How proven is a ‘proven intervention’? Ethics of placebo controls in light of conditional approval programs

Randomized, (placebo-)controlled, phase III trials, if well-designed, are recognized as the standard approach in confirming the efficacy of a new therapy [1]. They are often required by leading drug registration agencies, such as the US FDA and EMA, as evidence that a treatment is effective. However, the placebo-control design raises ethical concerns when the scientific interest to generate new knowledge is weighed against trial participants' need for medical care. This balance is often problematic. On the one hand, placebo-controlled trials are needed to confirm or deny the efficacy of the therapy under investigation. On the other hand, an important concern is that patients randomized into a placebo group might miss a therapeutic benefit.

According to ethical guidelines, an experimental therapy has to be compared with a best-known proven therapy, where one exists. Application of this guideline entails a number of challenges, complicating the evaluation of the ethical justifiability of placebo use. A number of factors can cloud the judgment about the efficacy of a therapeutic intervention and thus the judgment regarding whether a best-known proven therapy exists. Without claiming completeness, this article discusses a number of such factors with a focus on regenerative medicine. Although estimation of treatment efficacy can be difficult for many kinds of therapeutic interventions, it often is especially challenging in regenerative medicine for a number of reasons. As discussed further in this article, full information about the efficacy of therapeutic approaches in regenerative medicine (e.g., gene therapy, cell therapy) is unavailable until a long-term follow-up (often outside the scope of a standard clinical trial) has been performed. Results from such follow-ups are often unavailable at the time of decision-making as to the ethical justification of placebo controls. Moreover, therapeutic approaches in regenerative medicine can often have different pathways toward the same goal: for example, a cell therapy using the same type of cells and aiming to treat the same condition can be administered a varying number of times, with varying concentrations of cells and at different intervals. It can be difficult to judge which of these therapeutic approaches should be considered the best-proven therapy. Moreover, patients considering regenerative therapeutic approaches often tend to be suffering from dangerous conditions for which other therapeutic alternatives have already been exhausted (e.g., persons suffering from a version of graft-versus-host disease that cannot be treated with steroids). The use of a placebo-control design can become especially controversial in cases where such patients participate in randomized controlled trials (RCTs).

‘Proven’: varying degree & quantity of evidence

For placebo-controlled trials to be ethically justifiable, they have to meet certain criteria outlined in ethical guidelines such as the International Ethical Guidelines for Health-related Research Involving Humans prepared by the Council for International Organizations of Medical Sciences (CIOMS) [2] or the World Medical Association Declaration of Helsinki [3], an important ethical code for medical research and a key document in the ethics of international research involving human participants [4]. Provided that the need for a placebo-control design is scientifically justified, the main requirement in these criteria (Table 1) is the lack of an existing proven intervention. When this requirement is met, the use of such a design can be ethically justified and there is no need to consider other requirements, which become important when a proven intervention does exist. These other requirements fall outside the scope of this article since the focus here is on the criterion of ‘proven intervention’.

There are different degrees and quantities of evidence: the safety and efficacy of a therapy can be proven in different phases and different numbers of trials. The quality of these trials can also vary. Internationally known frameworks, such as Grading Quality of Evidence and Strength of Recommendations (GRADE), suggest that evidence obtained in randomized trials should be considered 'high’ quality and evidence in observational studies as ‘low’ quality [8]. Meanwhile, any other evidence should be considered as ‘very low’ quality, with the grade of evidence decreasing based on, for example, study quality, inconsistency, imprecise or sparse data or high probability of reporting bias [8]. Similar recommendations can be found in different national guidelines. For instance, the Swedish Agency for Health Technology Assessment and Assessment of Social Services grades the evidence for interventional studies as strong, moderately strong, limited and insufficient, and suggests lowering the grade if there are flaws in, for example, study quality, conformity among studies, transferability/relevance and precision or if there is a high likelihood of publication bias [9]. Intending to assist in the decision-making process of those involved in cancer care [10], the National Comprehensive Cancer Network guidelines [11] take into consideration even levels of consensus and disagreement [12]. Regulatory agencies such as the US FDA or EMA may also differ concerning the level of evidence required to consider an intervention ‘proven’ [10].

