Past CHE seminars 2008

15 December 2008

Title: The contribution of smoking and obesity to socioeconomic inequalities in health in England
Speaker: Dr Stephen Morris, Reader, Health Economics Research Group, Brunel University.

Abstract: There is a growing interest in the reduction of avoidable inequalities in health in England, and in policy adjustments to reduce these inequalities between areas. In this paper we investigate the extent of health inequality in England and the contribution of two causes of avoidable morbidity and mortality, smoking and obesity, to this inequality. The aims of the paper are to measure socioeconomic inequalities in health both within and between areas of England over a nine year period from 1998 to 2006, and to quantify the contribution of smoking and obesity to this inequality. Our analysis is based on nine rounds of the Health Survey for England (n=100,317). We construct a health variable using the predicted values from an interval regression model regressing self-assessed health against a comprehensive set of health indicators, and using EQ5D scores to set the cut-points on the self-assessed health categories. We then use a concentration index approach to measure national health inequality over a nine year period and the proportion of this that is due to health inequalities within and between government office regions of England. We use inequality decomposition methods to quantify the contribution of smoking and obesity to health inequality both nationally and within regions. Our analysis shows that national income-related health inequality appears to have remained fairly stable over time. There is greater health inequality, and lower health, in the North of England compared with the South, and inequality within regions accounts for a greater proportion of total income-related inequality than inequality between regions. Smoking and obesity account for a small but statistically significant proportion of income-related health inequality and the extent of this is greater in regions with higher health inequality.

4 December 2008

Title: Is clock-watching productive? The effect of hours harmonisation on hospital doctor outputs
Speaker: Matt Sutton, Professor of Health Economics, Health Methodology Research Group, School of Community Based Medicine, University of Manchester.

Abstract: Background: Previous studies have demonstrated very substantial variations in the activity rates of hospital consultants in the UK. The new contract for hospital consultants introduced across the UK from April 2004 involved greater harmonisation of hours worked across consultants. This was intended to increase overall productivity but has been dubbed ‘clock-watching’.

Aim: To show that increased productivity could have been achieved by increasing mean input, by a mean-preserving reduction in input variation and/or targeting of increases in input on individuals with higher marginal products. To estimate the effect of hours worked on the number of patients treated by hospital doctors and examine whether the change in the distribution of hours worked increased aggregate output.

Data: Data on 330 hospital consultants from surveys undertaken in 2001 and 2006 linked to hospital administrative records.

Methods: Econometric models for panel data produce marginal product estimates for individual doctors.

Results: Levels of output are highly variable across individual doctors and show marked persistence over time. Hours are a significant determinant of output. Across individual doctors, there is a weak positive correlation between changes in hours worked and the marginal product of hours.

Conclusions: In isolation, the harmonisation of hours produced by the new consultant contract led to an increase in the aggregate number of patients treated. Wide variation in output across individual consultants suggests there is considerable scope for further increases in productivity.

Date: 25 November 2008

Title: Measuring Hospital Inefficiency: Do Different Methods and Measures of Patient Severity of Illness and Quality Make a Difference?
Speaker: Professor Michael D Rosko, Professor of Health Care Management, Widener University, Chester Pennsylvania, U.S.A.

Abstract: This presentation reviews the impact of employing different methods and measures of hospital quality and patient severity of illness on the mean estimated inefficiency and relative ranking of hospitals generated by stochastic frontier analysis (SFA). The study included urban U.S. hospitals in 20 states operating in 2001. We used hospital data for 1,290 hospitals from the American Hospital Association Annual Survey and the Medicare Cost Reports. We employed a variety of controls for hospital quality and patient severity of illness. Among the variables used were a subset of the quality indicators generated from the application of the Patient Safety Indicator and Inpatient Quality Indicator modules of the Agency for Healthcare Research and Quality, Quality Indicator software to the Healthcare Cost and Utilization Project (HCUP), State Inpatient Databases. Measures of a component of patient burden of illness came from the application of the Comorbidity Software to HCUP data. SFA was used to estimate hospital cost-inefficiency. We tested key assumptions of the SFA model with likelihood ratio tests. Two-stage estimation was compared against the preferred simultaneous estimation procedure.

Our results indicate that methods sometimes matter (i.e., different conclusions are reached when two-stage and simultaneous estimation procedures are used). However, results tend to be robust across variations in assumptions about the structure of technology and the distribution of the error term.

