Physician and patient predictors of evidence-based prescribing in heart failure: A multilevel study

Background: The management of patients with heart failure (HF) needs to account for changeable and complex individual clinical characteristics. The use of renin angiotensin system inhibitors (RAAS-I) to target doses is recommended by guidelines. But physicians seemingly do not sufficiently follow this recommendation, while little is known about the physician and patient predictors of adherence. Methods: To examine the coherence of primary care (PC) physicians' knowledge and self-perceived competencies regarding RAAS-I with their respective prescribing behavior being related to patient-associated barriers. Cross-sectional follow-up study after a randomized medical educational intervention trial with a seven month observation period. PC physicians (n = 37) and patients with systolic HF (n = 168) from practices in Baden-Wuerttemberg. Measurements were knowledge (blueprint-based multiple choice test), self-perceived competencies (questionnaire on global confidence in the therapy and on frequency of use of RAAS-I), and patient variables (age, gender, NYHA functional status, blood pressure, potassium level, renal function). Prescribing was collected from the trials' documentation. The target variable consisted of >= 50% of recommended RAAS-I dosage being investigated by two-level logistic regression models. Results: Patients (69% male, mean age 68.8 years) showed symptomatic and objectified left ventricular (NYHA II vs. III/IV: 51% vs. 49% and mean LVEF 33.3%) and renal (GFR < 50%: 22%) impairment. Mean percentage of RAAS-I target dose was 47%, 59% of patients receiving >= 50%. Determinants of improved prescribing of RAAS-I were patient age (OR 0.95, CI 0.92-0.99, p = 0.01), physician's global self-confidence at follow-up (OR 1.09, CI 1.02-1.05, p = 0.01) and NYHA class (II vs. III/IV) (OR 0.63, CI 0.38-1.05, p = 0.08). Conclusions: A change in physician's confidence as a predictor of RAAS-I dose increase is a new finding that might reflect an intervention effect of improved physicians' intention and that might foster novel strategies to improve safe evidence-based prescribing. These should include targeting knowledge, attitudes and skills