Expenditure on pharmaceuticals is rising more rapidly than other components of health care with costs of cancer products growing at 21% per annum in recent years [15]. This will continue with ongoing demographic changes, instigation of stricter clinical targets and the continued launch of new expensive drugs [58]. Increasingly, though there are concerns with the level of health gain of new drugs with a median of just 0.097 Quality Adjusted Life Years (QALYs) versus current standards in an analysis of 281 recent submissions by pharmaceutical companies to the Scottish Medicines Consortium [9]. This compares with the public perception that many new drugs are seen as breakthroughs [10].

European governments and health authorities have introduced successive reforms and initiatives to address these challenging resource issues including funding new premium priced innovative drugs. These include measures to engineer low prices for generics and originators as well as interchangeable brands within a class [2, 1116]. They also include steps to improve transparency in classifying new drugs as innovative, linking the perceived degree of innovation of new products to reimbursed prices, and limiting payer exposure to new expensive drugs given their potential significant budget impact [11, 13, 1723]. Not surprisingly, optimising the use of new expensive technologies is seen as the major challenge faced by health services in Europe as they strive to continue providing comprehensive and equitable healthcare [6].

Current schemes to limit the growth in pharmaceutical expenditure, ensure health gain is maximised within finite budgets, or both including limiting 'off label' prescribing and prescribing outside identified sub-populations where the value of the technology is greatest. They also include price-volume agreements (PVAs), patient access schemes where typically drugs are provided free for a period of time, 'coverage with evidence development' schemes as well as payment by result schemes based on outcomes achieved in practice. These latter schemes are also known as performance-based contracts, efficiency stipulation schemes or effectiveness guarantee schemes [13, 1922, 2435]. All of these schemes have been included within the general term 'risk sharing schemes' either by payers or pharmaceutical companies.

The number and range of risk sharing schemes have grown in recent years no doubt enhanced by patient and physician pressure on governments to accelerate access to new and more costly medicines despite often significant uncertainty surrounding their likely health benefit [10, 36]. This may well be exacerbated by pharmaceutical companies keen to address lost revenues from patent expires, which have recently been estimated at over US$100 bn per year over the next four years [37]. Payers though need to consider the opportunity costs of risk sharing schemes if available resources are not used wisely [36, 38].

Further supply side and demand side initiatives are essential if Europe is to maintain the ideals of socially funded and equitable healthcare. This could mean further expansion of risk sharing schemes. There have been a number of papers published recently that have comprehensively reviewed existing and historic risk sharing schemes offering suggestions for the future [5, 10, 20, 34, 36, 39]. However, there is still considerable confusion surrounding the terminology, which urgently needs to be addressed. There are also concerns among health authority personnel with the level of administrative intensity associated with some of the current schemes and that they could end up contributing substantially to the development costs of new molecules [5, 40, 41]. Certainly in the past, a great deal of the risk associated with outcomes of coverage decisions have been borne by health authorities and insurance companies [34]. This is starting to change given the number of new expensive technologies being launched coupled with their budget impact [6, 23, 34]. These challenges are magnified by the lack of scientific studies evaluating the implementation and outcome of many existing schemes for pharmaceuticals in terms of their overall costs and benefits [10, 34]. This has not been the case with non-pharmacological technologies where 17 policy outcomes have been derived to date from 32 principally non-drug technologies, with in approximately 60% of cases the coverage decision consistent with the outcomes of the study conducted as part of the schemes [34].

Benefits of risk sharing schemes include enhancing health gain within available resources as well as addressing issues such as the safety of new products in practice [10]. These benefits though have to be balanced against concerns for patient care if new drugs are launched too early with considerable uncertainty regarding their safety as well as potentially paying for cost-ineffective technologies (type I error) [36].