Results suggest that the measures produced by the Comorbidity Software appear to account for variations in patient burden of illness that had previously been masquerading as inefficiency. Outcome measures of quality can provide useful insight into a hospital’s operations but may have little impact on estimated inefficiency once controls for structural quality and patient burden of illness have been employed.

Choices about controlling for quality and patient burden of illness can have a non-trivial impact on mean estimated hospital inefficiency and the relative ranking of hospitals generated by SFA.

Key Words. Hospital efficiency, stochastic frontier analysis, hospital quality, patient safety, severity of illness

6 November 2008

Title: Designing contingent valuation scenarios for environmental health: The case of childhood asthma
Speaker: Sylvia Brandt, Assistant Professor, Department of Resource Economics, University of Massachusetts, Amherst.

Abstract: Valuation of morbidity associated with childhood asthma is significant both to policy and to non-market valuation methodologies. Our results show that household perceptions and beliefs, such as belief in one’s ability to predict and control asthma attacks, and relative perceptions of the overall burden asthma places on a family, have a larger impact on valuation than traditional measures of asthma severity. More generally, our approach can be applied to other chronic illnesses as well, such as diabetes or chronic pain.

9 October 2008

Title: The impact of market structure on GPs’ referrals to specialist care
Speaker: Professor Tor Iversen, Director of Research HERO, Institute of Health Management and Health Economics, University of Oslo, Norway.

Abstract: There is a considerable variation in referral rates of general practitioners (GPs). Typically, this variation cannot be fully explained by variation in the composition of listed patients. We study to what extent variation in market structure has an impact on the number of referrals to radiology examinations that a GP makes. In particular, we are interested in whether or not competition for patients among GPs influences their role as gate-keepers. We find that an increase in excess capacity among GPs has an unintended cost in terms of an increased number of referrals. To what extent patients benefit from the increased cost is unknown.

2 October 2008

Title: State priorities, economic systems and allocations of resources to health in the USSR and Russia
Speaker: Dr Christopher M Davis, Reader in Command and Transition Economies, University of Oxford, and Fellow, Wolfson College.

Abstract: Priority setting in health based on economic evaluation and political economy factors has been insightfully assessed in Klein, R., Day, P. and Redmayne, S. (1998), Hauck, K., Smith, P.C. and Goddard, M. (2004) and Goddard, M., Hauck, K., Prekker, A. and Smith, P.C. (2006). However these authors primarily examine health sectors in market economies with democratic governments in advanced and developing countries. This seminar presentation will examine the different cases of priority setting and implementation related to health in the command (shortage) economy of the USSR during 1965-91 and in the transition economy of Russia in the periods of economic collapse (1992-99) and recovery (2000-08). It will review the experiences of the Putin and Medvedev governments in using priority programmes to channel supplemental resources to meet urgent needs in health, such as prevention and treatment of HIV/AIDS. The presentation will be based on past and ongoing research of the presenter (Davis 1989, 1990, 2001).

Related Reading

Davis, C. (1990) “National health services, resource constraints and shortages: a comparison of Soviet and British experiences” in Manning, N. And Ungerson, C. (1990) Social Policy Review, London, Longman

Davis, C. (1989) “Priority and the shortage model: the medical system in the socialist economy” in Davis, C. and Charemza, W. (1989) Models of Disequilibrium and Shortage in Centrally Planned Economies, London, Chapman and Hall

Davis, C. (2001) “The health sector: illness, medical care and mortality” in Granville, B. and Oppenheimer, P. (2001) Russia’s Post-Communist Economy, Oxford, OUP

Goddard, M., Hauck, K., Prekker, A. and Smith, P.C. (2006) “Priority setting in health – a
political economy perspective”, Health Economics, Policy and Law, 1:79-90

Hauck, K., Smith, P.C. and Goddard, M. (2004) “The Economics of Priority Setting for Health Care: A Literature Review”, World Bank HNP Discussion Paper

Klein, R., Day, P. And Redmayne, S. (1998) Managing Scarcity: Priority Setting and Rationing in the National Health Service, Oxford, OUP

22 September 2008

Title: Variation in end-of-life care in the U.S: a biopsy of health system performance
Speaker: David C. Goodman, MD MS. Professor of Pediatrics and of Community and Family Medicine, The Dartmouth Institute for Health Policy and Clinical Practice; Co-Principal Investigator, The Dartmouth Atlas of Health Care, Hanover, New Hampshire. USA

Abstract: The U.S. health care system is faced with the formidable challenge of improving access, quality, and outcomes without spending a greater proportion of GDP. Previous studies by Wennberg and colleagues have shown wide regional variation in capacity (i.e. hospital beds and physicians per capita) and utilization unrelated to patient wants or needs. Fisher has further demonstrated that higher intensity of health care services in the elderly is not associated with better technical quality or lower mortality.

