Health Economics: 8 - The role of economic evaluation and priority setting in health care decision making.
8.1 The uses of economic evaluation
Economic evaluation is used to varying degrees in different countries. How widespread its use is, and the purposes for which it is used, depend to a great extent on the country’s dominant health system, whether public, social insurance or private insurance based. In developing countries, this also depends on the extent to which the country is in receipt of aid from international bodies. Morris, Devlin, Parkin and Spencer (2012) classify the reasons why economic evaluation is used in health care decision making as:
- To maximise the benefits from health care spending.
- To overcome regional variations in access.
- To contain costs and manage demand.
- To provide bargaining power with suppliers of health care products.
Of these, the most widespread is to contain costs and manage demand through a system which controls the introduction of new technologies, particularly those that are high cost technologies, through technology assessment. High cost does not necessarily mean that their unit cost is high; technologies are seen as expensive if the overall cost of using them in their target patient population is high. As a result, technology assessment has been dominated by pharmaceuticals. Many countries now use some form of economic evaluation in deciding whether their national health systems should fund particular pharmaceutical products. Broadly, the issue is whether the system will reimburse patients or service providers if they use a drug.
In England, the assessment of new health technologies takes place within the National Institute for Health and Care Excellence (NICE). This was established in 1999 to address geographic variations in access by providing national-level guidance on the effectiveness and cost effectiveness of different interventions in the NHS. NICE therefore has an equity goal in ensuring that everywhere in the NHS provides any new technologies that are cost-effective; however, by implication it also means that any new technologies that are not cost-effective will not be adopted.
NICE produces guidance in three areas: public health, health technologies and clinical practice. Public health guidance concerns the promotion of good health and the prevention of ill health and is intended not only for the NHS, but also for local authorities, the wider public sector and the voluntary sector. Health technology guidance concerns the use of new and existing medicines, treatments and procedures within the NHS. Clinical practice guidelines concern NHS treatment and care for people who have specific diseases and conditions.
NICE provides a large amount of information on all of its work and how it is carried out, which is available on its website www.nice.org.uk including details relevant to the evaluation of public health interventions. The other countries of the United Kingdom have arrangements that differ to a varying extent from those in England. Details may be found on the websites of their national health services, and for Scotland: www.scottishmedicines.org.uk
8.2 Cost effectiveness league tables and threshold cost-effectiveness
As discussed in section 5.3.4, the results of an economic evaluation of an intervention are typically expressed as an Incremental Cost-Effectiveness ratio (ICER). Section 5.3.5 discusses how this may take the form of a cost per Quality-Adjusted Life Year (QALY) gained. However, this simply shows what the cost is of buying a QALY using that intervention. This ICER does not in itself tell us whether the intervention is good value for money. One way in which value for money might be assessed is to see what the cost per QALY gained is of other interventions. When economic evaluations were first conducted and published, it became routine to include such comparisons in the form of a list of different interventions that had been appraised along with their ICERS. This evolved into the suggestion that all interventions that have been evaluated might be compared in the form of a ‘league table’, in which interventions are ranked from lowest to highest cost per QALY gained (Mauskopf, Rutten and Schonfeld, 2003).
Problems with league tables include inconsistency in the methods used for different evaluations, difficulties in transferability of results, the assumption that point estimates of ICERs determine rankings, and the implication that all interventions in the table are independent alternatives. It is also not clear how they should be used by decision-makers. It may seem reasonable to say that if an intervention is more cost-effective than another that is already funded, it should also be funded. But that simply raises the question of whether this comparator was itself cost-effective. In general, it is not possible to make any such decisions without reference to a threshold, an actual value which acts as the criterion for decision making - if an intervention’s ICER is below this it should be funded, if above it should not. (This threshold is the same concept as the ceiling ratio referred to in section 5.3.4.)
