Article Text

Original research
Heartwatch: an Irish cardiovascular secondary prevention programme in primary care, a secondary analysis of patient outcomes
  1. Robyn Homeniuk1,
  2. Fintan Stanley2,
  3. Joseph Gallagher3,
  4. Claire Collins1
  1. 1Research, Policy and Information, Irish College of General Practitioners, Dublin, Ireland
  2. 2Research Hub, Irish College of General Practitioners, Dublin, Ireland
  3. 3ICGP HSE Primary Care Lead for Integrated Care Programmes (Cardiovascular Disease), Irish College of General Practitioners, Dublin, Ireland
  1. Correspondence to Dr Claire Collins; claire.collins{at}icgp.ie

Abstract

Objectives To investigate patient follow-up data from Heartwatch: Ireland’s secondary prevention programme for cardiovascular disease delivered in general practice.

Design Retrospective descriptive study based on secondary analysis of routinely collected data from Heartwatch.

Setting Heartwatch targeted 20% of general practices in Ireland and recruited 475 general practitioners across 325 practices.

Participants The patient population included people with a history of acute myocardial infarction, percutaneous transluminal coronary angioplasty or a coronary artery bypass graft. Over 16 000 patients entered the programme however, to assess the long-term progress of patients, we identified a cohort of 5700 patients with at least 8 years in the programme.

Interventions A standard protocol for continuing care of patients for the secondary prevention of cardiovascular disease was administered by general practices. The programme was designed using WHO and European Society of Cardiology guidelines on secondary prevention.

Outcome measures A Continuing Care (CCare) score out of eight was the primary outcome measure used. It was calculated based on programme targets for well-known cardiovascular risk factors: exercise, systolic blood pressure, LDL cholesterol, optimally controlled glucose, smoking status, and pharmacological treatment.

Results After 1 year, 37% of the 8-year cohort had achieved a CCare score >5 increasing to 44% after year 8. Patient sex was predictive of better scores; male patients had almost a half-point advantage (0.432, 99% CI: 0.335 to 0.509). Patients who enrolled earlier following their qualifying event and patients with more frequent visits were also more likely to achieve higher CCare scores.

Conclusions Overall, patients are not likely to meet all targets set by secondary prevention guidelines, however, supporting patient self-management may impact on this. Early enrolment after a cardiac event and frequent structured care visits should be priorities in the design and implementation of similar programmes. Ongoing evaluation of them is necessary to improve outcomes.

  • primary care
  • cardiology
  • coronary intervention
  • protocols & guidelines
  • preventive medicine

Data availability statement

Data may be obtained from a third party and are not publicly available. The Heartwatch INDC acts as the primary collection point for all Heartwatch data returned by practices. Within this structure there is a data management committee, which has the responsibility of reviewing access requests to the aggregated, anonymous version of the collected data. Data used here were released after such an access request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

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.

Strengths and limitations of this study

  • The key strength of Heartwatch is the volume of patient data and the length of time the programme has been in existence with patients retained.

  • The data come from an active clinical programme. There is no comparative or control group.

  • LDLc (Low-density Lipoprotein cholesterol) guideline targets changed during the programme so guidelines initially would have suggested a higher target, although we retrospectively applied the most recent guideline target. Hence, as a stricter target is applied, we have underestimated the number achieving the desired active target prior to 2016.

  • There is likely a survivor bias on the available long-term information, as those with worse scores may have exited the programme earlier than 8 years and we do not have access to outcomes such as death or major adverse cardiovascular events.