This seminar will present recent analysis from the Dartmouth Atlas of Health Care group that extends studies of unwarranted variation through an analysis of the last two years of life in patients > 65 years with chronic illness. These hospital specific cohorts are highly standardized and address previous concerns about case-mix differences across regions. The patterns of care revealed though this study design provide a detailed view of substantial differences in care across identifiable health care systems, with the outcome held constant. The measures correlate highly with more general patient populations of the hospital region.

The end-of-life findings provide information that can help guide health system improvement, reimbursement schemes, and medical workforce policy.

15 September 2008

Title: Quality or quantity? Markets or management? Evaluating the performance of public hospitals in the New Zealand reform period
Speaker: Professor Peter Davis, Social Statistics Research Group, University of Auckland, New Zealand.

Abstract: In the last ten years, government funding of health in New Zealand has nearly doubled in nominal terms. And yet, if we take the media reports at face value, the health system is still beset with the usual stock of shortages and quality issues. Is this accurate? Can we expect more of our health system? Or is it just a matter of rising public expectations? The core of this paper is an assessment of the New Zealand health reforms of the 1990s, with data on hospital productivity and patient outcomes. The results may surprise, and could have a lesson for our current discontents. There are also implications for debates on the future of the NHS, particularly in the light of both recent and mooted reform.

The Papers Abstract can be downloaded here

4 September 2008

Title: Evaluating a public health intervention: Is your home radioactive, is it worth fixing, and how should we decide?
Speaker: Professor Alastair Gray, Health Economics Research Centre, Dept. Public Health, Oxford University.

Abstract: The remit of NICE has recently been extended to include public health interventions, the evaluation of which may present particular methodological and practical difficulties. A good example is control of radon gas, the second most important cause of lung cancer in the UK. Policies to reduce radon problems in homes are currently under review by the Health Protection Agency and DEFRA, providing an opportunity to inform decision makers with economic analyses. In this presentation I will review the analyses I have performed for this review, and discuss the difficulties of adhering to a standard NICE technology appraisal in this type of evaluation, when the costs and benefits are very unevenly spread across different agencies and population groups.

28 July 2008

Title: Are you what you eat? Experimental evidence on health habits and economic behaviour
Speaker: Matteo Galizzi. University of Brescia and Centre for Health Economics, York. An Alan Williams visiting fellow.

Abstract: In order to analyze the interrelations among individual behaviour in strategic interaction, psychological and cognitive attitudes and health and life style habits, we run a pilot experiment on a sample of 150 subjects at EXEC experimental economics laboratory in York to whom we administrate a two-hours session of both questionnaire and experimental tests. First, we consider a salient questionnaire to self-assess health characteristics and life style, containing detailed qualitative and quantitative questions on individual behaviour in nutrition, eating and drinking, smoking and physical exercise. Based upon such self-assessed data, we construct a number of individual health and nutritional indexes, including the Healthy Eating Index computed according to the USDA official guidelines. We then combine the nutritional and health indexes with psychometric tests for a number of individual profiles including, among others, risk aversion, time preferences, overconfidence - in terms of mis-calibration, illusion of control and “better-than-average” effect -, impulsiveness, self-monitoring, cognitive reflection. Finally, we replicate a repeated version of the Trust game. Micro-econometric analysis is then developed, through cross-section and panel estimations, for assessing whether and how nutritional and life style habits may be correlated with psychological and cognitive profiles, and which, among health indexes and individual profiles, may represent statistically significant determinants of economic behaviour.

11 July 2008

Title: Health technology assessment in Australia
Speaker: Professor Jane Hall, Professor of Health Economics and Director, Centre for Health Economics Research and Evaluation Faculty of Business, University of Technology, Sydney.

