Past CHE seminars 2009

3 December 2009

Title: Inconsistent health state valuations
Speaker: John Cairns, Professor of Health Economics, London School of Hygiene & Tropical Medicine

Abstract: Inconsistent valuation of health states, where higher scores are assigned to what appear to be unambiguously poorer health states, is unsatisfactory for health economists in a number of ways. If the scoring system or tariff contains such inconsistencies a move to a better health state could be associated with a loss of QALYs. This would tend to undermine any claim that economic evaluation of health care interventions can usefully inform health care decision making. Inconsistent valuations by individuals are also unsatisfactory because they might suggest a failure of the valuation process to capture the individual’s preferences. As the number of inconsistent valuations an individual provides increases doubt increases regarding the values elicited. Has the individual understood and fully engaged in the task?

This study has two aims: to examine a number of hypotheses regarding the determinants of logical inconsistency; and to explore the effect of individual inconsistency on the overall tariff for valuing health states. The empirical analysis uses a database of TTO valuations of the EQ-5D elicited from a sample of the Thai general population. Such an enquiry is important because it could give some insight into the treatment of inconsistent valuations when trying to establish a tariff with which to value health outcomes. It might, for example, suggest whether inconsistencies arise largely from a failure in understanding. Also knowledge of what generates more or less inconsistency could be useful in the design of studies to elicit health state preferences.

5 November 2009

Title: Budget allocation and the revealed social rate of time preference for health.
Speaker: Karl Claxton, Centre for Health Economics, University of York.

Abstract: Appropriate decisions based on cost-effectiveness evaluations of health care technologies depend upon the cost-effectiveness threshold and its rate of growth as well as some social rate of time preference for health. The concept of the cost-effectiveness threshold, social rate of time preference for consumption and social opportunity cost of capital are briefly explored before the question of how a social rate of time preference for health might be established is addressed. A more traditional approach to this problem is outlined before a social decision making approach is developed which demonstrates that social time preference for health is revealed through the budget allocations made by a socially legitimate higher authority. The relationship between the social time preference rate for health, the growth rate of the cost-effectiveness threshold and the rate at which the higher authority can borrow or invest is then examined. We establish that the social time preference rate for health is implied by the budget allocation and the health production functions in each period. As such, the social time preference rate for health depends not on the social time preference rate for consumption or growth in the consumption value of health but on growth in the cost-effectiveness threshold and the rate at which the higher authority can save or borrow between periods. The implications for discounting and the policies of bodies such as NICE are then discussed.

5 October 2009

Title: The VPF-QALY relationship and end-of-life treatment: should there be a 'pure value' of life?
Speaker: Shepley Orr, Department of Civil, Environmental and Geomatic Engineering, and Centre for Philosophy, Justice and Health. UCL

Abstract: The problem of how to determine the monetary value of a QALY has recently become a central concern in health economics. In addition to the threshold-searcher approach, and the willingness-to-pay for a QALY, another approach has been to infer the monetary value of a QALY by deriving values from the value of preventing a statistical fatality (VPF) figure as used by the Department for Transport. This paper examines the VPF-QALY relationship from both directions: inferring the value of a QALY from the VPF, and the possibility of inferring the value of the VPF from the monetized value of a QALY. Our approach asks whether either approach is acceptable. In particular we ask whether the monetized value of health should value only the stream of health provided (i.e., QALYs gained), or whether we should value life itself. Some have argued that such a “pure value of life” should be taken into account in any VPF-QALY relationship, and similar concerns seem to underlie the recent approval of deviations from the NICE threshold for end-of-life treatments. This paper considers the issues arising in relating VPF and QALY monetary values, and whether a “pure value of life” is normatively justified.

28 September 2009

Title: Inequalities in Health and Health Care in Serbia: An Inverse Health Care Law; but also Inverse SMR!
Speaker: Professor Roy Car-Hill, Centre for Health Economics, University of York.

Abstract: As part of work to develop a resource allocation formula for funding the decentralized Serbian Primary Health Care System, it was necessary to identify reliable and valid measures of need at decentralised geographical levels. In particular, the associations between a wide variety of proxy need indicators and inequalities in health (as measured by mortality) and inequalities in health care (as measured by current budgetary allocations and staffing provision) were examined.

The analysis of variations in health care showed that there are large inefficiencies in the provision of health care with rural areas over-staffed and large metropolitan areas relatively under staffed. These variations (inequalities) are largely unrelated to income or mortality, generating an Inverse Health Care Law.