Introduction

Globally, cardiovascular diseases (CVDs) are the leading cause of death, with 32% of all global deaths—17.9 million—being attributed to them in 2019.1 In Europe, cancer and circulatory diseases have been the leading causes of death since 2006.2 Recent statistics have shown that from 2006 to 2016 the number of deaths from ischaemic heart disease fell by 28.4% for men and 34.2% for women.2 In 2016, the standardised death rate from ischaemic heart disease in Ireland was 133 per 100 000 inhabitants, which was slightly more than the EU rate of 119.4 per 100 000 inhabitants.2

While the decline in deaths from ischaemic heart disease is promising, it is still a major cause of mortality in Ireland.3 The Central Statistics Office state that circulatory system diseases made up 28.9% of all deaths in Ireland in 2019.4 This was the second leading cause of death after malignant neoplasms.4 As highlighted by the evidence review in the Sixth Joint European Society of Cardiology (ESC),5 patients with a history of CVD may need long-term support to change their behaviour and limit the risk of further cardiac events. Furthermore, evidence from clinical trials have shown the benefits of secondary prevention following acute myocardial infarction (AMI), percutaneous coronary intervention (PTCA) or coronary artery bypass grafting (CABG).3 6 7 Provision of comprehensive cardiac rehabilitation, similar to the comprehensive approach in Heartwatch, was shown to have more patients achieve risk factor targets.8 As of 2013 nearly 80% of Irish patients were compliant with cardiac rehabilitation recommendations.9 Moreover, the WHO has suggested that providing CVD management of risk factors under universal health coverage and at a primary care level can reduce the burden of CVD.1 Ireland was the first country in the European Union to implement a standardised, national programme led by general practitioners (GPs)—called Heartwatch—that strategically implemented the ESC guidelines.10 Other countries have since also adopted secondary prevention programmes that monitor blood pressure, cholesterol, smoking status and physical activity.11–13

In Ireland, GPs work as private healthcare professionals charging private patients per visit and receiving government payments on a capitation basis for eligible public patients. Around 43% of people in Ireland qualify for free GP care, either through the General Medical Scheme card (32.4%) or GP visit card (10.4%).14 GPs have a central role in the Irish health system, and they are critical in the management of long-term conditions with 80% of all GP visits relating to chronic disease management.14 A fifth of Irish general practices were recruited to deploy the specially developed secondary prevention programme, which enabled patients to attend up to four specialised visits per year, with a payment made per visit to the GP from the state. The stated aim of Heartwatch was to reduce the morbidity and mortality caused by CVDs in Ireland. It has attempted to improve the care of patients with heart disease in the community at general practices across the country. It was an evidence-based programme, strategically designed to provide community-based care the integrates specialised disease management in general practice with referrals to other services as necessary. However, measuring the programme’s success has been multifactorial and complex. This paper examines the follow-up results of patients over 8 years to determine if there are long-term benefits of this secondary prevention programme and what factors may influence or predict the types of patients who benefit.

Methods

Data handling

Collection

Heartwatch is a national structured programme led by Irish GPs with a standard protocol for the continuing care of patients for the secondary prevention of CVD. The programme has been reported on in 2004, 2005, 2008, 2011 and 20143 15–18—this paper is the first that reviewed patients who have attended the programme for at least 8 years.

In 2003, 475 GPs in 325 practices were recruited19 to provide this national secondary prevention care programme. Heartwatch targeted 20% of GPs to review patients on a quarterly basis with care implemented according to defined clinical protocols. Practices from each health board area were recruited with the aim of having national coverage of the programme. Each area employed a regional GP coordinator and nurse facilitator to assist with the deployment of Heartwatch care protocol. On been granted a contract to provide the programme, the clinical staff in the GP’s practice underwent specific training to enable a standardised approach to applying the care protocol and performing the required checks at each visit. The patient population included men and women who had a history of AMI, PTCA or a CABG.15 16 In addition, patients with diabetes from an established diabetes structured care programme were also invited into the programme—however, these patients are not included in the analysis in this paper due to differing treatment requirements. Heartwatch, was introduced as a collaborative national pilot programme15 but was not expanded beyond 20% of GPs. The programme and its continuing care protocol are based on the internationally recognised cardiovascular prevention guidelines from ‘Prevention of Coronary Disease in Clinical Practice 1998’20 and those updated in 200320 and 2016 after the sixth Joint Taskforce guidelines were released.5 In 2016, the target level for lipoprotein cholesterol (LDLc) for very high-risk patients was changed to 1.8 mmol/L (from 3.0 mmol/L) as a direct result of the change in guidelines.5