We believe these concerns need to be discussed and debated before there is a continuing surge in 'risk sharing' schemes. This paper aims to stimulate this debate by:

  • Proposing logical definitions which can act as a reference for the future

  • Documenting exemplars from the published literature to illustrate proposed definitions supplemented with additional examples from the co-authors

  • Documenting the published impact, concerns and outcomes from current schemes from a payer perspective

  • Summarising considerations that should be borne in mind by payers when contemplating future schemes

  • Proposing guidance on key criteria for considering future risk sharing schemes again from a payer perspective


A literature review was undertaken by one of the authors (BG) in PubMed, MEDLINE and EMBASE between 2000 and February 2010 using key words. These were 'conditional coverage', 'conditional reimbursement', 'risk sharing', 'coverage with evidence', 'price volume agreements', 'value-based pricing, pharmaceuticals', 'no cure no pay', 'pay back schemes', 'health impact guarantee' and 'outcome guarantee'. This was supplemented by unpublished or "grey literature" references known to the 16 co-authors from European countries and locations involved with considering, evaluating and/or implementing such schemes. The focus is principally on Europe although other schemes relating to pharmaceuticals are also documented for Australia, Canada and the US to provide additional examples.

Only papers documenting the nature and content of actual schemes were considered for possible inclusion in this paper. There was no attempt to assess the quality of published papers using established criteria such as the Cochrane Collaboration criteria [42]. This was because this is a 'correspondence' article undertaken by health authority and health insurance personnel evaluating and implementing such schemes rather than a comprehensive review of all risk sharing schemes undertaken as such reviews have recently been published [10, 20, 34].

The schemes were subsequently collated under the proposed definitions to provide guidance to health authority and health insurance personnel as these are the principal intended audience for this paper. Gaps in the knowledge base have been identified with suggestions for the future.

Where possible, Euros are quoted; however, some figures are quoted in GB£s and US $ (€1 = US$1.36, GB£0.90, SEK9.70 - 3 March 2010).


We propose 'risk sharing' schemes for pharmaceuticals should be considered as agreements concluded by payers and pharmaceutical companies to diminish the impact on the payer's budget of new and existing medicines brought about by either the uncertainty of the value of the medicine and/or the need to work within finite budgets. In practice, the agreement lies in setting the scope and realizing the mutual obligations amongst both payers and pharmaceutical companies depending on the occurrence of an agreed condition - the "risk". The "risk" varies by situation, and can include pharmaceutical expenditure higher than agreed thresholds or health gain from a new product lower in practice reducing its value.

We recognise our definition is different to the definition recently proposed by Towse and Garrison - namely these are 'agreements between a payer and a pharmaceutical company where the price level and/or revenue received is related to the future performance of the product in either a research or real-world environment' [36].However our efforts have been directed towards finding a definition coherently overarching the many agreements introduced in previous years as well as one acceptable to health authority and health insurance personnel. Consequently, we believe it is essential to create a definition according to rules of a logical division. These are firstly that there is only one basis of the division. Secondly, every subject belongs to only one group, where there may be sub-categories of equal rank. Thirdly, every example form the superior group, i.e. risk-sharing schemes, has to fall into either category, i.e. either financial based or performance based.

All the schemes discussed and proposed should have a common denominator that is the "risk shared by the payer and the pharmaceutical company". What differentiates the various schemes is the nature of the risk, i.e. "a probable situation in future". Based on our proposed definition, the many existing 'risk-sharing' schemes can be subdivided into either financial/financial-based models or outcome/performance-based models. This is broader than the definition proposed by Towse and Garrison, which just covers performance-based schemes. However, it is a workable definition from a payer perspective.

Even this division, despite being based on common naming, may be misleading though because it suggests that outcome or performance-based models have little to do with expenditure. In fact, the opposite can be true with price-volume agreements and other financial schemes often having little to do with patient outcomes, concentrating more on keeping expenditure within agreed limits and/or enhancing the value proposition of the new product. Price capping schemes do though bring in an outcome element.

The various examples of risk sharing scheme included within the proposed definition are:

  • Financial based schemes:

    • Price-volume agreements (PVAs)/budget impact schemes. These focus on controlling financial expenditure with pharmaceutical companies refunding over budget situations

    • Patient access schemes. These typically involve either free drug or discounts for an agreed period to enhance the value of new medicines and improve the possibility of their funding/reimbursement. Patient access schemes also include price-capping schemes, which focus on controlling the financial impact but from an individual patient perspective. Typically drugs are provided free once patients have exceeded an agreed utilisation limit to again enhance reimbursement/funding within finite resources

∘ Performance based/outcome-based models. These can include schemes whereby companies refund agreed monies or provide free drug if the desired outcomes are not reached. Alternatively, a price reduction if the new drug fails to deliver the desired health gain in practice. In reality, the latter is likely to lead to price shifting from payers to manufacturers as new drugs may well not be able to fully reproduce the desired benefits once prescribed in a wider population than those in the clinical trials, i.e. the net monetary benefit is lower in reality

Existing schemes

As discussed, we believe all existing schemes can be collated into two groups; namely financial/financial-based models or performance-based/outcomes-based models.

