Health economic evaluation

The proposed primary endpoint and methods for the economic evaluation follow the reference case set out by the National Institute for Health and Care Excellence (NICE) [80]. The primary economic analysis will be a cost-utility analysis over 16 months presenting incremental cost-effectiveness ratios (ICER) for intervention (UC + DCM™) versus control (UC), with effects expressed in terms of quality-adjusted life years (QALY). The analysis will adopt the health care and personal social services perspective. Analysis of the uncertainty surrounding the ICER will be undertaken using non-parametric bootstrap simulation (10,000 simulations) and presented on a cost-effectiveness plane and a Cost-Effectiveness Acceptability Curve (CEAC) [81]. A net benefit regression approach will also be employed using the model selected for the clinical effectiveness analysis [82]. Where Net Monetary Benefit (NMB) = (£20,000 × QALYs) − Costs NMB regression will enable covariate control and clustered analysis. There will be no modelling forward of benefits and discounting (at 3.5 %) will be conducted for values post 12 months. We will use the NICE willingness to pay per incremental QALY threshold range (lambda [λ] = £20,000-£30,000) to determine cost-effectiveness.

Utility values will be captured using the EQ-5D-5 L (primary) and DEMQOL-Proxy-U [83] (secondary). Residents will self-report where able on the EQ-5D-5 L or the EQ-5D-Proxy will be used. Resident- and proxy-reported data will be reported and analysed separately. However, given that only a proportion of residents will be able to complete the questionnaires, we will explore whether it is valid to use one source of data as a substitute for the other.

The total cost of DCM™ will incorporate the costs of training staff and staff time spent delivering the intervention as well as travel costs and any other expenditure (e.g. on training materials). The assumption for the analysis will be that the local authority pays for the provision of care home care for residents. We will include a sensitivity analysis where a proportion of residents are considered to pay toward their care home costs. Researchers will collect health care resource use data for each resident participating in the trial at baseline, 6 and 16 months using individual care home records and care plans. This will be supplemented by care home-level data collection, which will enable some validation of individual-level data. Unit costs for health service staff and resources will be obtained from national sources such as the Personal Social Services Research Unit, the British National Formulary and NHS reference cost database.