Introduction Outpatient based management of heart failure is gaining popularity in the UK. Observation data and a small pilot randomised controlled trial suggested it is safe and clinically effective. However, the cost-effectiveness of this novel treatment approach is not known. Patient level costing is thought to be more accurate providing extra granularity when reviewing the cost of patient pathways.

Methods We used the Trust’s patient level costing models from financial years 2018/19 and 2019/20 to calculate total treatment costs for the patients in the pilot prospective trial which randomised acute heart failure patients to receive inpatient vs outpatient based therapy. Patient level costing for inpatients would include hospital bed, nursing cost, consultant /ward doctor cost, non-pay costs (pharmacy including iv drugs, diagnostics e.g. pathology, radiology, ECG, echocardiography).

Where patient level costs are unavailable, for example Community contacts, we have used a national average cost.

• Community centre visit - HRG code of N29AF 2018/19 National average cost £71

• Home visit - HRG code of N29AF 2018/19 National average cost £71

• GP visit - National average cost of £30 [1]

All other costs used are from the Trust costing model. We also tested the feasibility of collecting quality of life data using EQ5D-5L and the visual analogue scale (VAS) where 100 means the ‘best health you can imagine’ and 0 is the worst.

EQ-5D-5L

5 items each with a different domain: mobility, self-care, activities, pain and depression. Each is scored from 1-5 where 5 is the worst (severe limitation/unable to do). Value sets are used to transform the health profile into an index value that can be interpreted as a health utility; these range from -0.594 to 1.000 where a value below zero is taken to describe a health state whose quality is ‘worse than death’.

Results 24 patients were randomised to 13 outpatients and 11 inpatients between July 2018 and June 2019: Inpatients received 5.8 ± 2.8 days of treatment, outpatients received 8.5 ± 5.2 (p=0.12). No patient was in CCU, HDU or ITU or received dialysis/haemofiltration. 1 patient randomised to inpatient care died during the index episode, but none died in the outpatient group. There was one extra A&E visit. There were no extra GP visits during the index episode for outpatients, and the cost of extra visit for consultant /HF clinic (12 extra visits) has been factored into the equation (£1536).

Inpatients costed £55892 (i.e. mean £5081 per patient); Outpatients costed £31435 (£2418 on average). Therefore total average saving was £2663 per patient.

Treatment groups varied little in their quality of life. EQ-5D-5L VAS scores are significantly higher for outpatient. [Figure 1]

Indeed a linear mixed model confirms that treatment arm is significantly associated with higher VAS scores; an outpatient reports a VAS score 13 points higher on average than an inpatient, and there is evidence that the score increases slightly over time (0.18 points per day, equivalent to ≈ 11 points over the 60 day period) for all patients. [Table 1] This is not true of EQ-5D utility indices; estimates for treatment and time are non-significant (p=0.78 and p=0.88 respectively).

Conclusion Outpatient based AHF management appears cost-effective. Total average saving was £2663 per patient. Treatment groups varied little in their quality of life. EQ-5D-5L VAS scores are significantly higher for outpatient.

Conflict of Interest No conflict of interest