Examples of financial-based models include price-volume agreements (PVA) - sometimes referred to as budget impact schemes - and patient access schemes including price capping schemes are included in Table 1[12, 13, 28, 36, 4353], Table 2[20, 29, 33, 34, 5457] and Table 3[5, 1921, 24, 31, 32, 34, 5868]. This is not a complete list but aims to give a number of examples.

Table 1 Examples of Price Volume Agreements (PVAs)/budget impact schemes in Australia and Europe
Table 2 Examples of patient access schemes in Australia and Europe including free drug or discounts
Table 3 Examples of patient access schemes involving price caps in operation in Europe and US

Any further discussion on specific patient access programmes in the US is outside the scope of this paper.

Table 4 gives details of performance-based or outcome-based models in Canada, Europe and USA [10, 19, 20, 2527, 34, 40, 51, 6977]. Again, this is not an exhaustive list but aims at giving a number of examples.

Table 4 Examples of performance-based or outcome-based models in Canada, Europe and US

In addition to these schemes, the Italian Medicines Agency (AIFA) has recently instigated two different approaches to accelerate reimbursement for new drugs especially where there is limited data available at launch. These include variations on patient access schemes as well as performance-based/outcome contracts [5, 34, 51]. The AIFA Oncologic Working Group suggested two risk sharing arrangements for new anti-cancer medicines to enhance their reimbursement potential based on:

  • Epidemiological data for the disease

  • The possibility to clearly define a subset of the population responsive to the treatment

  • Results from the submitted clinical trials

Six products were included in the scheme with a further product added in 2009 [5, 34, 51]. These were:

  • Bevacizumab - Metastatic carcinoma of the colon or rectum, breast cancer, NSCLC and advanced and/or metastatic renal cell cancer

  • Dasatinib - Chronic myeloid leukaemia and acute lymphoblastic leukaemia

  • Erlotinib - NSCLC after failure of at least one prior chemotherapy regimen

  • Nilotinib - Chronic myeloid leukaemia

  • Lapatinib - HER2+ breast cancer patients

  • Sorafenib - Second line treatment for renal cell carcinoma and hepatocellular carcinoma

  • Sunitinib - Metastatic renal cell carcinoma (first and second line treatment)

Inclusion in either risk sharing scheme depended on the characteristics of the drug and the tumour type and included:

  • Erlotinib, sunitinib and sorafenib - provided at 50% discount from current prices for an agreed number of cycles. In the case of erlotinib this is 2 months; 12 weeks for sunitinib and sorafenib. As a result, with erlotinib part funding the 50% of patients who would be expected to have disease progression at or before 8 weeks of treatment

  • Dasatinib and lapatinib - the Italian Health Service fully covers the cost of drugs for the responders following assessment; manufacturers refunding the costs in the case of disease progression

Four new drugs are also currently contained within the Italian conditional reimbursement scheme [34, 51]. This is similar to the US Centre for Medicare and Medicaid conditional coverage scheme under which conditions are set for the continued reimbursement of new technologies [5, 34, 39]. The current scheme in Italy covers ivabradine for chronic angina pectoris, as well as exenatide, sitagliptin and vildagliptin for patients with Type 2 diabetes resistant to current oral treatments [34, 51]. Under this scheme, AIFA fully reimburses the new drugs until further re-evaluation of their actual level of innovation. The main objectives being to evaluate utilisation in routine clinical practice, collect epidemiologic data as well as obtain data on the effectiveness and safety of the new drugs in practice. The measures used to assess effectiveness of the new drugs for diabetes are HbA1c levels. It is the number of angina episodes for new drugs for chronic angina pectoris. By the end of 2008, over 17,000 patients had been enrolled with 7% withdrawals due to therapeutic failure [51]. More recent schemes for new anti-diabetic drugs in Italy have concentrated on co-utilisation with either metformin or sulphonyl ureas. This is because these new drugs should only be prescribed if HbA1c levels are not controlled with existing regimens, or there are unacceptable side-effects with existing drugs at prescribed doses.

