Heart failure (HF) affects nearly 5 million Americans and is the leading discharge diagnosis in the Veterans Administration (VA). HF is the most common indication for readmission, lowers quality of life (QOL) and greatly increases mortality. Behavioral modifications such as diet, physical activity, and medication adherence are crucial to decreasing HF recurrence and improving QOL. We completed an interdisciplinary theory-based prospective quality improvement program (QIP) intervening at the patient, provider, and hospital system level and compared it to current best practice (CBP) evaluated in the pre-intervention phase.
The overarching hypothesis driving this study was that a comprehensive QIP will lower HF recurrence compared to CBP. The primary aim was to determine the impact of QIP on HF-specific and general quality of life. Secondary aims examined impact of OIP on 30 day emergency room visits, medication adherence at 3 months, diet adherence at 3 months, satisfaction, and intervention acceptability. Exploratory aims assessed the effect of QIP on health care utilization and outcomes at 6 months. We hypothesized that the QIP group will have better outcomes compared to the CBP group for the various aims. In addition, this study will provide preliminary data for a rigorous effectiveness trial testing this promising intervention among HF patients.
The VALOR in HF study is a prospective pretest-posttest control group study in two VAMC's with primary and secondary patient outcomes being measured at baseline and at 3 month. The QIP included patient counseling prior to discharge at the inpatient teachable moment with advice on self-care and self-monitoring using a personal checklist, use of provider checklists and default quality improvements to improve provider performance and three monthly telephone-delivered Transtheoretical Model-based stage-matched counseling sessions (discussed self-monitoring, use of checklist, diet, medication adherence, and other self-care items) for patients. The study participants were veterans discharged after an inpatient stay where they received a diagnosis of heart failure and had access to a home phone. Exclusion criteria included poor short term survival, recent major surgery, being discharged to a long term care facility, severe psychiatric illness, not living in the area and other logistic reasons. The primary aim, QOL was measured using the SF-36 and the disease-specific Minnesota Living with Heart Failure Questionnaire (MLHFQ). Other measures were assessed by either validated instruments from the field or ones that were developed from previous research studies at the VA. Unadjusted evaluation between QIP and CBP were conducted by Fisher's Exact test or Chi-Squared test for discrete variables, by Student's T-test (two-tailed) for normally distributed continuous variables and by Wilcoxon rank-sum test for continuous variables not normally distributed. Readmission, emergency room visits and death were analyzed using time-to-event analysis for 30, 90 and 180 days.
We enrolled 136 participants (QIP 68 and CBP 68). Overall, the response rate for the 3-month follow-up was 72% (CBP 75.0% and QIP: 69.1%). The average age was 73 years, 98.5% were male, 61.8% were White, 33.09% were Black/African American, 51.5% had more than 12 years of education, 24.3% were currently married, 54.4% had diabetes, 41.2% had renal failure, 59.6% had coronary artery disease, 87.5% had normal cognitive functioning, and 55.1% reported living alone. These baseline characteristics and comorbidities were not significantly different among CBP and QIP (all p-values > .05). The CBP to QIP comparisons at 3 months for General QOL, as measured by the SF-36 with higher values representing better QOL, were: physical role function scale (45.05 vs. 58.51; p=0.046), general health (40.58 vs. 54.16; p=0.009), standardized mental component scale (68.60 vs. 71.30; p=0.5825) and the standardized physical component scale (31.84 vs. 36.66; p=0.0484). The CBP to QIP comparisons at 3 months for disease-specific QOL, measured using the MLHFQ (lower scores represent better QOL), were: Overall (45.96 vs. 38.78; p=0.2555) and physical dimension (20.96 vs. 17.57; p=0.1473) and emotional dimension (9.64 vs. 8.15; p=0.3575). Change in disease specific QOL (change in MLHFQ score) for the QIP versus CBP comparisons were: overall (-8.96 vs. -12.15; p=0.4304) and physical dimension (-4.22 vs. -7.70; p=0.0995). Average days to emergency room (ER) visits for the CBP group were consistently shorter than for the QIP group: at 30 days (9.88 vs. 12.89), at 90 days (31.73 vs. 36.48) and at 180 days (50.44 vs. 51.19). Average days to readmissions for CBP and QIP were: at 30 days (12.19 vs. 16.06), at 90 days (28.59 vs. 38.59) and at 180 days (57.10 vs. 56.88). Of the three monthly calls, 88.2% of QIP received at least one phone call and 54.4% received all three calls. Medication adherence at 3 months was higher for CBP than QIP for the Morisky Score (52.08 vs. 57.50; p=0.6705) and equal for refill compliance (0.89 vs. 0.89). The CBP to QIP comparisons for adherence to treatment were: diet (56.25% vs. 75.00%; p=0.0593), medication (93.62% vs. 99.00%; p=0.2449) and physical activity (46.81% vs. 71.43%; p=0.0186). There were zero (0) deaths at 30 days, four (4) deaths at 90 days and ten (10) deaths at 180 days; average days to death for the CBP and QIP were: at 90 days (54 vs. 59) and at 180 days (86 vs. 110.14).
There were improvements in the primary outcome, both general and disease-specific QOL, for the intervention (QIP) group. For all other outcomes, there is a general trend toward improvement for the QIP group. This study will provide the foundation to launch a SDP whose primary aim will be to evaluate the effect of QIP on readmission rates. It will be conducted in the same pre-discharge and post-discharge settings as the pilot study, where veterans admitted with HF are treated. QIP holds promise as an effective way to modify HF-related health behavior, improve HF management and lower readmissions.
- Natarajan S, Rodriguez M. Predicting hospital readmission in patients with heart failure: Usefulness of psychosocial factors not included in established risk scores. Paper presented at: American Public Health Association Annual Meeting and Exposition; 2013 Nov 5; Boston, MA.