Mixed palette

Nearly 20 European countries now claim to use cost-effectiveness information when making reimbursement decisions on new pharmaceuticals. The practice began in Australia in 1992, followed by Canada, but it received a major boost in Europe with the founding of the National Institute for Health and Clinical Excellence (NICE) in England and Wales in 1999. Economic evaluation had been used for some years before in the UK but NICE integrated cost-effectiveness analysis into a formal decision process, with a standard analytical approach and the clear objective of facilitating the cost-effective use of healthcare resources.

The NICE process was not the first, but it has been seen as the most comprehensive in the prescriptiveness of its procedures and the extent of its consultation with stakeholders during an appraisal. Variants on the NICE model have been adopted in other tax-funded health systems, but now the major insurance-based health systems of Europe, namely France and Germany, are introducing cost-effectiveness analysis into their historically very different decision-making processes.

This article examines the use of economics in the health systems of the UK, Germany and France and draws out some lessons for those trying to address the European market with a single product evidence dossier.

Different emphasis
All three systems are concerned with the efficient use of health budgets, by focusing on the clinical benefits and costs of using pharmaceuticals, but their emphases are slightly different, partly owing to the timing and nature of the decisions they make. In France, the Haute Autorité de Santé (HAS) reviews all drugs at the time of launch. In Germany, the Institute for Quality and Efficiency in Health Care (IQWiG) looks at a selected number of drugs after they have been in use for some time, although there are new proposals for earlier and more extensive review. NICE began by reviewing only a limited number of drugs, but now it appraises most new drugs within a short time of their UK launch. NICE also undertakes multi-product appraisals of treatments for a particular disease and patient group, and HAS is beginning to use cost-effectiveness analysis in a similar way.

In Germany and the UK, the main decision by NICE or IQWiG relates to reimbursement status. Product prices are set by the companies in advance of the appraisal, although in Germany the rate of reimbursement may be limited by the reference pricing system. In France, reimbursement status (and the proportion of price reimbursed) is decided on the clinical evidence, but the actual price on which reimbursement is calculated is the subject of negotiation between the company and the Economic Committee.

Although NICE has no remit to negotiate on price, companies now have the opportunity to introduce a 'patient access scheme' (PAS), which reduces the cost to the NHS in particular circumstances. NICE must carry out an assessment of the impact of a PAS on the cost-effectiveness of a product, but the schemes themselves can only be approved by the Department of Health. HAS and IQWiG consider reimbursement for drugs in their licensed indication. NICE may recommend the reimbursement of a product in a small segment of the patients for whom it is licensed, if cost-effectiveness can only be demonstrated in this group.

Because of these differences in the types of decision made, the emphasis on particular decision criteria varies. NICE decisions are based on the merging of the clinical benefit and cost data in the cost-effectiveness framework. NICE is the only organisation to demand outcome data, in the form of quality adjusted life years (QALYs), and has a threshold range against which it assesses cost per QALY ratios. In France, the negotiation on price makes the overall impact on the healthcare budget a major factor in decisions. The German approach also gives more weight to budget impact on health insurance organisations.

Cost effectiveness
NICE uses a standard approach to cost-effectiveness analysis, which is generally accepted by academic health economists in Europe. The main deviation from this is the use of a National Health Service (NHS) and social care cost perspective, rather than a societal perspective. This relates partly to the terms of the legislation by which NICE was established: its remit covers the efficient use of NHS budgets. NICE has adopted the principle of 'best available evidence' in making decisions, rather than applying a quality hierarchy and using only the results of certain types of study. Modelling is seen as an essential part of the assessment process, which is focused on the need to make decisions.

As stated, NICE has an explicit cost-effectiveness threshold range, which it uses to formulate recommendations. Generally, a drug with an incremental cost-effectiveness ratio (ICER) of less than £20,000 per QALY is likely to be approved. Those with an ICER between £20,000 and £30,000 can be approved, but must demonstrate contribution to other government health policy goals, such as reducing inequalities in health, to justify the higher cost. Products with an ICER in excess of £30,000 are unlikely to be approved.

Although the absolute level of the ICER is important, the quality of the evidence on which it is based and the uncertainties in the calculations are also taken into account. This is especially the case in the Single Technology Appraisal (STA) format, which is now commonly used, as the decision is based on a review of the evidence submitted by the manufacturer, with no alternative independent model produced. On the occasions when NICE has rejected technologies with apparently low ICERs, a major issue has been the reliability of the evidence.