By employing the standard continuing care protocol of Heartwatch, eligible patients may have attended up to four visits per year with their GP practice after signup. Measurements of key risk factors were recorded at the signup visit and at each subsequent visit along with data related to medication changes and referrals21 (see online supplemental table 1). This information is securely uploaded directly from the practice patient management server in an anonymised format to the Independent National Data Centre (INDC).21

Whether all factors were measured at every visit was dependent on whether the value was within target or not at the previous visit. For example, the target for total cholesterol is <5 mmol/L—if a patient is within target, their GP only needed to measure at every other visit whereas if they were outside of the target their GP must repeat the test at the subsequent visit. However, the GP/practice nurse may have chosen to repeat all tests at each visit.21

During the analysis of medication data, patients were categorised as receiving or not receiving specific prescriptions—‘decreased dose’, ‘increased dose’, ‘maintained’ and ‘new’ were considered as receiving; and ‘not prescribed’ and ‘discontinued’ as not receiving.

Access

The Heartwatch INDC acts as the primary collection point for all data returned by practices.15 Within this structure, there is a data management committee, which has the responsibility of reviewing access requests to the aggregated, anonymous version of the collected data. In the reporting, the Strengthening the Reporting of Observational studies in Epidemiology cohort reporting guidelines were used.22

Processing

For this paper, data from all consultations January 2003 to March 2020 were extracted in December 2021. For patients to be included in the overall analysis (figure 1A–E), they must have had at least one valid initial visit (baseline) between January 2003 and March 2020. For the 8-year follow-up analysis, only those individuals who also had valid 2-year, 4-year and 8-year follow-up visits were included. None of the patients recruited through the diabetes programme were included in the analyses presented here. Patients needed a minimum of one visit per year for those 8 years, but the intervals were not always the same because some patients attend more frequently than others. However, as part of the automated checks undertaken by the system, there must be a minimum of 10 weeks between visits for a practice to schedule a visit and upload data.21

Figure 1

Heartwatch overview 2003–2020. (A) All Heartwatch visits graphed by year of visit. (B) Each year of follow-up with total number of patients graphed. (C) Population pyramid of all patients. (D) All patients grouped by earliest qualifying event type. (E) All patients grouped by interval from earliest qualifying event to date of first Heartwatch visit. AMI, acute myocardial infarction; CABG, coronary artery bypass grafting; PTCA, percutaneous coronary intervention. *January 2003 to March 2020.

Patients could have attended their GP up to four times per year under Heartwatch. However, as the number of visits per year and time between visits varied, the definition of what was the first visit of each year of follow-up was applied retrospectively. The first-ever visit was defined as Year-1: Visit-1. The earliest date 1 year after this was defined as Year-2: Visit-1, however, given the variation in attendance a 30-day variance was given, so it would have been the earliest visit at least 335 days after the first visit. Later years were calculated similarly—Year-3: Visit-1 was the earliest visit 2 years (±30 days) after Year-1: Visit-1.

Some patients had more than one recorded qualifying event (QE)—AMI, PTCA or CABG. In these cases, counts and intervals were calculated based on the earliest recorded QE occurrence.

Outcome measure development and calculation

Rather than relying on individual targets to determine the success of the patient, a preliminary care outcome score was developed. It was based on EUROASPIRE studies7 23 24 and the methods used by Ergatoudes et al.11 They scored patients across six outcome measures derived from guidelines—exercise, systolic blood pressure (SBP), LDLc, optimally controlled glucose, smoking status and pharmacological treatment—then considered the number patients who met six guidelines, five guidelines and so on.