The analysis of variations (inequalities) in death rates and SMRs showed that they are unrelated to income or any other measure of poverty; and specifically that inequalities in death rates across Serbia were unrelated to inequalities in income across Serbia.

This paper goes on to explore three possible hypotheses as to how this apparent anomalous situation might have arisen:

Cross border migration and internal displacement of people, following the break up of the former Yugoslavia; and/or tight bureaucratic control over internal migration;
Most of the poverty measures are based on information/records about those employed in the formal sector, and these are a poor proxy for the actual level of poverty;
Alcohol, for an Eastern European country, is relatively expensive and so is more available to the wealthier.

Roy Carr-Hill & George Boulton

7 September 2009

Title: Sensitivity analysis in economic evaluation: what policy impact? what do policy makers say?
Speaker: Stirling Bryan, Associate Director, Centre for Clinical Epidemiology & Evaluation, Vancouver Coastal Health Research Institute, Canada.

Abstract: Objectives: Cost-effectiveness analyses (CEAs) are increasingly guiding health care reimbursement decisions, and so methodological rigour, including the appropriate characterisation of uncertainty, is essential. This research sought to: define ‘good practice’; audit sensitivity analyses; understand their policy impact; and elicit policy maker views. This paper focuses mainly on the latter two objectives.

Approach: A sample of CEAs (n=15), undertaken to inform recent technology appraisal judgements by the National Institute for Health & Clinical Excellence (NICE), was audited to report the sensitivity analysis methods used. NICE policy/guidance documents relating to the 15 topics were reviewed, with reference to any aspect of the sensitivity analyses noted, and an assessment made of the ‘policy impact’. Further, qualitative data from interviews with NICE appraisal committee members were analysed, using a thematic approach, to understand how policy makers use sensitivity analysis and to assess the value they attach to that component of the CEA.

Results : Sensitivity analysis practice is highly variable and often not reported clearly. Parameter ranges in deterministic analyses and parameter distributions in probabilistic sensitivity analyses (PSA) are rarely justified, and the choice of distribution varies between analyses. Uncertainty is discussed explicitly in policy guidance documents. The predominant focus is parameter uncertainty. The cost-effectiveness ranges cited in policy documents are mainly from deterministic analyses. This might reflect poor understanding of PSA or the value attached to deterministic approaches, especially in sub-group searching. Analyses demonstrating high levels of uncertainty tend to be associated with negative reimbursement decisions. The interview data echo support for deterministic analyses as highlighting decision-critical model parameters. Strong support was also expressed for PSA. Communication of results is sub-optimal; more detail and greater clarity is required.

Conclusions: Both deterministic and probabilistic analyses have their place; the former being particularly helpful in the search for sub-groups. The association between high levels of uncertainty and negative decisions requires further discussion. The challenge is one of effective communication between analysts and policy makers; highlighting training and presentational issues.

3 September 2009

Title: Capability measurement within health care: theoretical and methodological issues.
Speaker: Joanna Coast, Professor of Health Economics, School of Health & Population Sciences, University of Birmingham.

Abstract: The standard welfarist and non-welfarist approaches to economic evaluation are to estimate outcomes either in terms of utility (willingness-to-pay) or health. This seminar explores an alternative method which follows the capability approach developed by Amartya Sen. The seminar will explore the theoretical perspective brought by the capability approach, and the extent to which it can be utilised in economic evaluation. It will then move on to consider work currently being undertaken to develop means of estimating capabilities for use in economic evaluation, focusing particularly on the development and use of the ICECAP index of capability for older people, and the more general index of capability currently being developed for use in an adult general population.

26 August 2009

Title: Long run returns to education: does schooling lead to an extended old age?
Speaker: Owen O’Donnell

Abstract: While there is no doubt that health is strongly correlated with education, whether schooling exerts a causal impact on health is not yet firmly established. We exploit Dutch compulsory schooling laws in a Regression Discontinuity Design applied to linked data from health surveys, tax files and the mortality register to estimate the causal effect of education on mortality. The reform provides a powerful instrument, significantly raising years of schooling, which, in turn, has a large and significant effect on mortality even in old age. An extra year of schooling is estimated to reduce the probability of dying between ages of 81 and 88 by 2-3 percentage points relative to a baseline of 50 percent. High school graduation is estimated to reduce the probability of dying between the ages of 81 and 88 by a remarkable 17-26 percentage points but this does not appear to be due to any sheepskin effects of finishing high school on mortality beyond that predicted linearly by additional years of schooling.

