Search

This study empirically describes the activity and costs associated with the psychiatric services in England & Wales, Scotland & Ireland from 1845 to 1985. Basic economic concepts are employed to structure data. Key hypotheses from the three historical schools (Progressive, Social Control and Epidemiological) are tested, using a neo-Popperian methodology. The literature is reviewed in Chapter 2 (general and socio-economic) and Chapter 3 (epidemiological). Chapter 4 sets the comparative legal and financing context. Chapters 5, 6 and 7 outline activity and financial data for each country. Major points common to each included: - relative stability in admissions and costs up to around 1945, - a major change after 1945 when admission rates (both first and total) rose sharply, - inflows exceeded outflows until around 1955 when the number of resident inpatients began to fall, - despite the fall in resident inpatients, spending, which was predominantly publicly financed and hospital orientated, continued to increase. Chapter 8 suggests that three hypotheses were not obviously falsified by the data: - a pharmacological hypothesis linking new drugs in the mid-1950s to the changes which occurred then, - Wagner's Law, which predicts continued high levels of spending despite reductions in hospital places, and - Hare's epidemiological hypothesis which relied on the increase in first admission rates between 1850 and 1900. Four other hypotheses related to costs were rejected. Three paradoxes are identified for further investigation. Chapter 9 shows that the decline in the stock of inpatients occurred almost entirely in the long-stay (over 5 years) group, due mainly to deaths rather than live discharges. This pattern helps explain the continued high levels of spending. Ireland's outlier status is linked to both demand and supply side factors. Chapter 10 draws some overall conclusions and offers priorities for further research.

The thesis examines methods to measure or attribute resource use and costing for economic evaluations in health care. The literature review found minimal evidence comparing top-down and bottom-up (micro) costing. To cost nursing inputs for patients, studies rarely measured staff time and poorly reported methods. In chronic haemodialysis (HD), 'case mix' variations in nursing between patients were ignored. The empirical work evaluated nurses' self-recording using barcode scanners and observer work sampling to measure the nursing time per patient. Initial piloting eliminated patient-level work sampling due to problems linking data to patients. Barcode scanning captured 80% of nurses' hours; data quality was acceptable. It covered 4 weeks for 169 patients. Costs, in 2006, included employers' National Insurance and superannuation. Relative to the 'top-down' nursing expenditure per HD session (£44.56 to £50.79), the bottom-up cost was underestimated by up to 10%: 4% due to the unit cost using expected rather than actual working hours, and 6% due to missing patient-level resource use data. Multiple linear regression clustered by patient found those ineligible for care at satellite units needed extra nursing input (mean 8 minutes, 95% Cl 4-11, or £2.30 to £7.22 per session) compared with those eligible. Conclusions were that top-down (expenditure based) and bottom-up estimates of staff costs cannot reconcile due to averaging at different points, their attribution of resource use or costs to patients, and valuation of unit costs. More guidance is required on which unit cost of staff time (per hour paid, worked or patient-related) best reflects the opportunity cost of staff time. Barcode scanning successfully captured data, but required considerable research effort, making it impractical for most multicentre studies. Cost differences between patients were 5-14% of the nursing cost per session or 1-5% of the overall cost per session. Hence, they had minimal effect on results of economic evaluations.