The NICE methods are constantly under review and two changes have eased the path to approval for some products. These are PAS and special decision criteria, introduced for products that benefit patients being treated at the end of life. It is now possible for products, usually cancer treatments, which offer an average survival gain of at least three months to patients with no more than two years of expected life, to be approved with an ICER of around £50,000 per QALY. This recognises that a QALY may not be regarded by society as having the same value in all circumstances, and opens the door to other adjustments to the value of benefits.

Another issue highlighted by the new UK Government is the use of value-based pricing (VBP) to ensure that the costs of drugs to the NHS reflect their value to patients. It could be argued that the current NICE process follows this principle already, as the approval decision is based on the ICER, which reflects the relative benefit (value) of products. The restricted approval of some products for certain sub-groups of patients is also in line with this approach. The fact that UK price levels are not determined independently of those in other European countries limits the scope for full VBP in a single country. One interesting development may be more use of the 'only in research' recommendation and a form of conditional reimbursement while better research evidence is collected.

French system
The French system uses a sequential process, which rates the clinical benefits of new products against those of existing treatments and determines reimbursement prices on perceived relative value. New products giving a significant additional clinical benefit, known as Amelioration du Service Medical Rendu (ASMR), can achieve more rapid reimbursement approval at premium prices. Products with no significant ASMR cannot expect to be reimbursed at a level higher than existing treatments.

The clinical and economic decisions are made without the explicit use of cost-effectiveness analysis. However, HAS has established a specialist, multi-disciplinary committee to carry out multiple product appraisals for whole classes of drugs, in which 'community benefit' must be considered. This is done using a more traditional cost-effectiveness framework.

Rejection
In Germany, the use of cost-effectiveness analysis with QALYs as the measure of benefit has been rejected. IQWiG examines each disease area separately and compares the treatments in that disease area with one other using disease specific benefit measures. A graphical presentation of the cost and value of each treatment, beginning with the most cost-effective, can be used to draw an efficiency frontier. The slope of the frontier declines as each additional treatment is less cost-effective than the last.

Once the value/cost relationship is known for a new product, it can be compared with the frontier in different ways. For example, it can be compared with the least cost-effective product already reimbursed, and if it is demonstrated to be at least as good, it can be given reimbursement. A stricter criterion would be to require it to be as good as the most cost-effective existing product, ie with a slope equal to the steepest part of the frontier.

Criticism
This approach has met with considerable criticism from the academic health economics community. The choice of a single indicator, which captures all the outcomes and adverse events in a disease area, may be difficult. QALYs were designed to capture the multiple dimensions of patient outcomes. Sourcing clinical benefit data from only randomised controlled trials (RCTs) may limit the relevance of the data to actual practice and may exclude some treatments.

Depending which decision criterion is used, the approach may depend on the assumption that all past decisions were efficient and reflective of the social weighting of health benefits in differing contexts. Proponents of the approach say that it fits well with the structure and financial mechanisms of the German healthcare system, in which explicit decisions to move resources between disease areas and patient groups are not taken centrally.

Central dossier
From this brief review, it can be seen that a central dossier, which provides local marketing companies with all the data and arguments needed to satisfy the reimbursement authorities will be an extensive document. In spite of the work of projects like EUnetHTA, the development of a standardised approach to the use of economics in the reimbursement decision process in Europe is some way off. Further, approaches at national level continue to diverge, with regional variations emerging in decision making in Italy, Spain, Sweden and the UK.

Nevertheless, there are some common elements to a dossier, which have general relevance. The core clinical data on the new product will apply in all countries. Efforts continue to exploit the success in the harmonisation of licensing decisions to help remove unnecessary variations in the reimbursement decision process. Though comparator treatments will vary locally and costs will always be system-specific, the basic model structure needed for analysis will be determined by the nature of the disease. Central development of these key elements will continue to be valuable.

Defining value
As many stakeholders accept the concept of the value-based price the spotlight turns on to the definition and measurement of value, but the quickest way to change the cost-effectiveness of a drug is to change its price. Keeping pricing and reimbursement decisions separate, when both are comparing costs and benefits though an economic framework, will become increasingly difficult.

The Author
John Hutton is Professor of Health Economics, York Health Economics Consortium, University of York, UK.

To comment on this article, email pme@pmlive.com