This initial method was applied to a subset of the Heartwatch dataset to estimate the number of people meeting each metric. However, Ergatoudes et al11 only focused on patients in the 2 years after AMI, and the included targets needed to be adjusted for the Heartwatch context to include care guidelines for patients with PTCA and CABG. The methodology and preliminary results of this process were then scrutinised and refined with input from GP specialists in cardiology and diabetes, researchers and data experts.

Following this agreement, patient care outcome scores were calculated within the cohort with 8 years of follow-up recorded. Statistics were run on this cohort on the baseline and 8-year data. The metrics selected and their metric score varies from 1 to 2 (table 1). For optimally controlled glucose, patients without a QE of diabetes mellitus (DM) were give two points because of the high prevalence of comorbidity of CVD and DM.25 These scores have been called the Continuing Care score (CCare score).

Table 1

Components, target levels and scoring used to calculate the Continuing Care (CCare) score outcome measure

Statistics

Statistical analysis was carried out using R (V.4.1) and RStudio (V.1.4).26

Given the repeated measures structure of the data (repeated visits per patient) a linear mixed-effects model was used, with maximum likelihood estimation of fixed and random effects. Fixed effects included patient level demographics (sex, age), programme adherence (average visits per year, visit number), signup context (QE type (dummied) and qualifying even interval), as well as possible two-way interactions. The model also considered random effects at the patient level, which allows individual patients to vary randomly in terms of their intercept (accounting for differing baseline readings). The mixed-effects model was estimated using the lme4 package,27 confidence intervals were calculated at 99% and conditional and marginal coefficient of determination were also calculated (see online supplemental table 2 for full model estimates).28 29

Patient and public involvement

Heartwatch was developed in collaboration with the Irish Heart Foundation Irish a national heart and stroke charity which supports and advocates for people who have been affected by heart and stroke. However, it was not possible to involve patients in this later secondary analysis due to data protection restrictions.

Results

Heartwatch overview: 2003–2020

Looking across all validated Heartwatch records between 2003 and March 2020. By the end the second year of Heartwatch, there were over 20 000 visits per year; annual attendance remained above 20 000 until 2012 (figure 1A). While overall attendance has decreased since the peak in 2008, the proportion of patients attending once, twice, three or four times a year remained stable between 2004 and 2020 (figure 1A). Over 16 000 patients had entered the programme, patients remained in Heartwatch for an average of 7 years (figure 1B). Over 7000 (45%) patients have been in the programme for 8 years or more.

There were more male (76%) participants compared with females. The majority of Heartwatch patients were over 60 years old when they signed up, with 27% aged <60 years old and 33% of all participants aged between 60 and 69 years old at signup (figure 1C). The median age at signup across all years of the programme was 67 and has not differed much over time (range: 63–67). The female group were typically older, with a median age of 70 compared with 65 for males (figure 1C).

An AMI was the most common QE (40%), with PTCAs and CABGs accounting for 35% and 25%, respectively (figure 1D). Overall, 18% of patients were enrolled within 1 year of their QE (figure 1E). Another 32% of patients enrolled between 1 and 2 years after their QE; with the rest singing up between 3 and 6 years (25%) or more than 6 years (26%) after their QE. Early signups on programme commencement tended to have longer intervals between event and signup (QE-Interval), (2003: mean 6 years) but the interval shortened and by 2006 stabilised (2006–2019: range: 2–3 years).

The 8-year cohort from 2003 to 2020

In the assessment of patients’ progress, we identified a cohort of 5700 patients with at least 8 years in the programme (figure 2A). The included patients had a minimum of one visit per year for 8 years between 2003 and 2020. The remainder of the analyses presented pertains to this cohort.

Figure 2

Heartwatch 8-year cohort overview 2003–2020. (A) Patient records graphed by year of follow-up. The records of the 8-year cohort are highlighted in pink. (B) 8-year cohort grouped by earliest qualifying event type. (C) Population pyramid of 8-year cohort. (D) 8-year cohort grouped by interval from earliest qualifying event to date of first Heartwatch visit. 8-year cohort, n=5729. AMI, acute myocardial infarction; CABG, coronary artery bypass grafting; CCare, Continuing Care; PTCA, percutaneous coronary intervention.

