All patients admitted to a metropolitan tertiary referral hospital in Sydney with an ACS between April 2003 and February 2004 were identified using diagnosis-related groups (DRG) codes14 for unstable angina (F72A, F72B), myocardial infarction (F41A, F41B, F60A, F60B) and chest pain (F74Z). Hospital medical records were reviewed to determine secondary prevention eligibility. Reasons for ineligibility included living more than 20 km from the hospital; insufficient understanding of English to give informed consent; and uncontrolled congestive heart failure, severe comorbidity, or death.

Of the total of 824 admissions, 378 patients (46%) were ineligible for participation in a formal cardiac rehabilitation program (reasons given in Box 1). All 446 eligible patients were contacted by mail and telephone within 12 months of admission and invited to volunteer for research investigating CHD prevention; 235 declined (reasons given in Box 1) and three were excluded because they had previously attended cardiac rehabilitation. Altogether, 208 patients attended for an assessment (Box 1), 64 of whom were about to commence cardiac rehabilitation (standard cardiac rehabilitation [SCR] group) and 144 were not (no cardiac rehabilitation [NCR] group). Reasons for not attending cardiac rehabilitation included: transport problems (19%); work commitments (16%); doing own exercise (8%); not interested (10%); and not referred (47% — due to patient/physician preference and/or scarcity of resources). Risk factors for all eligible patients who did not attend face-to-face assessment were extracted from medical records to permit comparison of risk factors with those who attended assessment.

Standard cardiac rehabilitation is offered onsite to outpatients in a gym at Concord Hospital. The program includes two 60-minute gym sessions and a 2-hour education session each week for 6 weeks. Weekend or evening sessions are not available.

Ethical approval for the study was obtained from the then Central Sydney Area Health Service, Concord (Repatriation General Hospital Zone) and the University of Sydney. All participants gave written informed consent.

Data were analysed using SPSS for Windows (version 12.01; SPSS Inc, Chicago, Ill, USA) and are presented as mean and SEM or as proportions. Differences in outcome measures between groups were compared using t tests for continuous variables and either χ2 tests or Fisher’s exact tests for proportions of categorical variables. Two-tailed P values of < 0.05 were considered significant.

Most of the patients were male, born in Australia and were not working (Box 2). Fewer patients in the NCR group than in the SCR group had undergone coronary revascularisation (56% v 86%; P < 0.001), and the mean (± SEM) time between ACS admission and risk-factor assessment was longer in the NCR group (6.3 ± 0.2 months) than the SCR group (3.3 ± 0.2 months). Socioeconomic status was not significantly different between the NCR and SCR groups. Fewer patients in the NCR group than in the SCR group reported taking a statin (69% v 83%, P = 0.02). Other cardiovascular medications were similar in the two groups (Box 3).

A significantly greater proportion of patients in the NCR group than the SCR group had total cholesterol levels > 4.0 mmol/L, low-density lipoprotein (LDL) cholesterol levels > 2.5 mmol/L, were physically inactive, obese, depressed, and current smokers (Box 4), and had four or more cardiac risk factors. The mean (± SEM) LDL cholesterol level, cardiac depression scale depression score, number of modifiable risk factors and LIPID score were also significantly greater in the NCR than the SCR group (Box 5). Further, the proportion of patients with a LIPID score ≥ 5 (medium-to-high risk) was significantly higher in the NCR than the SCR group (49% v 22%; P < 0.001). Quality of life measured using the SF-36 questionnaire was significantly worse in the NCR group compared with the SCR group in the dimensions: physical functioning (71.6 ± 1.5 v 76.5 ± 2.0; P = 0.008); vitality (50.5 ± 1.5 v 57.3 ± 2.6; P = 0.04); and role-emotional (78.3 ± 2.9 v 93.7 ± 1.8; P = 0.001).

Of the patients in the total cohort with at least one modifiable risk factor, most (137/194, 71%) were unable to state any risk factors. A significantly greater proportion of patients in the NCR group than in the SCR group were unable to state any of their risk factors (75% v 58%; P = 0.013). Also, significantly fewer patients in the NCR group than the SCR group could state nationally recommended targets for total cholesterol level, blood pressure and physical activity (Box 6).

To assess whether the number of modifiable risk factors present affected an individual’s ability to state their own risk factors, we examined group difference for patients with one or more risk factor and for those with two or more risk factors. For patients with one or more risk factor, there was no difference in the proportion who could state at least one risk factor between the NCR (47/142, 33%) and SCR groups (20/55, 36%). However, among patients with two or more risk factors, those in the NCR group were significantly less likely to state at least one risk factor than those in the SCR group (31/128, 24% v 14/37, 38%; P < 0.001). In both the NCR and SCR groups, patients with two or more modifiable risk factors were significantly more likely to state at least one risk factor than patients with only one risk factor present.

Of the 235 eligible patients who declined attending for assessment, significantly fewer had experienced myocardial infarction than patients in the NCR group (20% v 58%; P < 0.001). However, there was no significant difference between those who declined assessment and those in the NCR group in the proportion with a total cholesterol level > 4.0 mmol/L (73% v 78%); systolic blood pressure ≥ 140 mmHg (53% v 43%); diabetes (28% v 28%); current smoker (21% v 21%); mean (± SEM) total cholesterol level (4.8 ± 0.1 v 4.7 ± 0.1 mmol/L) or mean (± SEM) systolic blood pressure (137.1 ± 1.5 v 136.9 1.2 mmHg).