Many evidence based cardiological treatments reduce coronary heart disease (CHD) deaths. These treatments together explained over 40% of the substantial fall in CHD deaths between 1981 and 2000.¹ However, the CHD National Service Framework (NSF) recognised in 1999 that barely half of all eligible patients actually received effective treatments for myocardial infarction (MI), angina, or heart failure. Uptake rates were consistently worse among women, the elderly, and the deprived.^2,3 This study therefore examined the reduction in CHD deaths potentially achievable through increasing treatment levels in England and Wales.

METHODS

The previously validated cell based IMPACT model was used to combine data on (1) numbers of patients in specific CHD groups; (2) the prescription rates for all standard CHD treatments in 2000; and (3) the effectiveness of these treatments, defined as survival benefit over a minimum of one year, from the largest and most recent meta-analyses or randomised controlled trials.¹ Cumulative benefit from polypharmacy in individual patients was estimated by the Mant and Hicks formula, where relative benefit  =  1 − (1 − treatment A) × (1 − treatment B) × (1 − treatment C), etc. Compliance (concordance) for medical treatment was assumed to be 100% while patients were in hospital, 70% among symptomatic patients with angina or heart failure, and 50% among patients with hypertension or increased cholesterol. Uptake level was defined as prescription rate times adherence.¹

Having estimated the actual reduction in CHD deaths in 2000, we then used the IMPACT model to examine the consequences of increasing the uptake (prescription) rates of specific medical treatments in each disease category to reach 80% of all eligible patients (100% was considered unrealistic).²

The corresponding calculation was performed for revascularisation, assuming that coronary artery bypass graft (CABG) surgery and percutaneous transluminal coronary angioplasty (PTCA) procedures in 2000 were increased by 80% (substantially more than the NSF targets).²

Multiway sensitivity analyses were then performed by using the analysis of extremes method.¹ Minimum and maximum mortality reductions were generated by 95% confidence intervals from meta-analyses for treatment efficacy and from minimum and maximum plausible values for patient numbers, treatment uptake, and adherence.¹

All data sources, clinical definitions and International classification of diseases codes are detailed on our website (www.liv.ac.uk/PublicHealth/sc/bua/IMPACT-Model-Appendices.pdf).

RESULTS

In 2000, specific medical and surgical treatments in England and Wales were estimated to prevent or postpone about 25 805 deaths for at least one year (minimum estimate 17 110, maximum estimate 49 040) (fig 1, table 1). However, uptake (prescription) rates were generally mediocre. For instance, treatment rates among MI survivors averaged 56% for aspirin, 34% for β blockers, and 25% for statins. Similarly, patients with heart failure managed in the community averaged just 56% for angiotensin converting enzyme inhibitors, 17% for statins, and 15% for β blockers (fig 1, table 1). Increasing treatment rates to reach 80% of eligible patients could have prevented or postponed about 20 910 additional deaths (minimum estimate 11 030; maximum estimate 33 495). Of the 20 910 fewer deaths, 4680 (22%) would have resulted from increasing secondary prevention after an acute MI or revascularisation, and 7285 (35%) fewer deaths would have resulted from increases in heart failure treatments for patients in the community and in hospital (fig 1, table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Estimated coronary heart disease mortality reductions in 2000 and potential gains if specific treatments reached 80% of eligible patients. CABG, coronary artery bypass graft; MI, myocardial infarction; PTCA, percutaneous transluminal coronary angioplasty.

Extending primary prevention statin treatment to 80% of the 7.6 million “healthy” people with total cholesterol concentrations above 6.2 mmol/l would have prevented about 3295 deaths, representing 16% of the total gain, compared with 2370 (11%) fewer deaths from initial treatments for acute MI, 945 (4%) from treatments for hypertension, and 1475 (7%) from increases in aspirin and statins for patients with angina managed in the community (fig 1, table 1).

Only 400 (2%) additional deaths would have been prevented by an 80% increase in revascularisation procedures in 2000, and just 305 (1%) fewer deaths would have resulted from increased treatments for unstable angina. Irrespective of whether best, minimum, or maximum values were used in sensitivity analyses, the major potential gains consistently came from secondary prevention and heart failure, followed by statins and initial infarction treatments (fig 1, table 1).

DISCUSSION

In 2000, barely half the patients with cardiac disease actually received the appropriate treatment in England and Wales, much as elsewhere in Europe.³ If just 80% of eligible patients with CHD had received the medical treatments indicated, then over 20 000 extra deaths could have been prevented or postponed, almost doubling the mortality reduction actually achieved, consistent with older studies.⁴

Furthermore, almost two thirds of the total potential additional benefit would have come from focusing on secondary prevention and heart failure in primary care. Because absolute benefit is greater in older groups, they have the most to gain. The 2003 general medical services contract will now reward the identification of eligible patients and the creation of CHD registers in every general practice. Such incentives may substantially increase treatment uptakes. The increasing enthusiasm for chronic disease management programmes and nurse led primary care clinics focused on secondary prevention and cardiac rehabilitation should also help. The situation in 2005 may therefore be substantially better than that in 2000.

We generously assumed that CABG surgery and PTCA procedures in 2000 were increased by 80%. This was substantially more than the NSF had achieved by 2003 (some 6000 additional procedures over 1999 rates).² Relatively few deaths were prevented. However, revascularisation is being increasingly seen as a symptomatic intervention for improving quality of life, rather than simply for saving lives.²

All analytical models have limitations.¹ The IMPACT model was confined to CHD and did not explicitly consider patients with stroke or peripheral disease. Patients with diabetes were considered only in terms of their established CHD. The IMPACT model also assumed that efficacy, the mortality benefits reported in randomised controlled trials, can be generalised to effectiveness in unselected patients in clinical practice. A constant relative risk reduction, independent of the level of risk, was also assumed. Overestimation of the true treatment benefits therefore remains possible. Further explicit assumptions were required to cover deficiencies in the UK CHD data, which remain lamentably patchy and mixed.⁵ Sensitivity analyses were therefore essential to examine the effect of varying these underlying assumptions and hence test the robustness of the model.¹ Maximum and minimum estimates were generally narrow. Furthermore, the relative contribution of each intervention remained remarkably consistent. This study focused on mortality reduction. Further research is now required on life years gained, symptom relief, quality of life, cost effectiveness, and the potential reduction in serious non-fatal events such as recurrent MI, stroke, or heart failure often leading to repeated hospitalisation.²

In conclusion, future national strategies should maximise the delivery of appropriate treatments to all eligible patients with CHD and prioritise secondary prevention and heart failure.