In this cohort, 38% of patients in the 8-year cohort had a PTCA as their first QE and CABG was again the least common type of QE (26%) (figure 2B); 77% were male and the median age at signup was 65 years old (figure 2C). A third of patients in this cohort were referred to the programme within 1–2 years of experiencing their QE (34%), the median QE-interval was 2 years (figure 2D).

The CCare scores

After 1 year in Heartwatch, the median CCare score was 5 (33% of patients), 30% of patients scored lower than this and the remainder scored >5. After 4 years, 37% of patients had individual-level improvements in their score, 36% of scores had not changed and 27% decreased. The number of patients who achieved scores >5 increased to 44% at this point (figure 3A). After 8 years of follow-up, 40% of patients scored ≥6. By the eighth year, patients’ scores were higher, although the ratio of higher–same–lower narrowed (36%:32%:32%) (figure 3A). At each time point, the median CCare score for the cohort was 5.

Figure 3

The CCare for the 8-year cohort in follow-up years 1, 2, 4 and 8. (A) The CCare scores for the 8-year cohort; proportion of cohort by number of metrics met. (B) The CCare scores by grouped age bands. (C) The CCare scores by grouped recorded sex. (D) The CCare scores by grouped year of first visit. (E) The CCare scores by grouped by number of visits per year. (F) The CCare scores by grouped by interval from earliest qualifying event to date of first Heartwatch visit. CCare, Continuing Care; QE, qualifying event.

Components of the CCare score and other metrics

At the start of the first year, 64% of the patient cohort was within the target for SBP (online supplemental table 1). This increased to 70% at year 4 but reverted to 67% in year 8. Exercise and antiplatelet/anticoagulant treatment showed a similar pattern of improvement up to year 4, with a degree of decline from year 4 to year 8 (table 2).

Table 2

Percentages of 8 year cohort within target under each score component, and by year of follow-up (n=5729)

The number of patients with comorbid diabetes that had HbA1c readings within target increased over time. However, this occurred in tandem with an increased new diagnosis of diabetes in the rest of the cohort (table 2). LDLc started with only 21% of patients being in target but steadily increased to 29% in year 8. Smoking rates likewise improved through to year 8. The rate of waist circumference within target did not change much through follow-up.

The values comprising the CCare score—as well as some not included in the score—demonstrated a similar pattern. Some risk factors showed continued improvement with follow-up—Total and LDL Cholesterol—others improved for the first 4 years of follow-up—Systolic BP, Exercise—while others did not improve—BMI, weight—and the prevalence of diabetes increased (online supplemental table 3).

In terms of medication changes, fewer patients were prescribed aspirin from year 1 (91%) to year 8 (87%). This was the only recorded prescription that had an overall reduction, however, it remained among the most frequently prescribed items. Beta-blockers and statins were also frequently prescribed items (year 8: 92%–73%). Prescriptions of diuretics started at 18% and increased to 24% over time. There were three medications that had a 6% increase in being prescribed; ace inhibitors (started at 45%), Ca channel blockers (started at 16%) and angiotensin II inhibitors (started at 10%). The prescription of diabetic medications also increased through follow-up (table 3).

Table 3

Medications prescribed among the 8 year cohort by year of follow-up (n=5729), with change from year 1 calculated for each later year of follow-up

Patient demographics and the CCare score

The patients’ sex was predictive of better scores; male patients had almost a half point advantage on females (0.432, CI: 0.335 to 0.509, p<0.0001). Female patients had lower CCare scores across all 8 years of follow-up, 26% had scores >5 in year 1, which rose to a maximum of 33% in year 4 and fell again to 28% in year 8, which was 15% points lower than the equivalent in male patients (41%, 47%–44%, respectively) (figure 3B). The increase among both sexes from year 1 to year 4 were similar (females: +7%, males: +6%) (figure 3B).

