risk factors (BP, LDL-cholesterol, HDL-cholesterol, triglycerides, diabetes diagnostic test for patients without diabetes or HbA1c for patients with diabetes, smoking status and BMI or waist circumference where BMI was missing) assessed (women 41.0%, men 47.8%; n=58 406), 4.1% of women and 6.7% of men achieved <3 targets, while only 1.8% of women and 2.1% of men achieved all targets (figure 3B).
Age-standardised proportion of patients who achieved recommended treatment targets according to the General Practice Management Plan for coronary heart disease based on the most recent measurement in those who had risk factors assessed in the last 2 years: (A) participants who were assessed for the specific risk factors; (B) participants who were assessed for all seven risk factors (blood pressure, low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, triglycerides, diabetes diagnostic test for patients without diabetes or HbA1c for patients with diabetes, smoking status and body mass index (BMI) or waist circumference where BMI was missing; n=58 406); p<0.001 unless stated in the figure.
Overall, a higher proportion of women achieved target for BMI, HDL-cholesterol, smoking and for HbA1c in patients with diabetes than men in the same age group (table 4). A lower proportion of women compared with men in the same age group achieved targets for waist circumference and LDL-cholesterol. For BP and triglycerides, women in younger age groups and men in the oldest age group were more likely to have achieved targets compared with the opposite sex in the same age groups.
Proportion of patients who achieved recommended treatment targets according to the General Practice Management Plan for coronary heart disease based on the most recent measurement in those who had risk factors assessed at least once in the last 2 years by sex and age group
Compared to patients with risk factors assessed more than once, those who were only assessed once were more likely to have achieved targets, except for LDL-cholesterol and current smoking (table 5). For patients who had risk factors assessed more than once, the proportion who achieved targets increased by >5% for LDL-cholesterol, smoking and waist circumference in women and for BP, LDL-cholesterol, current smoking and BMI in men.
Age-standardised proportion of patients who achieved recommended treatment targets according to the General Practice Management Plan for coronary heart disease
Prescription of medications and achievement of related treatment targets
Among patients assessed for LDL-cholesterol, substantially higher proportions of patients prescribed with statins achieved target compared with patients not prescribed with statins (online supplementary table 1). Among patients assessed for HDL-cholesterol and triglycerides, the proportions who achieved targets were higher in those not prescribed than those prescribed with statins. Similarly, among patients assessed for BP, higher proportions of patients achieved target were observed in those not prescribed compared with those prescribed with recommended antihypertensive medications. Patients prescribed with relevant recommended medications were mainly much older than those not prescribed with the medications. Within each prescription and lipid risk factor subgroup, patients who did not achieve targets were generally younger than those who achieved them. For BP, patients who did not achieve target were older, especially so in women not prescribed with antihypertensive medications. Similar patterns were observed in men and women.
Discussion
Our study showed that underprescribing of recommended medications, lack of cardiovascular risk factor monitoring and failure to achieve treatment targets remained prevalent among patients with CHD. Furthermore, women continue to be underprescribed and underassessed compared with men. Small-scale studies from as far back as two decades ago illustrated assessment, prescribing and treatment gaps in patients at high risk of, or with established, CVD.7 9 16 17 Although GPMP was introduced in 2013 and preparation of a GPMP is available for Medicare rebate, the CHD management gap remains wide.
In Australia, women accounted for 44% of CHD deaths and 52% of all CVD deaths in 2015.2 Yet, the historical assumption that CHD is predominantly a condition affecting men still persists with women with a history of CHD being less likely to be managed according to guidelines compared with similarly affected men.18 Based on primary care data in 2008, the AusHEART Study reported women were less likely to be prescribed with a statin and an antiplatelet agent than men.19 Restricting our analysis to patients aged ≥55 years as in AusHEART showed higher prescriptions of statin and antihypertensive medications but lower prescription of antiplatelet agents in our cohort. The AusHEART study design may have biased towards including 322 GPs who have an interest in cardiovascular risk management who each recruited 15–20 consenting adult patients.19 MedicineInsight incorporates a much wider sample; hence, our results provide a better indication of the prescribing gaps and treatment disparities in Australia. Contrary to findings from international studies,20 younger patients were substantially underprescribed. This may be due to a misperception that younger patients are at lower risk; hence, lifestyle modification would be sufficient to achieve treatment targets.21 Moreover, some of the youngest women may have been pregnant around the study period. This may have contributed partly to the lower proportion of women aged <45 years prescribed with ACE inhibitors or ARA and statins, which should be avoided during pregnancy.22 Nevertheless, prescription of antiplatelet agents and beta-blockers were also much lower in this group compared with other age and sex subgroups. Furthermore, younger patients, particularly women, were unlikely to have been misidentified as having a history of CHD since 2.9% of patients did not have a recorded diagnosis of coronary artery disease, past myocardial infarction or past coronary artery bypass grafting and formed 0.6% of those who were never prescribed with any of the recommended medications. These 2.9% of patients were also older than the remaining cohort (mean age 73.0 vs 66.4 years).