Abstract: Many countries are developing health technology assessment processes which evaluate new technologies in terms of safety, effectiveness and cost-effectiveness. Australia was the first country to introduce a mandatory requirement of economic evaluation linked to public funding as part of its national pharmaceutical scheme. There have substantial developments since this was introduced in 1993, notably the development of a similar approach for funding new medical services and procedures. Although the medical services approach was modeled on the pharmaceutical process, there are interesting and distinct differences. The Pharmaceutical Benefits Advisory Committee (PBAC) requires the pharmaceutical company applicant to prepare the evaluation, admits unpublished evidence, and provides only summaries of its decisions. Its recommendations to the Minister are binding in that a new drug cannot be subsidized through the Pharmaceutical Benefits Scheme without a positive recommendation. In contrast, the Medical Services Advisory Committee prepares its own evaluations, restricts evidence to published results, and publishes full reports. Its recommendations do not bind the Minister.

The diffusion of new technologies and the application of formal HTA within public hospitals is different again. Medical services and pharmaceutical benefits are funded by the national government under open-ended (fee for service) arrangements. The funding of public hospitals, another major area of expenditure, is more readily subject to budget constraints. While this should provide a stronger impetus for assessing the efficiency (cost-effectiveness) of new technologies, in practice the approach is more diverse and less structured.

I am a member of the Medical Services Advisory Committee.

3 July 2008

Title: Using the time trade off to value health states that are ‘worse than being dead’
Speaker: Professor Nancy Devlin, Department of Economics, The City University, London.

Abstract: The York MVH TTO protocol has been widely replicated in studies to elicit values for EQ-5D states, and its value set is routinely used in economic evaluation. Yet it has a known problem: the procedure it uses to elicit values for states 'worse than dead' (have a value < 0) is fundamentally different, both conceptually and empirically, from that used to value states better than dead (have a value > 0). It is questionable whether values > 0 and < 0 can meaningfully be aggregated; the procedure also produces arbitrarily extreme negative values that require equally arbitrary transformation to -1. The aim of our study was to develop and test a TTO protocol that overcomes these and other issues and to explore the characteristics of the valuation data it generates. In this seminar we report results from the use of a 'lead time time trade off' to value EQ-5D health states, and assess its merits and limitations as a valuation approach. Our results include findings from a pilot to test the effect of alternative durations on valuation.

30 June 2008

Title: Advertising and strategic entry deterrence in pharmaceutical markets
Speaker: Ismo Linnosmaa, Senior Lecturer, Department of Health Management and Policy, University of Kuopio.

Abstract: In this study we examine the incentives of incumbent firms selling branded pharmaceuticals to deter entry of generic pharmaceuticals by means of advertising. In the pharmaceutical market this possibility arises due to the first-mover advantage that the patent protection gives to the producers of branded pharmaceuticals. First results of our theoretical analysis show that the entry of the generic pharmaceutical can be blockaded, deterred or accommodated by the advertising strategy of the incumbent firm. Accommodated entry takes place when the cost of entry for the generic firm is low and advertising is costly. When the entry cost of the generic firm increases and advertising becomes less expensive, the incumbent firm either deters or blockades the entry of the generic pharmaceutical. Welfare analysis indicates that the entry of the generic pharmaceutical is socially optimal if the additional utility that patients obtain from the consumption of the generic pharmaceutical exceeds the fixed entry cost of the generic firm and possible net reduction in welfare associated with advertising. The comparison of social optimum and the market equilibrium points show that, although the entry of generic pharmaceutical also takes place in the market, the market equilibrium is inefficient due asymmetric and possibly excessive advertising caused by the first-mover advantage of the incumbent firm.

Date: 5 June 2008

Title: Designing regimes of hospital performance assessment to have an impact: characteristics, evidence & implications
Speaker: Professor Gwyn Bevan, Professor of Management Science,Department of Management, London School of Economics & Political Science.

Abstract: Results of systematic reviews of the evaluation of regimes of hospital performance assessment that have concluded that there is little or no evidence that these result in improvement from the two expected pathways. One is that patients use information on hospital performance as consumers to switch hospitals; the other that hospital staff are keen to use such information when it shows there is room for improvement. There is, however, evidence of a third pathway to improvement: that people running hospitals do respond to regimes that are designed to inflict reputational damage on those shown to be performing poorly. Unfortunately such systems can have unintended consequences. Hence there are interesting questions about design of regimes of performance assessment and the generation of high-powered incentives.

