Health technology assessment for pharmaceutical regulation in the European Union: do we need another body?

Health technology assessment has long been considered the most comprehensive and suitable approach to support health policy decision-making for specific technologies, pharmaceuticals included, and Europe is no exception. The expected outcomes of this multidisciplinary approach – also acknowledged by the World Health Organization − are recommendations bridging evidence-based medicine and economic evaluation and including the social, legal and ethical implications. ¹

The first move for coordination of health technology assessment in Europe was a project funded by the European Commission in the mid-1990s, ² followed by two other projects which led to the establishment in 2006 of EUnetHTA, a large network of organisations designed to back health technology assessment as a decision-making tool for the European Union and its Member States.

Here we first summarise the health technology assessment experience on pharmaceuticals in the European Union in the very last years. Then, we raise an alternative proposal to that recently issued by the European Commission.

Health technology assessment in the European Union

The European Commission recognised the validity of the EUnetHTA Core Model for health technology assessment analyses in 2016. ² The model includes nine ‘core domains’, which can be split into two groups according to the aspects analysed ³ : (i) ‘technical’ (including technical aspects, safety and clinical effectiveness); and (ii) ‘other’ (economic, ethical, organisational, social and legal aspects).

The European Parliament called on the European Commission to propose legislation on health technology assessment in Europe for assessing the therapeutic value of new medicines in 2017, ⁴ recommending that future European Union rules should be limited to clinical assessment.

Finally, the European Commission released a legislative proposal for health technology assessment regulation in 2018, supporting a European single assessment of relative effectiveness for pharmaceuticals through a Member State Coordination Group, ⁵ at the same time leaving pricing and reimbursement decisions to nations. In practice, Member States would remain responsible for the second group of more context-specific health technology assessment domains. ²

Health technology assessment from theory to practice

The multidisciplinary approach was probably the main factor in the success of health technology assessment, regardless of the real contribution of all the domains potentially explored. For instance, the contribution of ethical aspects has always been important in theory and scant in practice, ⁶ with many more articles on methods than on applications. ⁷ The most likely reason is that it is very problematic to test ethical implications of single technologies empirically as ethics reports to more general issues. ⁸ Roughly the same holds for legal aspects, ⁹ although they have been much less analysed in the literature. In practice, health technology assessment has always been much more narrowly defined in Europe, ¹⁰ and the main outcomes have been relative effectiveness and cost effectiveness analyses.

Relative effectiveness analysis

Relative effectiveness analyses of new therapies should evaluate the clinical relevance of their potential added therapeutic value compared with the best available treatments. ⁴ Randomised clinical trials are unanimously considered the gold standard for assessing both efficacy under ideal conditions and effectiveness in healthcare practice.^11,12 Since randomisation minimises statistical bias on the basis of the play of chance, the primacy of randomised clinical trials for internal validity is well established (and not a matter of opinion); thus randomised clinical trials are still the preferred source for relative effectiveness analyses. However, robust randomised clinical trials designed and based on hard clinical endpoints are very expensive and sometimes hard to organise, ¹³ so less rigorous studies are often considered when no head-to-head randomised clinical trials with the comparators of interest are available. Despite the well-known limits of routine data taken from administrative and health databases, there is now a growing argument in favour of real-world ‘big data’ as a cheap and easy alternative to randomised clinical trials. ¹⁴ More, practical and ethical issues against head-to-head randomised clinical trials are often raised for orphan drugs for rare diseases and drugs for life-threatening diseases when preliminary data suggest a positive clinical effect. However, generating unconfounded evidence from sources other than randomised clinical trials remains questionable, and all the alternative studies may at best strengthen the external validity of the results of the existing randomised clinical trials in the long term, possibly expanding them to subgroups of patients originally not included.