Concerns with current schemes

There are already a considerable number of 'risk sharing' schemes in operation globally. However, there are concerns with existing schemes that need be addressed before they should become a routine part of future reimbursement or contracting negotiations especially given the suspicion among payers that a number of proposed schemes are extensions of pharmaceutical company marketing activities.

Whilst price-volume agreements (PVA) shift cost considerations from the payer to the pharmaceutical company, which is important especially if there are concerns that new medicines will be prescribed in a wider population than envisaged, there are concerns that the patients initially prescribed the drug will not always be those most likely to gain the greatest benefit [10, 36]. The growing use of health informatics can help here. In addition, PVA schemes may not always consider issues such as compliance, which is a growing issue with the recent study by Cramer and colleagues showing that only 59% of patients take their cardiovascular and antidiabetic medication for more than 80% of days on therapy each year [78]. There are also concerns that pharmaceutical companies may additionally benefit from early access of new technologies with as yet unproven efficacy and safety even within such schemes unless the value of the new drugs has been critically evaluated by trained professionals. However, it is recognised PVA schemes may be the optimal method to finance new medicines especially where there are currently limited demand side measures to control off-label prescribing or prescribing in patient populations where the new medicine is less cost-effective [36, 79].

Physician Association concerns in Germany with discount and rebate initiatives between Pharmaceutical Companies and the Sickness Funds in Germany include a lack of transparency. This can create problems for the Physician Associations in calculating and monitoring drug budgets as well as providing prescribing guidance to meet their responsibility to monitor and manage drug budgets and advise physicians when their allocated budgets are likely to be breached. This is especially important in Germany as physicians can be fined for being over budget [80]. Physician Association advice can be different to the advice given by Sickness Fund personnel to ambulatory care physicians to help improve the quality and efficiency of their prescribing especially if the value of discussed drugs is altered by discounts given by pharmaceutical companies. Matters can be further complicated when rebates and discounts are given for one drug on condition that there will be a predefined increase in utilisation and expenditure of a second drug.

This is different to the US where combined or bundling of rebates is now generally discouraged by Managed Care Organisations (MCOs) as it impedes transparency. This is particularly important where rebates are passed onto patients. As a result, contracts are increasingly focusing on individual drugs. Proposals for formulary inclusion are also increasingly being considered for drugs where there is support to improve adherence and compliance given concerns that poor compliance could severely compromise effectiveness potentially leading to increased medical and/or drug expenditure in practice to produce the desired clinical outcomes [78, 8183]. MCOs such as Horizon Blue Cross Blue Shield of New Jersey are conducting research to improve the understanding of barriers to enhance compliance, with the findings likely to be incorporated into future initiatives including formulary listing considerations. Having said this, the first consideration for formulary inclusion of new medicines should always be their comparative efficacy and safety versus current formulary drugs. Only when two medicines are seen as essentially similar will the inclusion of schemes aimed at enhancing adherence and compliance be considered for potential formulary listing. Alongside this, 'price protection' clauses are also now being written into some contracts where manufacturers have sought price increases in the past.

Concerns with patient access schemes involving free drug or discounts include opportunity costs, even if this reduces the overall cost/QALY for the new product, as this could still result in significant additional cost for the payers. Alongside this, the administrative burden associated with additional patient monitoring as well as collating and submitting information are also concerns. Concerns with price capping schemes again include the administrative burden such as proof that only appropriate patients have been incorporated into the scheme and that they have reached the agreed limit. This is especially important when the number of patients where these will be applicable in reality is low. For example, there is only one eligible patient in Serbia that would currently qualify under the price capping scheme for bevacizumab in operation in either Italy or Sweden (Table 3). Alongside this, obtaining replacement stock or credit can be a lengthy process especially if manufacturers operate via wholesalers. Such processes may also impact on the accuracy of local medicine utilisation systems making future comparisons difficult. Lastly, pharmaceutical companies may wish to help with data collection and verification, which has to be denied if there are issues with patient integrity.