Regular assessment of cardiovascular risk factors is important in determining whether adjustments are required in the management of a patient’s health. Similar to another Australian study, we reported assessment gaps exist with risk factors that require blood test were less likely to be assessed in comparison with BP and smoking status. Moreover, women were less likely to have risk factors assessed across all age groups.9 Although body size can be assessed easily during consultation, height, weight and especially waist are not commonly measured in primary care. A Melbourne study reported only 22% of patients had BMI recorded and 4% had waist recorded in the last 2 years; women were also less likely to have BMI recorded.23 The proportions of patients with BMI or waist recorded were higher in our study possibly because patients included have a history of CHD. Nevertheless, improvements in risk assessments are needed in primary care.
Interestingly, our cohort of women were more likely to have achieved treatment targets even though underprescribing was more prevalent than men. The over-representation of younger women may have contributed to these findings. Importantly, patients who undergo blood testing, which is normally conducted on a separate day to the GP visit, maybe more likely to be compliant with treatments (eg, adhere to medication, uptake of exercise and healthy diet). It is uncertain how many of the 20%–30% of patients without their lipids assessed during the study period met targets for lipids. For patients prescribed with statin or antihypertensive medications, less than half met the respective targets for LDL-cholesterol and BP, and treated women were less likely to have achieved targets than treated men. On closer inspection of the treated groups, patients who achieved targets were generally older than those who did not achieve these targets and the number of recommended medications prescribed increased with age. Therefore, younger patients were less aggressively managed, hence, less likely to have achieved targets.
On the positive, higher assessment frequency was seen in patients who were less likely to have achieved treatment targets. This may indicate GPs are monitoring the health of patients who they consider in need of additional care to prevent recurrent CHD or other CVD. For these patients, improvements were observed for LDL-cholesterol and smoking, although they were still less likely to have achieved targets than patients who were assessed only once. Improvement in the proportion who achieved target in this group of patients were also observed for waist circumference in women and BMI in men. Nevertheless, the proportion of patients who were considered to be at a healthy weight remained small. A UK primary care study on overweight and obese patients reported most patients did not receive a weight management intervention during the 7-year study period.24 A study on barriers to providing dietary counselling in primary care identified lack of time, patient non-compliance, inadequate teaching materials, lack of counselling training, lack of knowledge, inadequate reimbursement and low physician confidence as perceived barriers.25
The major strength of our study was the use of a large national general practice dataset in Australia. Nevertheless, a few limitations warrant mention. Patients with CHD were identified using an algorithm; however, patients with CHD are identified and used in regular feedback practice reports by the NPS MedicineWise visiting programme, so any major discrepancies (or misdiagnoses) would have been identified. Information on medication was based on prescription, not drug dispensing; we were unable to determine whether patients prescribed with recommended medications who did not achieve targets were due to ineffective treatment (ie, poor management) or medication non-adherence. Contraindication may also account for a small proportion of underprescribing. For risk factors assessed during consultation, we were unable to determine whether missing information was due to a lack of assessment or lack of documentation. We were unable to identify patients who sought healthcare from multiple clinics as linkage between general practices was unavailable. In the NHS 2014–2015, substantially more men than women self-reported having CHD and higher proportions of older participants reported having CHD in both sexes.2 The high proportion of patients in the youngest age group compared with the NHS self-reported data suggest possible non-representativeness. Moreover, our results may not be representative at a regional level especially in remote areas. The MedicineInsight data are representative of practices that have computerised records drawn from two large practice software companies covering 80% of the market.14 Crude and age-standardised figures were broadly consistent (data not shown).
Having GP visits and GPMP after hospitalisation with CHD reduce risk of CVD emergency readmission, which could potentially lead to cost savings to the health system. Greater reduction in risk seems possible if patients with CHD are managed according to guidelines. New approaches are needed to address the CHD management gap and related sex inequality. Working with GPs to better understand some of the factors influencing gaps in guidelines and practice would be a good start in narrowing the CHD management gap.
Key messages
What is already known on this subject?
Cross-sectional surveys conducted in Europe since the mid-1990s have reported an increase in the use of antihypertensive and lipid-lowering medications. Nevertheless, the majority of coronary patients surveyed were not meeting treatment targets.
Small-scale studies in Australia have reported coronary heart disease (CHD) management gaps in primary care from as early as two decades ago.
The General Practice Management Plan for CHD, a tool for the management of patients with CHD in primary care, was introduced in Australia in 2013.
What might this study add?
Using a large and contemporary general practice dataset in Australia, our study shows in 2014–2018 that female and younger patients with CHD are not managed optimally according to guidelines.
How might this impact on clinical practice?
We need to better understand the reasons for the persistence of these gaps in care and identify new ways to address them.
Acknowledgments
We would like to thank participating general practices for contributing primary health records to MedicineInsight.
References
Footnotes
Contributors CMYL conceived the design of the study, secured funding for study, obtained the data, drafted the manuscript and is guarantor of the study. GM analysed the data and is guarantor of the study. MW provided statistical oversight. CKC provided clinical advice. FS and SR secured funding for the study. RH conceived the design of the study and secured funding for the study. All authors contributed to the interpretation of the data and critical revision of the manuscript.
Funding This work was supported by a National Heart Foundation of Australia Vanguard Grant (101754).
Competing interests RH and CMYL have received financial support from the National Heart Foundation of Australia for the submitted work; MW has received personal fees from Amgen and Kirin outside the submitted work; CKC has received grants from National Health and Medical Research Council and National Heart Foundation outside the submitted work.
Provenance and peer review Not commissioned; externally peer reviewed.
Patient consent for publication Not required.