Analytical approach:
The economic evaluation was carried out alongside a clinical trial, with a one-year time horizon. The authors stated that they took the perspective of the UK NHS and Personal Social Services.
Effectiveness data:
The clinical data were from a randomised controlled trial (RCT), carried out at the Mater Hospital in Belfast (Northern Ireland). This enrolled 173 patients; complete data were available for 127 of these, with 64 (mean age 66.2 years; 57.8% women) in the programme group and 63 (mean age 66.6 years; 55% women) in the usual care group. The reasons for drop-out and loss to follow-up were provided. Randomisation was carried out using the minimisation method, and follow-up was one year. The primary endpoint was the change, from baseline to one year, in health-related quality of life.
Monetary benefit and utility valuations:
Health-related quality of life was estimated by the patients who were enrolled in the clinical trial, using the European Quality of life (EQ-5D) instrument, at baseline, six months, and one year. Preference weights were derived, using the time trade-off technique, from a representative sample of the general population.
Measure of benefit:
Quality-adjusted life-years (QALYs) were used as the summary benefit measure.
Cost data:
The economic analysis included the costs of hospital pharmacist’s time, general practitioner (GP) visits, emergency department visits, hospital bed days, and the programme, including steroids and antibiotics for exacerbations, the administrator’s time, printing, and overheads. Resource quantities were collected by a specific questionnaire, from patients in the clinical trial, and from their medical records. The unit costs were from national sources, such as the Personal Social Services Research Unit, the British National Formulary, NHS trusts, and the Belfast Health and Social Care Trust. All costs were in UK pounds sterling (£), at 2006 to 2007 prices.
Analysis of uncertainty:
Bootstrapping was used to calculate 95% confidence intervals around the projected model outcomes. A similar approach was used to assess the overall uncertainty and to produce cost-effectiveness acceptability curves. An imbalance in the mean baseline utility between groups was controlled for using multiple regression. One-way sensitivity analyses were carried out by varying the number of bed days and the cost of the programme. Missing QALY and cost data were imputed using linear regression, with treatment group, age, gender, and COPD severity as covariates.