There are a number of objections with performance-based or outcomes-based models. These include [4, 9, 10, 20, 34, 36, 41, 81, 8490]:

  • Whether the objective of the scheme is fully explicit

  • Whether the level of evidence is sufficient to make robust decisions initially. This includes whether there is a sufficiently strong correlation between any surrogate measures included in the phase III trials and the postulated health gain. This also includes concerns if meta analyses have been undertaken with only indirect comparisons. Limited outcome data also makes it difficult to undertake any sub-group analysis to ascertain potential patient populations where the value of new drugs will be greatest

  • Who will fund additional evidence generation given the high acquisition costs of most new technologies and concerns that considerable expenditure will be spent on marketing activities to promote new drugs rather than spent on further clinical trials

  • Who will fund drug provision if rebated drugs go out of date before suitable patients can be found

  • The reliability of registry and other data generation schemes unless comparable control groups. Concerns can be diminished if subsequent studies are undertaken by independent organisations

  • Accelerating the uptake of new drugs in practice, i.e. including patients in registry studies or provisional reimbursement schemes may accelerate their uptake in practice. The utilisation of new drugs may also be greater if companies provide additional support such as nurses to help with case finding and monitoring. This may be balanced though through reducing 'off label' use if physicians know prescribed indications are being monitored.

  • The length of follow up - especially if this is long (β interferon scheme in the UK - below)

  • The general administrative burden for all key stakeholders. This also includes the costs associated with the instigation of additional evaluation units such as those proposed by NICE in England. The PASAG (Patient Access Scheme Assessment Group) is already established in Scotland to help evaluate proposed risk sharing schemes; similarly the Patient Access Scheme Liaison Unit within NICE. Both units are already adding to the administrative burden of current risk sharing schemes

  • Compliance - especially for long term chronic conditions. This issue must be fully addressed where pertinent else all key stakeholder groups will lose out

  • Pricing, i.e. pharmaceutical companies may be tempted to initially over price their new drug in expectation of price cuts/cost shifting downstream as the evidence base grows

On the other hand, pay-for-performance schemes do encourage pharmaceutical companies to develop biomarkers or other methods that help target patient populations where health gain and hence value is greatest. This is especially important for new cancer drugs given the limited aggregate health gain of most new drugs including those for cancer and their appreciable cost per patient [9].

There have been a number of concerns with the β interferon scheme for multiple sclerosis in the UK, which need to be considered by payers when evaluating future schemes [25, 40, 9193]. These include:

  • The model

  • Flaws in the actual model due to difficulties in fully mapping out the quality of life and natural history of MS to the trial outcomes, which were based on changes in EDSS scores (Expanded Disability Status Score)

  • Concerns that the model was heavily influenced by assumptions about future discounting and did not account for example for the cost of azathioprine

  • The model did not appear to fully account for patients discontinuing treatment early because of side-effects

  • The length of follow-up

  • Concerns that within ten years the β interferons and glatiramer acetate may have been supplanted by newer drugs reducing the whole rationale behind the scheme

  • Funding and administration support

  • Primary Care Trusts generally did not receive any additional funding to cover the cost of these drugs

  • Hospitals also did not receive additional funding for more extensive follow-up consultations and for completing the necessary forms reducing their involvement in practice

  • Concerns generally with the necessary infrastructure required including specialist nurses as well as concerns where the costs of the additional administrative burden would come from

The administrative burden, lack of communication, and concerns with passing on savings have all been highlighted as key issues with current schemes for cancer drugs in the UK [41]. Recent research among hospitals yielded [41]:

  • 73% of hospitals reported they did not have the capacity to manage current schemes as these typically required additional staff to manage, co-ordinate and track them. This is especially the case if hospital personnel have to spend valuable time manually tracking patients, retrospectively adjusting stock control systems and ensuring necessary financial systems are in place to fully realise any savings

  • A need for greater flexibility around the time limits for processing claims

  • A need for good communication between key stakeholder groups, e.g. in the case of bortezomib every missed claim results in a loss of GB£12,000

  • The need to ensure savings are passed back to the payers - this is not happening in 47% of hospitals

These findings again highlight concerns with existing schemes that need to be adequately addressed in the future before risk sharing schemes become a routine part of future reimbursement considerations.

Conclusions and future proposals

We believe there are only a limited number of situations where 'risk sharing' schemes should be considered in the future, as well as key issues that need to be considered by payers before entering into future 'risk sharing' arrangements. These are summarised below and have arisen due to the increasing launch of new expensive technologies putting considerable strain on European healthcare systems. They do not apply to the classical PVA/budget impact schemes, which have been discussed earlier.