Various threshold values have been proposed. A commonly quoted value, especially in the United States, is US$50,000, though this is entirely arbitrary. NICE in the UK has an explicit threshold, though a complicated one, which also has no basis in theory or evidence. There are two ways in which such thresholds could be determined. One is the social value of a QALY approach, establishing what value society places on each QALY gained, so that the cost of obtaining a QALY by an intervention can be compared to how much society is willing to pay for one. The other is the shadow price of a QALY, which is the value of the least cost-effective intervention that just exhausts the health budget.
There have been some attempts to establish league tables. The WHO/World Bank have promoted this, particularly in the context of developing countries, in the form of ‘generalised cost-effectiveness analysis’. However, the most comprehensive attempt to use this as a means of priority setting was in Oregon in the United States. Problems with Medicaid, the publicly funded programme for poor people led Oregon to set up a Health Services Commission to develop what became the Oregon Health Plan. One aspect of this was a systematic and explicit basis for deciding which health services should be included, and which excluded, from the Medicaid benefit package list. All possible interventions are classified into ‘condition-treatment’ pairs and a list is produced of them in order of priority. They are funded in order of priority until the health budget is exhausted. Originally, the ranking was determined by cost-effectiveness, but this was so controversial that it was never used in practice. Rankings were determined by five-year survival and improvement in symptoms and the type of service, whether curative or not, for example.
NICE’s threshold is not defined with respect to a single value and involves considerations other than cost-effectiveness. It states that interventions with an ICER below £20,000 will be approved and those above £30,000 are unlikely to be approved. Above £20,000 other considerations are taken into account. However, although the other considerations are listed, it is clear neither how they are factored in to decision-making nor how they are treated differently above and below £30,000. This is entirely discretionary and there is no way to tell if it is applied consistently across interventions. Also, NICE has on significant numbers of occasions approved interventions with ICERs above £30,000 and occasionally not approved interventions with an ICER below £20,000.
8.3 Programme Budgeting and Marginal Analysis
Programme Budgeting (PB) and Marginal Analysis (MA) are two quite different planning tools that have become accepted as a consistent approach called Programme Budgeting and Marginal Analysis (PBMA). It is a pragmatic approach to resource allocation that attempts to divide decision making about priorities into a manageable set of activities (Mitton and Donaldson, 2001).
8.3.1 Programme Budgeting
One of the aims of PB is to focus the process of setting and using budgets on the outputs of health services, rather than the more traditional focus on the purchase of health care inputs. PB involves assigning costs to health care programmes, defined in terms of the types of output produced, which in practice is often organised by client or disease groups. The first stage in PB is to describe current resource allocations in this way, by showing how much is spent in each programme. It is therefore possible to see what budgets are currently being used for, rather than simply the services that are purchased. The budgets include everything that is spent on the programme in a particular time period, usually a year, including both revenue and capital, wherever it is spent.
PB may be undertaken at the macro level, which looks at the whole of expenditure, divided into large programmes, for example disease groups, or at the micro level looking at expenditure within these programmes, for example different care settings for a particular disease group, which involves a finer level of detail of information about services.
In theory, PB is part of a comprehensive planning system for decision making based on the intervention provided. However, in practice it refers only to the information that is provided, along with suggestions as to how that information might be used.
PB has been used in many countries, at national and local level, and has also been used internationally by the WHO. In the UK, a National Programme Budget Project was initiated by the Department of Health in 2002. All CCG expenditures on hospital, community and primary care services are mapped to 23 programmes of care based on medical conditions. Details of this can be found on this website:
https://www.england.nhs.uk/resources/resources-for-ccgs/prog-budgeting/
This information can be used for accountability purposes, assessing whether explicitly stated priorities are met. It can also be used as a basis for setting targets for changes in expenditure in line with priorities. If collected over time, it can be used to monitor the extent to which such targets are met.
8.3.2 Marginal Analysis
Section 1 described marginal analysis as a general economics approach to optimisation. However, in the PBMA context, Marginal Analysis (MA) means something specific. The idea is that once programme budgets have been established, decision makers will want to maximise the achievement of the aims of the programme given the budget allocated to it. This requires economic appraisal of health care at the level of service provision.
