Background Serial measurement of natriuretic peptides may guide management in heart failure (HF) patients. In previous trials, natriuretic peptides were infrequently monitored, which may undervalue the benefit of this approach. Methods and results HOME was an adaptive three-arm randomized clinical study to test whether home monitoring of BNP could reduce HF-related death, hospitalization due to acute decompensated HF (ADHF), and ADHF treated with intravenous diuretics in the emergency department or outpatient setting. Enrolment was terminated early because of slow enrolment, low event rates, and the belief that an algorithm for assessing BNP trends was needed. Justification for pooling data from all study arms was made and analysis as a single observational study was performed. The analysis resulted in 107 patients who were monitored for a median of 172 days with BNP measures on a median of 74% of days. BNP values were highly variable within a patient. Dispersion between serial BNPs was calculated to be 39.3%, 57.7%, and 73.6% for 1, 60, and 120 days between measures, respectively. A moving average filter (fBNP) was calculated to reduce day-to-day fluctuations and track changes from week to week. There were 27 primary events in 17 362 patient days of monitoring; the hazard ratio for time-varying fBNP was 2.22 (95% confidence interval 1.48-3.34) per unit natural log (corresponding to a 2.72-fold change in fBNP level). Conclusion The HOME HF study demonstrates the feasibility of home BNP measurement and shows the potential value of fBNP as an index of emerging clinical deterioration. Assessment of the clinical value of this is required.