A patient’s age at signup does not appear to predict CCare scores in the 8-year cohort. The effect size was small and not found to be significant (0, CI: −0.004 to 0.003, p=0.737). While not significant there were some differences across age groups. More younger patients had a CCare score >5 at signup (<60: 42%) compared with older patients (60–69: 38%, 70+: 34%) (figure 3C). All age groups had more scores >5 after 4 years of follow-up (0–59: 46%, 60–69: 45%, 70+: 42%), but more older patients had improved scores to reach those levels (0–59: +4%, 60–69: +7%, 70+: +8%). The difference in the number of patients with a score >5 narrowed after 8 years of follow-up, and fewer patients achieved scores of >5 than they had after 4 years (0–59: 42%, 60–69: 40%, 70+: 38%) (figure 3C).

Signup, attendance patterns and the CCare score

The year of patient’s signup does not appear to predict CCare scores in the 8-year cohort, but those who registered earlier in the programme had generally lower scores at signup (figure 3D). Patients who had a CABG as a QE were predicted to have better scores than patients qualifying from an AMI (0.106, CI: 0.028 to 0.183, p<0.0001). Patients qualifying from a PTCA events do not differ much from AMI (0.038, CI: −0.045 to 0.121, p=0.234).

Longer intervals between a patient’s earliest cardiac event and first visit under Heartwatch were predictive of worse CCare scores (−0.031, CI: −0.040 to −0.023), although the effect size is small; an interval of over 16 years (2.5% of patients) would be required to predict a half point lower score. Patients with shorter intervals were more likely to have scores >5 (figure 3E). Shorter QE-interval patients were stable in the first 4 years and then dropped, whereas those with longer QE-intervals saw continued improvement. Similar to the year of signup, the improvements that were seen in patients with longer QE-intervals to signup were insufficient to allow them to reach the same level of scores as the shorter QE-interval patients. A two-interaction effect of QE-interval and visit number predicts that the negative effects of long intervals diminish with attendance (0.001, CI: 0.001 to 0.002, p<0.0001), at that effect size it might take over 7 years of visits to erase the negative effect of a 1-year interval.

The number of Heartwatch visits per year was predictive of higher CCare scores (0.109, CI: 0.051 to 0.168, p<0.0001). Looking at frequencies, patients who visited more often were more likely to have scores >5 (figure 3F). However, when grouped by visit per year, all groups still saw improvements throughout years of follow-up, each achieving their highest CCare score after 4 years. A two-interaction effect of visits per year and visit number predicts that the positive effects of regular attendance compound with time (0.003, CI: 0.001 to 0.006, p=0.002).

While the descriptive statistics indicate improvements/stability up to year 4 followed by decline/return to baseline (figure 3A), the model of 8 years predicted marginally worse scores over time (−0.014, CI: −0.022 to −0.005, p<0.0001). The marginal R2, representing the variance explained by the fixed effects was calculated at 0.04. The conditional R2, representing the fixed effects and the random intercept (individual patient baseline effects) was 0.57. This would indicate that the patient level variables only account for a fraction of the variance compared with the patient’s initial health status on signup.

Discussion

In this investigation of patient data from a cardiovascular secondary prevention programme, we found that patients achieved moderate improvements in blood pressure (+3% within target), LDL cholesterol (+7% within target) and smoking status (+4% non-smokers). To develop a broader understanding of patient success, an 8-point CCare score was created to monitor changes over 8 years. Less than 2% of patients achieved all targets, however by year 8, 71% of patients had achieved between five and seven of the eight targets.

Using statistical modelling, we found that longer time intervals between the qualifying cardiac event and starting the secondary prevention intervention predicted worse scores even after 8 years, but these patients can and do improve overtime, just not as rapidly as other patients. Female patients were also more likely to have worse CCare scores, both in the baseline and 8 year model. Moreover, patients who attended three or more visits per year had higher average CCare scores and maintained higher scores while patients who visited less frequently saw a decline in their outcomes. This could be a reason to promote attending secondary cardiac prevention as soon as possible following a cardiac event and maintaining a good level of contact with that intervention.

