Analytical approach:
A decision analytic Markov model was used to assess costs and outcomes associated with the two interventions. The mode comprised three health states: pre-full time care, full time care and death. The time horizon was five years. The authors reported that the perspective was that of the UK National Health Service (NHS) and Personal Social Services (PSS).
Effectiveness data:
Clinical and effectiveness data were obtained from previously published studies that included randomised controlled trials and a UK epidemiology study. The main effectiveness estimates used in the model were the effectiveness of the interventions on cognition, functioning and behaviour. These estimates were obtained from a previously published meta-analysis of six large multicentre randomised controlled studies (see Other Publications of Related Interest for more bibliographic details). Transitions probabilities between the three health states were derived using a new published predictive equation of time to full time care. The predictive equation was derived using the London and South East Region (LASER-AD) UK epidemiological study.
Monetary benefit and utility valuations:
EQ-5D utility estimates were obtained by mapping relevant components of the 12-item Health Status Questionnaire, Ferm’s D-test and the quality of life scale in Alzheimer’s disease into each of the five domains of the EQ-5D.
Measure of benefit:
The measure of health benefit was quality-adjusted life-years (QALYs) gained. Benefits could be generated over the lifetime of the patient and future benefits were discounted using an annual rate of 3.5%.
Cost data:
The direct costs included in the study related to routine patient management, medications, hospitalisation, social community services and institutionalisation. These costs were obtained from the LASER-AD study. In this study the Client Service Receipt Inventory (CSRI) was adapted for use with elderly patients and used to record all healthcare and social care resource use. All resource use was then valued using UK published sources. The price year was 2009. Costs could be incurred over a five-year time horizon and future costs were discounted using an annual rate of 3.5%. All costs were reported in UK pounds sterling (£).
Analysis of uncertainty:
A series of one-way sensitivity analyses varied each model parameter to determine the key drivers of uncertainty in the model. A probabilistic sensitivity analysis was undertaken by fitting probability values alongside the model parameters. This analysis was run with 10,000 Monte Carlo simulations.