Situations where risk sharing schemes could be considered by payers in the future include where:

  • the objectives and scope of the scheme are explicit and transparent

  • the new drug is a novel treatment in a high priority disease area with expected net health gain, and there are currently few if any effective treatments. In addition, translational science suggests good effectiveness in reality and delaying treatment before all the outcome data is available may not be in key stakeholder's interest. This can subsequently be verified by independent studies

  • new drugs are seen as effective in priority disease areas but there are potential concerns with long term safety

  • new drugs could have a substantial beneficial impact on service delivery and patient safety in practice but it has been difficult to prove this within the confines of a phase III trial

  • the likely health gain can be determined within a limited time frame. This is especially important in patients with advanced cancers in order not to waste time and resources

  • proposed patient access schemes in priority disease areas substantially lower health service costs to enhance reimbursement having factored in all administrative costs. This must though take into account the possibility of rebated drugs going out-of-date before they can be used or whether current systems have the ability to fully track patients and potential rebates, e.g. the cost of missed claims with bortezomid (above). Overall, rebate schemes do appear to provide a more accurate record of drug usage than the provision of free drug, as well as help ensure only appropriate patients are prescribed the new drug, enhancing their appeal

We believe proposed risk sharing schemes should be rejected where:

  • effective and low cost standards already exist for the population in question with proven long term outcomes. This includes provisional reimbursement schemes which may encourage companies to launch new expensive products with only limited surrogate data

  • Health Authorities will end up funding a substantial proportion of a new drug's development costs

  • patient compliance is a key consideration to improve health, and this has not been adequately addressed in the proposed scheme

  • there is a high administrative burden for the health service versus the likely health gain and/or perceived financial benefits

In addition, proposed schemes must be based on robust evidence ('coverage with evidence') for potential consideration. This includes robust translational science which has suggested beneficial outcomes in reality. Schemes must also include unambiguous and easily measured 'evidence based' effectiveness criteria evolved from good biological research and comprehensive clinical trials. Where there is insufficient evidence to make robust decisions, we believe reimbursement or funding should initially be considered only at similar prices to current standards. Prices can subsequently be increased in all or sub-populations as more data becomes available demonstrating increased value in practice. In this way limiting the financial exposure of payers as well as limiting Type 1 errors.

We believe key considerations when payers are reviewing proposed schemes include transparency, ethical considerations, staffing, evaluation and exit strategies. Transparency and ethical considerations include:

  • Funding arrangements for any registries as well as administration costs must be transparent

  • Any ethical, legal and clinical governance considerations must be fully addressed when proposing and developing future schemes. This includes issues of ownership of the data, especially if schemes are operated within health services, intellectual property rights and opportunities for appeal

  • Future risk sharing schemes should be open to all pertinent companies and not just selected companies. This builds on the established procurement processes, and could be in the form of 'requests for risk sharing' schemes among competing companies in given disease areas/patient populations

Staffing/funding considerations include whether:

  • Appropriate professionals are in place to fully evaluate proposed schemes, otherwise decisions may be compromised. This includes clinical, clinical pharmacology, pharmacy, IT and economic experts

  • Continued funding of competent personnel (clinical and IT), as well as comprehensive IT infrastructures, is in place to effectively develop and implement follow-up registries where pertinent

  • There are high ethical standards in the evaluation of proposed schemes. This includes the declaration of any contacts and conflicts of interest between experts and pharmaceutical companies that could potentially jeopardise evaluations

Finally, we believe evaluations must be built into future schemes based on good science given the paucity of published studies to date and concerns with many current schemes. In the past, pharmaceutical companies may have been unwilling to broadcast their schemes especially patient access schemes. However, there are now established payer networks to address this, and publication is seen as a major way forward for health authorities to learn from each another. Any evaluation must include the overall costs involved with implementing and conducting the schemes as well as their outcome. 'Exit' strategies must also be considered in advance should the effectiveness and/or safety of new drugs turn out to be worse in reality leading to their possible withdrawal during the lifetime of the scheme.

Consequently we believe future schemes must have realistic time scales, must not involve appreciable administrative burden if part of routine clinical care unless addressed, and must take cognisance of any likely changes in care during their lifetime, i.e. standard drugs losing their patent and/or clinical standards changing.

We hope that this paper will stimulate further debate and discussion especially in Europe about future risk sharing arrangements for pharmaceuticals amongst all key stakeholder groups.