Castelli, A. and Smith, P.C. (2006) Circulatory disease in the NHS: measuring trends in hospital costs and output. Research Report. CHE Research Paper (21). Centre for Health Economics , York, UK.
1. Following the publication of the Atkinson Review of the measurement of government outputs in the National Accounts, there has been great interest in measuring the productivity growth of the National Health Service. Such macro measures of productivity are important when deciding how much public money to devote to the NHS, and in holding the NHS to account. However, it is also important to gain an understanding of the productivity of individual programmes of care, so as to ensure that resources are allocated efficiently within the NHS. Hitherto, such information has not been available. This report is an exploratory study of the feasibility and usefulness of developing measures of growth in outputs, costs and productivity of a single programme of care within the NHS: hospital treatment of circulatory diseases.
2. In this context, productivity is defined as the ratio of an aggregate measure of outputs to an aggregate measure of inputs for the chosen programme of care. The key methodological challenges are (i) choosing the appropriate measures of NHS activities, (ii) adjusting those measures for the ‘quality’ of care, (iii) aggregating the measures into a single measure of output, (iv) identifying the associated inputs, in the form of a single measure of costs, (v) tracking these measures consistently over time.
3. The building blocks of the study are the measures of hospital ‘continuous inpatient spells’ (CIPs) of hospital treatment, which include emergency, elective and day case treatment. These capture trends in the volume of activity of the NHS over a six year period, 1998/99 to 2003/04. The type of spell is indicated by diagnosis, as indicated by the Hospital Resource Group (HRG) for the spell. The analysis embraces all HRGs relevant to circulatory disease, including cerebrovascular disease, coronary heart disease, and associated investigations.
4. The value to the patient of these activities is crucially dependent on the quality of the outcome achieved. The only consistent, universal and reliable measures of outcome currently available in the NHS relate to mortality, either within the hospital spell, or within 30 days of admission. Where appropriate, we therefore use these outcome measures to adjust the crude measures of activity for the quality of outcome achieved.
5. Ideally, we should also like to incorporate other measures of health outcome into the analysis, such as gains in the quality of life following treatment. Unfortunately, the NHS does not routinely collect health outcome data. However, using a small sample of outcome data for two procedures collected by BUPA, we demonstrate how health outcome data could be used to augment measures of NHS output, and argue that the NHS should move rapidly towards routine collection of such data.
6. The diverse activities (CIPs) that make up this programme of care do not confer equal patient benefits. They must therefore be aggregated using some system of weights that reflects their relative contribution to aggregate NHS output. In principle, these weights should reflect the average ‘health gain’ of the treatment. In practice, this is rarely available. We therefore follow conventional practice in aggregating treatments according to their estimated costs,acknowledging that this is far from ideal.
7. It is relatively straightforward to identify the total physical inputs consumed by the NHS as a whole, for example in the form of capital, labour and drugs. However, identifying the part of these inputs that is attributable to an individual programme of care is a major challenge. In particular, specific measures of the physical inputs used by the circulatory disease hospital programme are not available. Instead, as an indicator of inputs consumed, we had to use the measures of reference costs developed by the NHS for the HRGs under consideration. These offer some insights into trends in the volume of physical resources consumed, but may suffer from arbitrary accounting choices and variations in methodology over time.
8. Chapter 5 of the report presents a large amount of material from the datasets used in this study, in particular trends in the volume of activity, costs and survival rates for selected high-volume HRGs. The general pattern is for activity to remain static or decline in the early years of the study, but to recover by the end of the six year period. Trends in costs are more difficult to describe, as much depends on how account is taken of price inflation over the period. For those treatments with significant mortality rates, there is generally an improvement in outcome over the study period.
9. Chapter 6 presents estimates of aggregate measures of outputs and inputs for the programme of care. Over the study period, activity alone (as measured by the cost-weighted activity index) has increased by 3.90% per annum. Adjusting this for the improvement in 30 day mortality rates increases the annual growth to 4.48%, reflecting the major improvement in outcomes over the study period. The experimental use of BUPA health outcome measures for coronary artery bypass surgery (CABG) and percutaneous transluminal coronary angioplasty (PTCA) suggests a further improvement of about 0.2% per annum, but we emphasize that these are highly speculative and partial estimates.
10. Although we are unable to develop measures of physical input growth, we have calculated total reference costs for the programme over the study period. These increased from £1.381 billion to £1.960 billion. Using the GDP index of price change, this implies real growth in expenditure of 5.3% per annum, whilst use of a specific NHS index of price change suggests more modest growth of 2.5% per annum.
11. Measures of productivity change in the hospital treatment of circulatory disease are therefore highly dependent on the measure of input growth used. A very tentative conclusion is that the NHS has used its physical resources more efficiently (to secure annual improvements in physical productivity of up to 2% per annum). However, because of the increased prices it has paid for its inputs, the cost-effectiveness of this programme has declined by anything up to 0.8% over the study period.
12. This study has demonstrated that it is feasible to develop models of productivity growth for a programme of NHS care. This is an important undertaking for informing resource allocation and purchasing decisions in the NHS. Our tentative conclusion is that, whilst there will always be uncertainty in the estimates derived, this represents an important extension of the work in progress at the Office for National Statistics in measuring whole system productivity change, and we advocate further investigation of other programmes of care, in particular those embracing significant community and prescribing activities.
|Copyright, Publisher and Additional Information:||© 2006 the authors. The full text of this report can be viewed free of charge from the Centre for Health Economics web site at: http://www.york.ac.uk/inst/che/pdf/rp21.pdf|
|Institution:||The University of York|
|Academic Units:||The University of York > Centre for Health Economics (York)|
|Depositing User:||Repository Officer|
|Date Deposited:||21 Feb 2007|
|Last Modified:||04 Oct 2010 16:50|
|Publisher:||Centre for Health Economics|
|Series Name:||CHE Research Paper|