Bristol homeopathic hospital costs study
Increasingly, NHS commissioners need information about the costs of services to help inform their decision making. Currently, commissioners rely almost entirely on crude cost data to evaluate the economic impact of services, if any cost data are available at all. The aim of the BISCUIT study was to pilot approaches to economic evaluation of homeopathic packages of care delivered at Bristol Homeopathic Hospital. We wanted to test the methods and outcome tools to facilitate the development of a larger study that could investigate the differences in resource use between those who use NHS homeopathic packages of care and those who do not. To answer this, we used a cost utility model and carried out the first steps of a discrete choice experiment. Using a prospective matched controlled cohort design, we recruited 29 Bristol Homeopathic Hospital patients from 315 possible candidates, but only received the relevant permissions to access medical records and subsequently found matched controls for nine. Nineteen matched controls returned data. Qualitative interview data was collected to learn about resource use, experiences of the study and views on the outcome tools for 23 participants. Quantitative data was collected through GP medical record extraction of data such as GP and practice nurse consultations, medications, hospital consultations, tests and investigations. In addition, participant questionnaires were administered to capture wellbeing (Warwick Edinburgh Wellbeing tool), quality of life (SF-36) and resource use (Client Service Resource Inventory). This latter tool collected data on personal costs such as private healthcare, over the counter products, travel and lost earnings. Qualitative data were analysed using qualitative content analysis and a framework approach. Descriptive analyses were carried out on the quantitative data by an independent statistician. We found that:• Because only 9 of the original 315 potential candidates from the Bristol Homeopathic Hospital were matched with controls, representativeness is questionable.
• The outcome tools were feasible, acceptable, sensitive to change and appropriate.
• Cases and controls were well matched at baseline for quality of life and wellbeing, but not for resource usage.
• Quality of life for case participants improved compared to controls. Wellbeing for case participants also changed significantly for cases compared with controls. However because this was not a randomised controlled study, we do not know if the improvements seen were due to the Bristol Homeopathic Hospital service.
• At baseline, costs were substantially higher for case participants. There was no difference in resource utilisation in the 12 months after baseline between cases and controls.
• Individual variability in resource utilisation for both cases and controls was substantial.
Although undeniably challenging, the importance of designing good quality economic evaluations for generalist services like the Bristol Homeopathic Hospital is important. Therefore, we would encourage future researchers to work through the methodological difficulties of conducting economic evaluations of services to ensure NHS commissioners - and the wider public - have the information necessary to make appropriate judgments. For further information, please see full project report (PDF, 1,549kB)