Strengths and limitations

The key strength of Heartwatch is the volume of patient data and the length of time the programme has been in existence with patients retained. Further to that, the data are geographically spread across general practices areas across Ireland.

However, this is an active care programme and not a randomised controlled trial, so no comparative or control group exists. Moreover, data are collected primarily for clinical monitoring, not research purposes, thus, some variables were calculated. For example, in our statistical models, visits per year is calculated retrospectively and so its value in a predictive model is constrained. A key strength is that this is real-world data.

ESC guidelines on LDLc changed during the programme, so patients may have been treated towards different target levels. We retrospectively applied the most recent recommendations ergo some patients may have been designated out of target who would have been in target at the time. Hence, we may have underestimated the number achieving the target as a stricter more recent guideline is being applied.

Another possible limitation could be a survivor bias on the available long-term information, as those with worse scores may have exited the programme earlier than 8 years.30

Heartwatch does not collect outcome information such as mortality or further cardiac events, nor does it collect patient-reported outcomes. This was a limitation which we have attempted to overcome by developing the CCare score method.

Comparison to other literature

It is difficult to draw a direct comparison with other literature because the programme is not a trial nor a survey of patients with a history of CVD but an ongoing care programme with real-world data. The programme, Heartwatch, presents a much longer-term view of secondary prevention compared with most literature which focuses on the first 6 months,31 first year or first 2 years following a cardiac event.8 32 There is substantial research on secondary prevention of CVD and identifying and managing risk factors for these patients which we have compared with our results. In a review of clinical trials looking at primary and secondary prevention of coronary artery disease, Kantaria et al found that reduction of LDLc, decrease in blood pressure and discontinuation of smoking resulted in reduced death rates and further cardiac events.33 This is promising, as these are the key areas in which Heartwatch patients improved.

An international review of risk factor management for patients with CVD in Asia, Europe and the Middle East also found sex-based differences where female patients were less likely to achieve total cholesterol, LDLc, glucose, physical activity and weight targets compared with male counterparts.34 These differences were smaller in Europe compared with Asia and the Middle East but persisted nevertheless which is congruent with our predictive model findings of sex-based outcome differences.

In 2009, the European Action on Secondary and Primary Prevention by Intervention to Reduce Events III (EUROASPIRE III)23 survey sought to determine whether the European guidelines were being followed in everyday practice. This survey found that large proportions of patients do not achieve the targets, more than half did not meet the blood pressure or cholesterol targets23 and they stated that European countries needed to raise the standard of preventive care. An Italian study of secondary prevention of coronary heart disease in primary care, which looked at health records of just under 6000 patients found 153 patients with CVD.10 This survey found that there was satisfactory adherence to guideline advice—46% of patients achieved LDL targets and 83% achieved the SBP target. They concluded that GPs are well placed to help people with a history of CVD to manage risk factors, but that care could be further optimised.

The Swedish study11 that inspired the development of the CCare score, found only 3.5% of people were achieving all targets 2 years after AMI. This is similar to our finding here where just 2% of patients achieved all eight targets at year 2. This highlights further the need for specific interventions of secondary prevention, as more Heartwatch patients achieved their SBP, 69%, and LDL targets, 23%, after 2 years compared with the those in Sweden where 57% achieved the SBP and 18.5% achieved LDL targets. However, it should be noted that Heartwatch has a more diverse patient group compared with the Swedish patients, who had suffered an acute incident. A Norwegian study of cardiac rehabilitation patients, showed that patients who had an acute incident were more likely to participate in secondary prevention.35

Finally, in the more recent EUROASPIRE surveys, smoking, obesity and exercise were persistent in their unlikeliness to change overtime but lifestyle changes were more successful if a patient was in a prevention programme.36 However, in the same Norwegian study as above patients who were overweight were more likely to participate in the programme, which shows a willingness to improve.35 In Heartwatch, exercise and waist circumference did not show much improvement over 8 years. Patients in the EUROASPIRE surveys stated lack of confidence as their main barrier to address unhealthy behaviour.