MA is therefore the appraisal of the added benefits and added costs of a proposed investment in health care, or the lost benefits and lower costs of a proposed disinvestment. Its basis is not a comparison of the costs and benefits of all uses of resources but an evaluation at the margin of different ways of using resources. MA can be used to analyse marginal changes between programmes, though it is much easier and more common to apply it within programmes. Although MA uses standard approaches to economic appraisal, it does have some distinctive features, mainly the screening of new services for potential investments and of existing services for potential disinvestments, often through an expert or stakeholder group.
8.4 Cost of illness and burden of disease
Cost of illness (COI) and burden of disease (BOD) studies are widely undertaken in many countries. Although it would be possible to outline principles that would distinguish between these, in practice there are no consistent definitions of them. The individual terms are therefore best reserved for the titles given to particular projects rather than using them to describe a particular type of study. Both aim to assess the overall impact of an illness - or in some cases all illnesses - on society, with the term cost used in a broad sense to mean not just the costs in terms of the resources used in dealing with a disease, but also the value of resources lost due to disease, and the value of lost health.
COI/BOD studies are descriptive, and their aims are usually expressed in terms of the potential uses of the information that they provide. First, by assessing the size of health problems, they indicate one aspect of its relative importance, and are therefore an input to priority setting, both in terms of the resources that should be devoted to health care and the resources that should be devoted to health care research. Secondly, they provide useful background or baseline data for economic evaluations of specific health care interventions, and are particularly valuable for identifying different types of cost (including health deficits) that might be affected by an intervention. Their aims are therefore similar in many ways to PBMA.
As with economic evaluation, there are different ways in which they can be measured, and there are guidelines for good practice. Because of its basis in population health and illness, COI/BOD has strong links with public health and epidemiology, and one of the key issues is whether the basis of costing is incidence or prevalence. Factors that affect this choice are whether the disease is chronic or acute and the practical availability of one or other measure.
On the economics side, COI/BOD studies tend to use a characterisation of costs that derives from an older tradition of cost-benefit analysis that has been considerably modified for use in economic appraisal of particular interventions. Costs are divided into direct costs, indirect costs and intangibles. An important issue is whether valuation is based on ‘avoidable costs’, which essentially means the value of resources used from a provider’s perspective or ‘willingness to pay’, which essentially means market valuations by consumers.
Many COI studies are funded by organisations that have a special interest in particular diseases, such as patient support and lobbying groups and pharmaceutical companies that have products in that area. There are also some local examples that cover individual countries. However, the most well-known, wide-ranging and ambitious example is the Global Burden of Disease study, which is funded and promoted by the World Bank and the World Health Organisation as a key element of its aid for health services in developing countries. One important feature of this project is the development of the Disability Adjusted Life Year (DALY), also discussed in Section 5.5.1, to capture the health impact problems in a way that would facilitate comparison of the size of a given problem across disease areas and between countries. DALYs, like QALYs, combine length and quality of life, but unlike QALYs they describe a deficit in health from a theoretical maximum of no disability for life with a maximal life expectancy. Further details of this project are on this website: http://www.thelancet.com/gbd
Although COI and BOD studies are widely carried out, they are controversial and, in particular, many economists are highly critical of them (Byford, Torgerson and Raftery, 2000). It is alleged that COI/BOD tells us nothing useful for policy or decision-making purposes. All that COI/BOD gives us is information on the size of a problem, which is irrelevant for priority setting. Priority setting should assess priorities based on effective and cheap solutions to problems rather than how big they are. COI/BOD gives information about what would be the result of removing a problem, but the more usual decision making context is of making marginal improvements to a problem. It is alleged that COI/BOD distorts resource allocation rather than improves it, leading to inefficiency and inequity. The argument that COI/BOD is an essential prerequisite for economic evaluation studies is persuasive, but this purpose requires only a costing study, not an estimate of global impact. Similarly, although COI/BOD might be used to set priorities for medical research, it is still the case that these need to be evaluated according to the marginal costs and benefits of research activities in different disease areas.
© David Parkin 2017