Introduction Cardiac Resynchronisation Therapy (CRT) and implantable cardiac defibrillators (ICDs) are a well-established, scientifically proven and guideline recommended treatment for selected patients with left ventricular systolic dysfunction (LVSD), reducing mortality and/or hospitalisation, with the latter being the most expensive part of the treatment of such patients. Attempts have been made to maximise the yield of CRT/ICD devices by use of sensors which predict the onset of CHF well before the patient deteriorates, hence avoiding the need for hospitalisation and in the process conserving valuable financial resources. HeartLogic is one such new software incorporated into Boston Scientific (BSC) CRT/ICD devices which uses multiple sensors to track physiological trends, combining them into one composite index alerting healthcare professionals of potential worsening CHF. The MultiSense Study demonstrated a 70% sensitivity in detecting CHF events, a median of 34 days prior to clinical heart failure. Aims In this study we describe our experience of the use of HeartLogic in a tertiary care cardiology centre with a well-established complex devices and heart failure service. Methods Retrospective analysis of patients (n=49) with BSC devices and in whom HeartLogic was activated in March/April 2019. Patient records (electronic patient records, including Latitude downloads) were reviewed and the results of those having alert(s) of ≥16 were compared with those who did not during a follow-up period of 288 ± 6 days (median 288, range: 282-327). Results Whole cohort: mean age: 68±11 years (median 72, range 33-87), 37/49 (75.5%) males. 35/49 (71.4%) had LVSD due to ischaemic cardiomyopathy. Devices implanted were: Resonate X4 CRTD 28/49 (57.1%), Resonate EL ICD 13/49 (26.5%) and the remainder, Momentum CRT-D 8/49 (16.3%). During the monitoring period number of alerts were as follows: ≥ one alert: 22/49 (44.9%) patients, 2 alerts: 10/49 (20.4%), 3 alerts: 7/49 (14.3%); 4 alerts: 2/49 (4.1%) and 5 alerts: 1/49 (2%). For each patient who had an alert, the maximum alert value was considered for analysis. The maximum alert score was 26.6 ± 12.3 (median: 23, range: 16-67), and the duration of alert was 32.7 days (median: 29, range: 4-83). For each of the 5 physiological parameters, a change in the expected direction at the time of alert was evaluated. S3 was the most sensitive (increase 15/22, 68.2%), followed by thoracic impedance (decrease 14/22, 63.6%) and S1 (decrease 13/22, 59.1%). Respiratory rate and night heart rate were less sensitive parameters (increase 11/22, 50% and increase 9/22, 40.9% respectively). Though not part of HeartLogic, activity levels showed a decrease in 13/22 (59.1%) of patients who had an alert. Characteristics of those who did and did not have an alert were compared (Table 1). Heart Failure Nurse contact and change in medication over monitoring period were statistically significantly more common in the group with alerts. In our real world experience with HeartLogic, expected change in values of S3, thoracic impedance, S1 and activity levels were the most sensitive markers of CHF and so predictive of alerts, leading to early and increased frequency of Heart Failure Nurse contact and significant intervention, possibly helping in prevention of hospitalisation and in turn conservation of valuable financial resources. Conflict of Interest None