Implications for policy and practice

In current secondary prevention research and guidelines, there is a focus on measurement and development of risk factor targets. However, research has repeatedly demonstrated that patients are not meeting these targets.37 38 For example, results for males and females diverge under similar targets. The current evidence base should be used as a foundation to refocus secondary prevention research away from target definitions and onto the implementation of these programmes with added public and patient involvement.

We, and others,39 have shown early enrolment after a cardiac event and frequency of structured care visits should be priorities in the design and implementation of similar programmes.

The evaluation of patient outcomes and cost-effectiveness should take into consideration that new programmes can experience an initial influx from a backlog of high-risk cases. Chronic disease programmes may evolve significantly in the initial few years. Planning and evaluation should take this into consideration.

While 20% of Irish GPs were recruited to Heartwatch, the demographics and location of the practices is not available for analysis. As mentioned above, recruitment of patients was at the GPs’ discretion. As Heartwatch was not primarily designed for research, it did not capture socioeconomic status or cultural behaviours such as diet. Better prior design of the recruitment and data collection would be required to fully assess how valid our findings would be to the wider Irish population. However, in other justifications, the model has been cited to inform the design of local programmes, particularly in areas with similar general practice systems (eg, Australia).19 40

Our results on sex-based differences and earlier recruitment into the programme may apply to the broader Irish population, or other similar populations with CVD, as these are previously demonstrated factors which impact secondary prevention of CVD.34 39 41 Another finding likely to be true for similar programmes is the need to carefully consider monitoring and evaluation of patient outcomes while the programme and its data collection is designed.17

Conclusion

Secondary prevention of CVD can have a positive impact even when patients start with poor outcomes. The sooner a patient can access a structured care programme, the better but even with delays it is worth enrolling patients with a history of cardiac events regardless of age. Overall, patients are not likely to meet all targets set by secondary prevention guidelines—especially those related to lifestyle factors such as exercise and waist circumference. However, supporting patient self-management may impact on this and the inclusion of factors such as a patient-centred approach and regular training of health professionals to deliver same, as noted elsewhere.24 36 Ongoing evaluation and improvement of secondary prevention programmes is needed to help more patients successfully reach targets.42

Data availability statement

Data may be obtained from a third party and are not publicly available. The Heartwatch INDC acts as the primary collection point for all Heartwatch data returned by practices. Within this structure there is a data management committee, which has the responsibility of reviewing access requests to the aggregated, anonymous version of the collected data. Data used here were released after such an access request.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.

Acknowledgments

We thank Drs Suzanne Kelly and Mike O’Callaghan, clinical leads in the ICGP for their support and clinical expertise. Our thanks also to Patricia Patton, ICGP library for bibliography assistance. This work uses data that have been provided by patients and collected by their GPs as part of their care and support, we acknowledge these contributions and thank Sally-Anne O'Neill and Colleen O’Neil for Heartwatch administrative support and advice.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • RH and FS are joint first authors.

  • RH and FS contributed equally.

  • Contributors All authors contributed to the research question development. FS carried out the recoding of variables and undertook the data analysis. All authors designed the analysis approach, interpreted the results and formulated the conclusions. JG contributed clinical expertise. CC is the PI and manager of the Heartwatch Programme and provided direction and oversight for this analysis and paper. CC is also the author responsible for the overall content and acts as the guarantor for the work. RH and FS prepared the first draft of the manuscript and contributed equally to this paper. All authors contributed to the manuscript and all authors read and approved the final manuscript. All named authors contributed sufficient work according to the COPE guidelines.

  • Funding There was no additional funding obtained to undertake this secondary analysis. The Heartwatch programme is funded by the Irish Health Service Executive. The ICGP Research Hub is funded through the Irish Sláintecare initiative.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.

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

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.