Methods	Parallel group RCT
Participants	259 patients randomised. Results available for 250. Mixed CHD patients, MI 52%, CABG 55%, PTCA 59%. Mean age 62 yrs (range 38 - 86), 62% male. Recruited from three outpatient clinics
Interventions	All participants received usual CR programme of structured exercise and individual and group classes (four) on diet modification and stress reduction. At the end participants were given an exercise prescription that included target heart rate (THR) zone and advice to exercise at least 5 times per week for 30 minutes INTERVENTION: CHANGE programme (,Change Habits by Applying New Goals and Experiences-), based on several cognitive behavioural frameworks (social problem solving model, self efficacy theory, expectancy value theory, relapse prevention theory) . CHANGE program given in three 1/2 hour sessions, once per week in the last three weeks of the CR programme. Two further sessions held at one and two months post CR programme. Sessions were provided by cardiac nurse in small group and centred on: small group social interaction, peer modelling, self assessment, goal setting and problem solving activities reinforced at later stages
Outcomes	Adherence to exercise amount: (10 hours of moderate intensity exercise a month - 150 min/week) Adherence to exercise frequency: (at least 5 times/week or 20 times/month) Both measured at 12 months.
Notes	Mean duration of an exercise session among those who exercised was longer than 30 min recommended (mean session length 52 minutes). Men were less likely to discontinue exercise than women. Participants with higher comorbidity scores or more muscle and joint pain were more likely to discontinue exercise
Risk of bias
Item	Authors’ judgement	Description
Adequate sequence generation?	Yes	Computerised minimisation stratification randomisation program used managed by program director in which participants were stratified on gender and site of recruitment
Allocation concealment?	Yes	The randomisation sequence was concealed until intervention was assigned
Free of other bias?	Yes
Blind outcome assessment? All outcomes	Yes	Exercise measured using portable wristwatch heart rate monitors, backed up by diaries mailed to investigators. Data collectors were blind to study group