Other factors complicating the estimation of treatment efficacy are the multitude of effective interventions and positive publication bias. It can be hard to choose which intervention to use as a comparator when there are a few that are effective. When it comes to bias, it is known that the failure to publish research findings can make it difficult to estimate what knowledge about a given therapy exists. A large proportion of unpublished results has the potential to distort evidence for treatment choices [13], leading to either overestimation of treatment effects or neglecting findings with potentially negative consequences for patients [14]. Investigators are responsible for the dissemination of trial results regardless of findings [3], but nonpublishing of results is a common problem in biomedical research in some areas (e.g., pediatric RCTs) more than others. One should thus consider publication bias when assessing whether a particular treatment is favored [9,15–17].

Some ethical guidelines (e.g., CIOMS guidelines) specify the scope of ‘proven intervention’ or ‘established effective intervention’, whereas others do not. According to CIOMS, an “established effective intervention for the condition under study exists when it is part of the medical professional standard. The professional standard includes but is not limited to, the best proven intervention for treating […] the given condition. In addition, the professional standard includes interventions that may not be the very best when compared with available alternatives, but are nonetheless professionally recognized as a reasonable option (for example, as evidenced in treatment guidelines)” [2]. In some cases it can be difficult to determine the level of evidence sufficient for an intervention to be classified as ‘best proven’ or ‘professionally recognized as a reasonable option’. For instance, when an experimental therapy is accessible outside the framework of clinical trials under special approval programs, it should not be taken for granted that the previously noted definitions encompass such therapies. CIOMS guidelines are clear that even “established effective interventions may need further testing, especially when their merits are subject to reasonable disagreement among medical professionals and other knowledgeable persons” or if a determination of efficacy in ‘rigorous clinical trials’ is lacking [2]. Conducting rigorous trials when a therapy is already made available outside the framework of clinical research can face a number of challenges. They are reviewed next.

‘Conditional’: can conditions be met?

Recently, programs speeding up access to medicines at the premarketing or postmarketing authorization stages [18] have been introduced in different countries to enable patients to access medicines outside traditional regulatory frameworks [19]. They have been initiated as a response to demands (e.g., from patients with life-threatening diseases or rare diseases) on regulators to provide access to treatments sooner, often on the basis of less robust evidence of safety, efficacy or cost–effectiveness than have been required before the introduction of such programs [20]. There are different kinds of programs, with some simply speeding up the registration or funding process (prioritization initiatives), and others changing the requirements for evidence and shifting thresholds for regulatory approval [20]. These initiatives have been implemented by major regulatory bodies such as the US FDA, EMA and Japan's Pharmaceuticals and Medical Devices Agency [21,22]. The initiatives have different names and imply different pathways. They usually take one of three forms. Pathways providing faster access to therapies may be achieved by making existing regulatory and funding processes more efficient, despite uncertainties surrounding safety or efficacy and despite uncertainties about cost–effectiveness [19]. The latter two types require suspending or overriding accepted standards of evidence of safety, effectiveness and/or cost–effectiveness [19].

While each type raises different issues, including sociopolitical and moral aspects [20], this paper focuses on the second category, allowing access to treatments despite uncertainties about their safety or efficacy. An example of this category would be the EMA conditional marketing authorization (CMA), which is based on incomplete benefit–risk assessment data. It remains conditional until confirmatory postmarketing measures are completed [23,24]. Non-European examples of this category are the United States accelerated approval (AA) program (not to be confused with the EMA's accelerated assessment [AA] program, which is an example of the first category [25]) and the Japanese conditional early approval (CEA) program [23]. These three systems exhibit similarities and differences. Pathways in the USA and the EU have similar conditions when it comes to which kind of therapies qualify for such pathways. They are typically reserved for treatments for severe, life-threatening or seriously debilitating diseases or for considerable unmet clinical needs [22]. Among the reasons for such reservation is earlier access to innovative therapies for critically ill patients, especially when they have no other effective treatment alternatives. Simply put, not having to wait for full approval of potentially efficacious therapies has a chance to improve the longevity and/or quality of life of desperately ill patients, an argument often heard in patient activism.

