Factors that predict compliance in a virtual cardiac rehabilitation program

Abstract

Background: Despite the well-established benefits of cardiac rehabilitation (CR) for patients with cardiovascular disease (CVD), participation in CR remain low. Virtual CR programs present a unique opportunity to promote utilization. To date, few virtual CR cohorts have been analyzed for compliance. This study aims to determine factors that predict compliance within a large virtual CR program in the United States.

Methods: We analyzed 1409 patients enrolled in the Kaiser Permanente Mid-Atlantic States Virtual CR program that consists of 12 CR sessions via telephone. Demographic characteristics, as well as body weight, blood pressure, HbA1c level, and smoking status were collected at admission. Patients were further classified by CVD diagnosis codes. Compliance was defined as at least 75% (9/12 sessions) attendance. Data was analyzed using simple and multiple regression models with significance defined as P < 0.05.

Results: Age was the single strongest predictor for virtual CR compliance (adjusted R² = 0.58; P < 0.001), and non-compliant patients were younger. HbA1C level, CVD diagnosis codes, and smoking status each moderately predicted compliance (adjusted R² = 0.48, 0.42, and 0.31, respectively; P < 0.001). Smoking and HbA1C level combined in a multiple regression model significantly improved prediction of compliance (adjusted R² = 0.79, P < 0.01). Sex, baseline weight or hypertension were not significant predictors of CR compliance.

Conclusions: Age, diabetes, CVD diagnoses, smoking status at admission are independent predictors of compliance in a large virtual CR program. Targeted intervention could be designed accordingly to improve CR compliance.

Keywords: Cardiac rehabilitation; Compliance; Predictor; Virtual.