Cost-effectiveness analysis

While relative effectiveness analyses are based on clinical evidence − global by definition − the main added value of cost-effectiveness analyses should be costs, which are dramatically affected by domestic features, so cost-effective analysis results can hardly be extrapolated from one setting to another. ¹⁵ That is why the European Commission guidance does not recommend European cost-effective analyses and leaves them to single countries. Cost-effective analyses themselves actually imply a multidisciplinary approach, with economic and organisational aspects added to the clinical ones employed for relative effectiveness analyses. ¹⁵ In general, different costing methods can lead to different cost-effective analysis results. In particular, unit costs can often be rough proxies of real healthcare costs, ¹⁵ the mainstay of a cost-effective analysis. This happens especially for hospital services – easily the biggest healthcare expense − when they are sourced from tariff lists, which is often the case in European cost-effective analyses. ¹⁶ Health scientists must therefore bear in mind that, unlike measurements of clinical variables, any cost estimate is early open to discretion and can give very different results depending on the accounting method. ¹⁷

The weakness of the present European Medicines Agency role and strategy

Despite almost 30 years of official endorsement, health technology assessment regulation in the European Union has still not taken off in practice. The most encouraging outcome of all the efforts to date still seems to be the collaboration among experts from several disciplines and countries with different expertise, ¹⁸ involved in well-financed projects throughout Europe. However, health technology assessment was ‘sold’ to the European Union as a reliable tool for health policy decision making, ³ otherwise it would have not justified the huge European Union investments.

A common European pharmaceutical market is still far off, and relative effectiveness analyses conducted by national agencies in parallel on the same drug are still common, resulting in duplication of work and inefficient use of resources. We feel that the real crux of the matter in European Union pharmaceutical policy is the present role of the European Medicines Agency for market approval. ¹⁹ This was limited from the outset to risk-benefit assessments of efficacy and safety, with no evaluation of the added therapeutic value of a new drug, which is then entirely devolved to each European Union State’s regulatory authority. This limit has been stretched further by the European Medicines Agency tendency now to fast-track drugs through ‘adaptive licensing’. Once preliminary efficacy and safety have been assessed for market approval, the European Medicines Agency passes the buck to national authorities for relative effectiveness analysis, with too weak evidence for reimbursement decisions. The underlying strategy is fostered by the (fashionable) ‘political creed’ that new drugs must be made available to patients as soon as possible, ²⁰ somehow implicitly attributing an incremental therapeutic value to all of them. New anticancer drugs are a blazing example. ²¹ Their added benefits vary widely and often rely on weak clinical evidence and surrogate endpoints, ²² leading to different decisions in different countries mainly on account of the high level of clinical uncertainty. ²³ Then too, in practice, as national health technology assessment agencies increasingly suffer the lack of robust clinical evidence for relative effectiveness analyses, ² many regulatory authorities slow down and/or restrict the national reimbursement of new (expensive) drugs approved under these fast-track regulatory pathways, because of the scientific uncertainty.^11,24

An alternative scenario for market approval and reimbursement

The very first priority of European pharmaceutical policy that needs to be tackled is assessing the added therapeutic value of new pharmaceutical therapies. This task has been shifted further to post-marketing approval, because of the current European Medicines Agency strategy of expedited market approval ¹⁹ − long on promise and short on delivery. At present, the European Medicines Agency seems to be acting more and more as a sort of highly specialised secretariat, with a growing number of well-paid employees but no drop in national workloads as a ‘trade-off’ ²⁰ ; this situation will be hard for the Member States to sustain in the near future.

We are not convinced that, as recommended by the European Commission, a new health technology assessment organisation would be the right solution for building up common relative effectiveness analyses. A more sensible strategy might be to radically shift the European Medicines Agency’s current trend by reinforcing its role and expanding its tasks to include relative effectiveness analyses. Although we acknowledge that comparative randomised clinical trials with active controls are challenging to conduct and call for very rigorous regulatory scrutiny, an ideal fast, inexpensive and unbiased study design for relative effectiveness analyses simply does not exist. ¹² Therefore, we believe it is time to return as far as possible to head-to-head Phase III randomised clinical trials for market approval of new drugs. This should substantially simplify reimbursement procedures at national level for the regulatory authorities and pharmaceutical companies. ²³ As to patients, we wonder why it should be considered less ethical to participate in randomised clinical trials designed for comparisons with the best treatment than in those designed merely to obtain market approval against placebo in non-inferiority studies. ²⁵ Eventually this might boost research investment towards still unmet needs in the long run, consistently with the European Medicines Agency’s roadmap to 2020.

To conclude, this proposal is aimed at filling the gap of clinical evidence between pharmaceutical market approval and reimbursement among European nations, envisaging a concrete step in the direction of a common, efficient European Union pharmaceutical policy.