2 July 2009

Title: Using the EQ-5D as a performance measurement tool in the NHS
Speaker: Professor Nancy Devlin, Director of Research, Office of Health Economics, London.

Abstract: In a landmark move, the UK Department of Health (DH) is introducing the routine use of Patient Reported Outcome Measures (PROMs) as a means of measuring the performance of health care providers in improving patient health. From April 2009 all patients will be asked to complete both generic (January 14, 2010y for four elective procedures; the intention is to extend this to a wide range of other NHS services. The aim of this paper is to report analysis of the EQ-5D data generated from a pilot study on PROMs commissioned by the DH, and to consider the implications of the results for the use of EQ-5D as performance indicators and measures of patient benefit. The EQ-5D has the potential advantage in the context of PROMs of enabling comparisons of performance across services as well as between providers; and in facilitating assessments of the cost effectiveness of NHS services. We present two new methods we have developed for analysing and displaying EQ-5D profile data: a Paretian Classification of Health Change, and a Health Profile Grid. Using these methods, we show that EQ-5D data can readily be used to generate useful insights into differences between providers in improving overall changes in health; as well as differences between surgical procedures. We conclude by highlighting a number of issues that remain to be addressed in the use of PROMs data as a basis for performance indicators.

4 June 2009

Title: Incentives and targets in hospital care: evidence from a natural experiment
Speaker: Frank Windmeijer. Professor of Econometrics, University of Bristol.

Abstract: Performance targets are commonly used in the public sector, despite their well known problems when organisations have multiple objectives and performance is difficult to measure. It is possible that such targets may work where there is considerable consensus that performance needs to be improved. We investigate this possibility by examining the response of the English National Health Service (NHS) to waiting time targets. Long waiting times have been a key issue for the NHS for many years. Using a natural policy experiment exploiting differences between countries of the UK, supplemented with a panel of data on English hospitals, we examine whether high profile targets to reduce waiting times met their goals of reducing waiting times without diverting activity from other less well monitored aspects of health care. Using this robust design, we find that targets led to a fall in waiting times without apparent reductions in other aspects of patient care.

7 May 2009

Title: How have waiting time targets in the UK's National Health Service affected waiting list management by hospitals and surgeons?
Speaker: Sofia Dimakou. City Health Economics Centre, Economics Department, City University, London.

Abstract: Waiting times for elective surgery in the United Kingdom's National Health Service (NHS) have been a key policy and political concern for many years. Waiting time targets have been the main means in recent years of dealing with this issue, and these targets have been made more stringent since 2000, reducing from 18 months between acceptance on the waiting list and admission, to the most recent target of 18 weeks between referral from a general practitioner and admission. The targets have been successful, in that NHS institutions have in general been able to meet them. However, it is less clear how this has been achieved and what the impact of this has been on the distribution of waiting times over all patients.

This paper investigates how hospitals and individual surgeons have responded to targets and have managed to reduce waiting times. It uses duration analysis applied to data covering every surgical operation carried out in the NHS for 9 years. Our previous work has established the usefulness of this approach to these kind of data, and this new work extends the analysis in two important ways. First, for the first time it allows analysis of waiting times for individual surgeons and individual referring general practitioners. Secondly, it examines how waiting times have evolved over shorter periods of time rather than simply on an annual basis. It also covers a longer time period.

Specifically, we analyse Hospital Episode Statistics (HES) data of around 25 million hospital admissions from 1997 to 2006 for three specialties - general surgery, trauma and orthopaedics and ophthalmology. These data include information on waiting times, patient characteristics, diagnosis, operation, admitting hospital and consultant and referring GP practice. We are able to analyse survival and hazard functions over time to show how changing waiting times relate to changing targets. These reveal great variations between hospitals and individual surgeons in the way that their waiting lists have been managed and have been affected by targets.

20 April 2009

Title: Comparative effectiveness in a world where ‘individual results may vary’
Speaker: John Mullahy, Professor of Population Health Sciences, University of Wisconsin-Madison, USA.

2 April 2009

Title: Assessing the influence of gestalt-type characteristics on preferences over lifetime health profiles.
Speaker: Dr Adam Oliver, RCUK Senior Research Fellow, LSE Health and Dept. Social Policy, London School of Economics.

