Purpose: Identifying factors associated with poor 30 day outcomes following hospital discharge for heart failure patients is important for improving quality of care. Discharge on appropriate medications is a critical factor to improve outcomes. Analyses generally have not differentiated discharge medication recommendations based on prognosis such as referral to hospice after discharge.
Methods: Hospitalized Medicare patients with an adjudicated acute HF diagnosis during 2006-2011 were identified from the Atherosclerosis Risk in Communities (ARIC) study of a biracial cohort of 15,792 participants from 4 US communities since 1987. Logistic regression was used to explore two questions: 1) whether using hospice within 30 days of discharge was associated with being discharged on at least one of three medications [ACE inhibitors/angiotensin receptor blockers (ACEI/ARB), beta-blockers (BB), or diuretics]; and 2) how inclusion of patients likely close to death affected associations of discharge medications (separately or in combination) with hospital readmission or death within 30 days. Hospitalizations for participants who used hospice and died within 30 days following discharge were excluded for the analysis of readmission or death. All regressions controlled for patient demographic and clinical characteristics.
Results: The study sample included 780 participants (mean age 76, 45% male, 35% black) with 1457 hospital discharges; within 30 days, 24% of discharges were readmitted and 5% died. Hospice was used within 30 days for 105 discharges (7.4%). Using hospice was strongly associated with not being discharged on at least one medication (OR 0.22; 95% CI 0.11,0.46). Forty discharges used hospice and died within 30 days but only 26 had complete data. Excluding these cases, regression with complete data for 1240 hospitalizations (684 patients) showed that thirty day readmission or death was lower (p<0.05) for patients discharged on ACEI/ARB (OR 0.71, 95% CI 0.53,0.95) or diuretics (OR 0.66, 95% CI 0.48,0.91) but not for BB. A regression assessing the joint effect of discharge on all three drugs had the strongest reduction (OR 0.56, 95% CI 0.41,0.76); including hospice users who died within 30 days without readmission showed only a slightly stronger effect (OR 0.53, 95% CI 0.39,0.72), possibly due to the addition of only 26 stays. Other predictors positively associated with readmission or death included race/site, atrial fibrillation, stroke, and longer hospital stay; hospice was weakly associated with a lower likelihood.
Conclusions: As heart failure progresses, transfer to palliative or hospice services is often appropriate. Including patients when death is imminent in outcome analyses could overestimate medication effects. Mechanisms for identifying such patients other than hospice use are limited but constitute an important goal for future research.