**Background:** The COVID-19 pandemic led to lockdowns and suspension of non-urgent healthcare services, including center-based cardiac rehabilitation (CR). Many CR programs switched to virtual delivery. This study aimed to understand the characteristics and correlates of disease-related knowledge and exercise self-efficacy among cardiac patients attending a virtual CR program during the first COVID-19 lockdown in Peru, one of South America's hardest-hit countries.
**Methods:** This prospective observational study was conducted at the CR unit of the Cardiovascular Institute (INCOR Essalud) in Lima, Peru, between August 2020 and December 2021. A total of 240 patients with coronary heart disease, heart failure, cardiomyopathy, valvular heart disease, or arrhythmias completed self-administered confidential surveys online at pre- and post-CR. Exclusion criteria included inability to attend online sessions, visual or cognitive mental disorders, severe conditions (LVEF<40%, NYHA III/IV, complex arrhythmias, decompensated heart failure, recent hospitalization, cardiac device carriers, severe pulmonary hypertension, post-surgery complications, decompensated diabetes). The virtual CR program consisted of 24 synchronous sessions (3 times/week for 8 weeks) via Zoom, with small groups of 8 participants. Each session included 30 minutes of education and 40 minutes of physical training. Educational content was based on the Spanish version of Cardiac College™, adapted for virtual delivery using slides, videos, and printed materials. Knowledge was assessed using the CADE-Q SV (20 items, max score 20, Cronbach's alpha 0.89). Exercise self-efficacy was assessed using a modified Bandura's 18-item ESE scale (5-point Likert, Cronbach's alpha 0.84). Paired t-tests and Pearson correlation coefficients were used for analysis, with partial correlations adjusting for pre-CR scores.
**Key Results:** The sample was predominantly men (75.4%), with mean age 61.9±13.2 years. Cardiac diagnoses included stable CAD (28.3%), valve disease (27.9%), and STEMI (27.5%). Most patients underwent PCI (45%), had hypertension (58.3%), and 95.8% had at least one cardiovascular risk factor. Mean total knowledge scores improved significantly from 12.9±2.4 to 15.6±2.0 out of 20 (p<0.001), with significant improvements in 4 of 5 knowledge areas (CV risk factors, exercise, nutrition, and psychosocial risk; p<0.001). Mean exercise self-efficacy improved significantly from 1.9±0.9 to 3.0±0.9 out of 5 (p=0.01). Post-CR knowledge was significantly correlated with gender (r=0.20, p=0.02), cardiac diagnosis (r=0.14, p=0.03), and surgical procedures (r=0.23, p=0.01). After controlling for pre-CR knowledge, correlations remained significant for cardiac diagnosis (r=0.17, p=0.02) and surgical procedures (r=0.27, p=0.02). Post-CR self-efficacy was significantly correlated with cardiac diagnosis (r=0.23, p=0.01) and surgical procedures (r=0.18, p=0.02), which persisted after adjustment (r=0.19, p=0.02 and r=0.20, p=0.02, respectively). Knowledge improved significantly in those younger than 65 years, with stable CAD, STEMI, congenital heart disease, PCI, no surgical procedures, 1 risk factor, and diabetes.
**Clinical Implications:** This is the first study in a low- to middle-income country to assess the effects and correlates of virtual CR with a focus on education. The virtual CR program improved both disease-related knowledge and exercise self-efficacy during a period of significant uncertainty and restrictions. Self-efficacy is a modifiable characteristic and a strong predictor of physical activity behavior in chronic disease. The findings suggest that cardiac diagnosis and surgical procedures are important correlates of post-CR outcomes, which can help healthcare teams—including vascular nurses—tailor educational interventions. Limitations include potential selection bias, single-center design with convenience sampling, multiple comparisons inflating error rates, low correlation coefficients (all <0.30), and inability to make causal inferences. Further research in other low-resource virtual CR settings is recommended, including assessment of sociodemographic characteristics such as educational level and family income.