Although features of heart failure disease management programs are broadly outlined, little is known about which interventions are actually used in the outpatient setting or which patients are most likely to require interventions.

Between September 2001 and June 2002, we enrolled 32 patients admitted to the Brigham and Women's Hospital Heart Failure Services, Boston, Mass, with decompensated heart failure. The intensity of care and outcomes of these patients were prospectively tracked for more than 90 days. During this time, there were 325 patient contacts (median 8.5 per patient), including 247 calls (median 7 per patient) and 78 clinic visits (median 2 per patient). Brigham and Women's Hospital clinicians adjusted diuretics a total of 109 times (median 2.5 times per patient). When frequency of diuretic adjustments was used to estimate the intensity of care, higher values of blood urea nitrogen at discharge predicted an increased intensity of care during the 90-day follow-up (relative risk [RR] 1.2, 95% confidence interval [CI] 1.0-1.3, P = .02). When frequency of clinic visits, telephone calls, and diuretic adjustments were used to estimate intensity of care, discharge creatinine (RR 1.03, 95% CI 0.99-1.06, P = .05), discharge blood urea nitrogen (RR 1.13, 95% CI 1.04-1.23, P = .004), and length of stay (RR 1.07, 95% CI 1.00-1.13, P = .04) were predictors of the composite end point.

Even after undergoing optimization of medications during admission for acute heart failure, patients in a comprehensive disease management program required frequent interventions to maintain clinical stability. Renal dysfunction was the strongest predictor of increased interventions and worse outcome.