Abstract: Introduction
In contrast to the basic tenets of economic theory, there is substantial evidence that people’s remembered and predicted utility of events systematically differs from the utility that they experience. These systematic differences are caused by ‘gestalt characteristics’. The objective of this study was to test whether people maximize ‘quality-adjusted life years’ (QALYs), or whether QALY maximization is compromised by them being influenced by factors that resemble the gestalt characteristics when choosing between lifetime health profiles.

Methods
Time trade-off values were elicited from fifty respondents, who were also presented with a series of hypothetical questions that each depicted two lifetime health profiles. The respondents were asked to choose which of the two profiles in each question they would prefer to experience. By inputting the values that the respondents placed on the health states into the lifetime health profiles, it was possible to observe whether their answers were consistent with QALY maximization, or with various hypothesized gestalt-type effects.

Results
Across decisions that involve a simple trade-off between the length of life and the quality of the health state, choices consistent with QALY maximizing were relatively common, although even here approximately half of the respondents violated this rule. Consistency with QALY maximization was lower in most of the other tests, and indicated that many people might, for example, prefer to trade off some lifetime health in order to experience a good end to life, or to avoid highly unstable lifetime health profiles.

Conclusion
The respondents’ answers were often consistent with the hypothesised gestalt-type effects, but it is probable that for some of the questions the characteristics themselves were not driving the respondents’ answers, and that factors such as complex rates of discounting might have played a role. However, whatever the driving motivation behind the respondents’ answers, the important point to note from this study is that QALY maximization is often substantially and systematically violated when people are offered a choice over the lifetime health profiles that they would prefer to experience.

5 March 2009

Title: Key principles for the conduct Of Health Technology Assessments: how does NICE stack up?
Speaker: Mike Drummond, Professor of Health Economics, Centre for Health Economics, University of York.

Abstract: Since its inception in 1999, NICE has been widely criticised, both from within the UK and without. However, it is relatively rare to see a comparative analysis of NICE and similar institutions worldwide. Is NICE as good as it gets, or can it learn from institutions in other jurisdictions? Recently, 15 key principles for the conduct of HTA were proposed, covering issues such as the structure of HTA programmes, the methods of HTA, the processes for conducting HTA and the use of HTA in decision-making (Drummond et al IJTAHC 2008;24:244-58). In this presentation the principles are used to assess the performance of NICE, as compared with that of other HTA programmes in Europe, North America and beyond.

5 February 2009

Title: The relative societal value of health gains to different beneficiaries
Speaker: Richard P Edlin. Lecturer in Health Economics, Academic Unit of Health Economics, Leeds Institute of Health Sciences, University of Leeds.

Abstract: Publicly funded health care systems aim to balance multiple objectives, including the maximization of population health and furthering notions of equity and fairness. An analytical tool that can be used to balance the possibly conflicting objectives is the health related social welfare function, defined as a function of population health.

A recent project (Dolan et al, 2008) elicited preferences from members of the general public and analysed these preferences within a SWF framework. The arguments of this SWF are a re-weighted QALY (the adult healthy year equivalent) considering both the degree and timing of ill-health. The project suggested that a year of full health to a child contributes twice as much lifetime health as a year of full health to an adult. Whilst there is some evidence that treating those in very poor health receives greater priority, it appears that most priority is given to treating those in poor-but-not-very-poor health. The parameters of the SWF are defined using equity, responsibility and rarity arguments. It appears that there is a strong preference towards obtaining more equal health, with small weights on responsibility and rarity.

This presentation will outline the project research from the preparatory stages to the main results and discuss possible implications for resource allocation.

12 January 2009

Title: If you want to assess hospital performance, consider every patient
Speaker: Andrew Street, Professor and Assistant Director of the Health Policy Team, Centre for Health Economics, University of York.

Abstract: Hundreds of studies have been published about hospital efficiency. Their practical impact has been minimal though because no two hospitals are alike, making comparisons unreliable, and because the analysis typically fails to indicate where corrective action is required.

Previously I have recommended comparing specialties rather than hospitals and analysing data about the individual patients treated in each specialty. Acting on these recommendations has been hindered by the absence of patient-level cost data. Recently, though, we have matched data in the Hospital Episode Statistics for every patient treated in English hospitals to the reference cost data reported by each hospital. I demonstrate how these combined data can be exploited by analysing all one million patients admitted to English obstetrics departments in 2005/6.

Analysis of this nature can be used to make more robust and focussed performance assessments; to fine-tune prices under Payment by Results; and to indicate where refinements to the Healthcare Resource Group classification system might be required. I outline the challenges involved in extending this research to other specialties and across longer time periods.