22 May 2008

T itle: Health and health care in Iran: successes and contradictions OR retreading the silk route: a travellers tale
Speaker: Professor Roy Carr-Hill, University of York

Abstract: Iran has the usual pre- and post-transitional problems of newly middle income countries; indeed their death rates from traffic accidents would be the envy of a WalMart undertaker. The fulcrum of the ‘axis of evil’ also has a unique health care system from which not only the States could learn; but so could we. A major peculiarity is their model of ‘integration’ between the Medical Schools and the Ministry of Health; but their primary health care in the rural areas – not in urban areas - is a WHO Alma Aty dream. Although they collect large amounts of data both routinely and through surveys, for anyone interested in inequalities, these databases are sadly underused.

8 May 2008

Title: Evaluating patient self-management interventions: some reflections on methodological challenges and opportunities
Speaker: Professor Stirling Bryan, Professor of Health Economics, University of Birmingham.

Abstract: Patient self-management and self care feature are key components of recent UK government policy, especially in the context of supporting people with long-term conditions. Department of Health policy documents suggest that self care and self-management can improve health outcomes, increase patient satisfaction and “empower patients to take more control over their lives”. If the principal benefits come in the form of greater autonomy and control for patients then conventional approaches to capturing health benefits in economic evaluation are unlikely to perform well.

This paper reports an economic evaluation study conducted alongside a clinical evaluation of a self care intervention: patient self-management of anticoagulation (warfarin) therapy (the SMART trial). The cost-utility analysis in the SMART study concluded that self-management was not a cost-effective use of resources, with no clear increase in QALYs and significantly higher costs. A willingness to pay approach was additionally used to capture some of the broader “control and autonomy” features of the intervention. The home-based equipment used in the clinical study was made available for patients to purchase at the end of the research and so data were gathered on both stated (hypothetical willingness-to-pay) and revealed (purchase decision) preferences. The data indicate some strong preferences for the self-management approach despite no additional health benefit being evident.

The paper also reports on the methodological comparison of stated and revealed preferences, with encouraging findings for the hypothetical willingness-to-pay approach in this self care context.

28 April 2008

Title: An investigation of the relationship between clinical quality in primary care and hospital re-admissions for Type 2 diabetes related complications: A multivariate failure time data analysis
Speaker: Giuliana de Luca. Department of Economics and Statistics, University of Calabria, Italy.

Abstract: It has been suggested that timely and effective primary care can reduce hospitalizations for diabetes specific and related complications. For instance, intensive blood glucose control may significantly reduce the risk of cardiovascular and micro-vascular complications.

Recently the UK NHS has implemented a number of policies intended to improve the quality of chronic disease management in primary care, and thereby reduce hospital admissions for patients with chronic diseases. These include the Quality and Outcomes Framework, Evercare case management, and the Patients at Risk of Readmission tool.

We investigated the relationship between clinical quality in general practices and hospital readmissions for patients with Type 2 diabetes. A cohort of over 4000 patients aged ≥18, from 60 GP practices in England, hospitalized with a diagnosis of a Type 2 diabetes complication between 1 April 1997 and 30 March 2006 were identified using the Hospital Episode Statistics database and retrospectively followed up. Practice level quality indicators for diabetes care were available from the Quality Assessment in Primary Care study for the years 1998, 2003 and 2005.

A Cox proportional hazard model for recurrent events was estimated to assess the impact of quality delivered in primary care on the risk of re-hospitalizations for diabetes related complications. The primary study outcome was the time in days from discharge following the first Type 2 diabetes related admission to the next and subsequent readmissions for a diabetes-related complication. Preliminary analysis suggests that better clinical quality was associated with a reduction in the likelihood of readmission.

10 April 2008

Title: Estimating the returns to the investment in cardiovascular research in the UK: a preliminary analysis
Speaker: Professor Martin J Buxton, Professor of Health Economics, Director: Health Economics Research Group, Brunel University.

Abstract: 'This talk will present the context and some initial findings from a study being undertaken jointly between HERG, RAND Europe and OHE. This research has been commissioned for a consortium, of the Medical Research Council, the Wellcome Trust, and the Academy of Medical Sciences, and sets out to critically examine methods for, and results from, analysis of the returns to medical research. It will present some new estimates of CVD research spending, and an initial analysis of the health gain in QALYs that might reasonably be attributed to that research, allowing for likely lags involved, and the international dimension to research in this area.

