Purpose of the study - The purpose of this study is to develop and evaluate the efficacy of a new, theoretically based intervention to improve medication adherence in persons with HF. The hypotheses include:

Hypothesis I. Poorly adherent patients with symptomatic HF who receive the intervention (n=40) will be more adherent to medicines during the12-month intervention than a control group (n=40) of poorly adherent patients with symptomatic HF.

Hypothesis II. Poorly adherent patients with symptomatic HF who receive the intervention (n=40) will have fewer hospital readmissions (HR) and emergency department (ED) visits during the 12-month intervention than a control group of poorly adherent patients with symptomatic HF.

Study Activities and population - A prospective, randomized controlled design will be used to pilot test the efficacy of a new intervention to promote adherence to the medication regimen in chronic heart failure in the clinical setting. The initial development and feasibility testing of the intervention is complete (Preliminary Work, Section 4.5).

Patients with HF symptom exacerbation (n = 80) who exhibit high likelihood of poor adherence, as determined by a validated screening measure, the Medication Adherence Scale (MAS)89 at baseline assessment, will be recruited and randomized to receive the intervention or usual care with attention control. Medication adherence, symptom frequency and intensity, hospital readmissions (HR) and emergency department (ED) visits will be assessed in both groups at 3, 6 and 12-month clinic visits. The intervals between visits are considered sufficient to minimize sensitization bias to psychometric measures. Longitudinal measurement is required to evaluate the magnitude of change in adherence and symptom-related events occurring over time. Efficacy will be measured as improved adherence (primary outcome) and decreased HR and ED visits (secondary outcomes) in the intervention group as compared to the attention control group at 12 months. The study will close when all patients have had a 12-month post-enrollment clinic visit.

Data Analysis & Safety/Risk Considerations - The investigators will use generalized linear models to test the primary and secondary hypotheses (McCullagh and Nelder, 1989). Logistic generalizations of the traditional, multivariate GLM are appropriate when the dependent variable is binary and the probability of an event is modeled, as is the case with adherence in this study. When the dependent variable is a count, a commonly used generalized model uses the natural logarithm of the count, as with re-hospitalizations and ED visits in this study. In compliance with NIH guidelines for data safety and monitoring activities, a number of quality control steps will be used to ensure the on-going safety of participants and the scientific integrity of this feasibility intervention study. The proposed study poses only very minimal risks to participants. All participants will receive usual care. Participation in the study will not replace patients' regular health care attention. Data safety will be monitored by the PI, the Data Safety Officer (Dr. Karl Swedberg), and the study co-investigators. Any adverse events will be reported immediately to the Duke IRB and the NIH.

Data Safety. First, a project database will be constructed, using participant ID numbers and including all demographic, pre- and post-intervention history and all assessment materials. A separate file, cross-referencing participant identification with project ID numbers, will be maintained with access limited to the PI or a designated member of the investigative team.

Inclusion Criteria:

• > 18 years old
• symptomatic heart failure
• telephone required

Exclusion Criteria:

• heart failure due to congenital etiology
• patients who cannot provide self-care