A Heart Failure Management Program Using Shared Medical Appointments

a nurse practitioner (NP), a dietitian, and a clinical psychologist, similar to what has been shown to be successful and cost-effective in nonveteran populations. ^6-8 Patients attended at least 4 sessions before graduating to the advanced HF SMA program where they could attend monthly booster sessions. The program promoted self-management by providing education about and support for the HF process, HF medications, diet adherence, physical activity, psychological well-being, and stress management via interactive presentations. During the visit, patients’ medication and food logs were reviewed. Patients were encouraged to discuss successes and obstacles in achieving their goals. All study procedures were approved by the institutional review board at JBVAMC.

Study Design

Data were collected by retrospective review of the JBVAMC EHR. The EHR was reviewed for all veterans scheduled for ≥ 1 SMA clinic visit within the HF specialty clinic using predetermined, convenient selection between January 1, 2012, and December 31, 2013. Outcome data were collected through 12-month follow-up (through December 31, 2014).

Patients in both treatment arms received HF care through the HF clinic, including one-on-one education regarding HF self-management provided by a NP. Patients were assigned to the HF SMA group if they also attended the HF SMA clinic within 3 months of their initial HF clinic consult. The number of SMAs attended was included as a covariate in the models. Patients who were scheduled for, but did not attend, the HF SMA clinic were assigned to the HF clinic group. Patients who attended the initial HF consult before September 1, 2011, were excluded, thereby ensuring that all patients included in the present analyses had the opportunity to attend the HF SMA appointment within the predetermined period of chart review.

Data for all VA hospitalizations that occurred between January 1, 2012 and December 31, 2014, were extracted from the EHR. Extracted data included admission date, discharge date, and discharge diagnoses. From these data, the authors assessed 4 hospitalization outcomes for each HF hospitalization and all-cause hospitalization within 12 months of the initial HF clinic consult date: hospitalization (yes/no), number of hospitalizations, number of days in the hospital, and days to first hospitalization.

Data Analysis

Demographic, HF characteristics, and HF outcome variables for the HF SMA and HF clinic groups were compared using t tests and chi-square analyses. Logistic regressions were used to predict 12-month hospitalization, linear regressions were used to predict number of hospitalizations and number of days hospitalized, and Cox proportional hazards regressions were used to predict time from initial HF consult to first hospitalization for each HF-related hospitalization variable and all-cause hospitalization variable. A separate logistic regression was conducted to predict 12-month all-cause mortality. The primary predictor variable of interest for all models was group membership (HF SMA vs HF clinic). Covariates