6 March 2008

Title: Estimating country-specific cost-effectiveness from multinational clinical trials
Speaker: Andrew Briggs, Lindsay Chair in Health Policy & Economics Evaluation, Public Health & Health Policy, University of Glasgow.

Abstract: This talk will examine some of the challenges and the suggested solutions to estimating regional cost-effectiveness results from clinical trial data, using recent examples of multinational clinical trials with an associated cost-effectiveness analysis. In particular, fixed and random effect approaches, that have been proposed in the literature, will be compared to an alternative approach based on separate estimation of the components of the cost-effectiveness calculus. In addition to analytic issues, consideration will be given to some of the practical challenges that analysts have to address, such as generating country-specific unit costs, translating to a common currency, and handling missing data.

11 March 2008

Title: The First Step to Measuring Health Capital Trends in Barcelona, 1992-2006: Quality of Life Weights Estimation
Speaker: Anna Garcia-Altes, Barcelona Institute of Public Health.

Abstract: The objective of this seminar is to present the results of the estimation of quality of life weights in Barcelona for 1992, 1994, 2000, 2002, and 2006, based on self-assessed health status reported in the Barcelona Health Survey.

Quality of life weights were estimated using an ordered probit model. In this model, self-assessed health status was related to the presence of chronic diseases, demographic characteristics (age and sex), instruction level, and a random error. Also, the possibility of interactions with age and chronic limitations in usual activities was allowed. Quality of life weights were derived by normalizing the regressors obtained.

Quality of life weights related to chronic diseases varied depending on the diseases. Quality of life decreased as age increased, and increased as educational level increased. For the same disease and adjusted for age, sex, and educational level, quality of life weights increase over time.

The proposed methodology allows quality of life weights to be calculated from health survey data, which has direct application in economic assessment, and analysis of socioeconomic health inequalities. The next step would be to add life years gained in the 1992-2006 period, and estimate health capital trends.

7 February 2008

Title: The impact of health insurance on access to care: evaluation of a social experiment in rural China
Speaker: Professor Winnie Yip, Adjunct Associate Professor of International Health Policy and Economics, Department of Population and International Health, Harvard, Cambridge, MA, USA.

Abstract: Rural Mutual Health Care (RMHC) is a community-based health insurance intervention which provides first dollar coverage for both inpatient and outpatient services, and introduces a number of supply-side interventions aimed to improve quality and reduce inefficiencies in health service delivery. The seminar will report on an evaluation of RHMC, examining its effect on access to health care and comparing the impacts of RMHC to an alternative government-supported program which combines a medical saving account and hospital insurance with high deductibles.

The evaluation method combines differences-in-differences with Propensity Score Matching. The data draw upon longitudinal household/individual surveys conducted one year pre-, and two years post-intervention. RMHC has had a positive impact on health care utilization (especially at the village clinic level), increasing the probability of an outpatient visit by 70% and reducing the probability of self-medication by similar percentages. Further, there is evidence of spillover effects in which non-enrollees of the RMHC sites increased the probability of a village visit and reduced self-medication. In contrast, there is very little impact of the government-sponsored program.

7 January 2008

Title: Debate on discounting: has NICE got it wrong?
Speakers: Hugh Gravelle and Karl Claxton, Centre for Health Economics, University of York.

Abstract: The way in which the costs and health effects of a healthcare project are discounted can be crucial in calculating its incremental cost-effectiveness ratio. NICE, along with almost all other regulatory and official bodies, calculates the ICER by using the same discount rate for costs and for health effects. Two recent papers in Health Economics (Claxton et al, 2006; Gravelle et al, 2007) disagreed about whether this approach is valid or whether a lower discount rate should be used for health effects. We will present, discuss, and refine the arguments and attempt to discover why such distinguished economists reached different conclusions about discounting.

Claxton K, M Sculpher, A Culyer, C McCabe, A Briggs, R Akehurst, M Buxton, J Brazier. Discounting and cost-effectiveness in NICE - stepping back to sort out a confusion. Health Economics 2006; 15:1-4.
Gravelle H, Brouwer W, Niessen L, Postma M, Rutten F. Discounting in economic evaluations: stepping forward to optimal decision rules. Health Economics 2007; 16:307-317.