Heterogeneity exists within the patient population with coronary heart disease and the cost effectiveness of treatment may vary across subgroups within the overall population. This study compares the cost effectiveness of a secondary prevention intervention for a combined patient population relative to three selected subgroups: patients aged over 70 years; patients with a diagnosis other than angina only (that is, patients with a history of myocardial infarction, coronary artery bypass graft and/or percutaneous transluminal coronary angioplasty); and patients with diabetes. The results for the general population have been published elsewhere, but ongoing budget constraints require consideration of the appropriateness of targeting resources to patient subgroups. We adopt a probabilistic model to combine within trial and beyond trial impacts of treatment to estimate the lifetime health care costs and quality-adjusted life years of two primary care-based secondary prevention strategies: SPHERE Intervention--tailored practice and patient care plans and Control--standardised usual care. In all cases, the intervention was associated with mean cost savings and mean QALYs gains, when compared to the control, though statistical significance was never achieved. However, the probability of the intervention being cost effective was higher than 85% in all analyses across a range of potential cost-effectiveness threshold values. There is no compelling statistical evidence to support the targeting of specific subgroups across the general population. However, if affordability constraints are binding, the results do allow a tentative ranking of priorities based on the probabilistic subgroup analysis.