Conditions concerning the seriousness of disease are not specified in the Japanese pathway [26]. The latter, unlike the American and European systems, contains a time-limited approval [26] and provides financial aid to people who have experienced an adverse reaction and developed serious side effects as a result of conditionally approved therapy [24,27].

For the sake of simplicity, in this article, ‘conditional approval' is used to refer to all pathways in the second category. Under such pathways, therapies with promising but incomplete efficacy data get market authorization on the condition that they will continue to be evaluated while on the market [22]. In other words, these pathways reduce the requirement for evidence of safety and efficacy, typically required in the premarketing stage of the development of therapies. Instead, these requirements become postmarketing obligations [26]. Depending on the data gathered during the postmarketing period, therapies may be withdrawn from the market, granted ‘standard’ approval or continue to be marketed conditionally [22].

The rationale of conditional approval programs is to make it possible for patients with unmet medical needs to benefit from rapid access to novel therapies, whereas uncertainty around their efficacy is supposed to be reduced by conducting further clinical research [28]. The moral concerns raised by this approach are not so much about its rationale, but rather the consequences it has on affected patients and on the possibility of conducting rigorous research in the future.

When it comes to patients, the concern is primarily about the ethical justifiability of burdening those lacking therapeutic options when clear evidence of clinical benefits is absent [28,29]. Vulnerable stakeholder groups with unmet medical needs would be the ones most burdened, especially if conditionally approved therapies come at a high cost. Vulnerable stakeholders are either large cohorts of patients suffering from serious or life-threatening conditions or small groups suffering from rare or orphan diseases affecting so few people that developers are not interested in investing in long-term clinical studies needed to gain market authorization through the conventional approval pathway [28]. Concerns have been raised that institutionalizing regulatory schemes that lower the evidentiary standards for clinical benefit does not align with the commitments of public health to protect and promote good health [28]. Likewise, the expected public health advantage of conditionally approved therapies is rarely strong enough to justify the risks of limited clinical evidence [30].

When it comes to the possibility of conducting rigorous research in the future, the collection of further evidence as to the efficacy of the conditionally approved therapy can, in reality, become limited by a number of factors, presented next.

‘Approval’: implication for efficacy?

That an ‘approved’ therapy can be seen by patients and medical personnel as efficacious may seem to be a purely semantic concern, but it can have a bearing on ethical decision-making. To those stakeholders who are likely unfamiliar with how conditional approvals are granted (e.g., patients or their families), the word ‘approval’ may be an implication of efficacy, thus, potentially leading to a therapeutic misconception. When a particular therapy is at least formally in clinical use in one or several countries, even clinicians, eager to help their patients, may be under the impression that this therapy should be interpreted as a best-proven one. In such a case, clinicians could regard placebo-controlled trials as inappropriate, and patients would not want to take part in them. With the focus on ‘approval’, it can be easy to miss that such products may end up not being used in practice. An example of this is the conditional approval of Prochymal, a mesenchymal stem cell (MSC) therapy product for the treatment of acute graft-versus-host disease (GvHD) in steroid-refractory patients under 18 years of age, which achieved Canadian regulatory approval using the conditional marketing approval system [31]. In light of the devastating outcomes of this type of acute GvHD, Health Canada deemed that the risk–benefit ratio of this cell therapy was acceptable and allowed its conditional approval [31]. A few weeks later, the Canadian approval was followed by Medsafe (medical regulatory agency) approval of the same product in New Zealand [32]. However, Prochymal was not reimbursed by insurance companies and ended up not being used [33]. Accessibility in practice, affected by aspects such as insurance coverage could, therefore, be seen as an additional condition for conditionally approved therapy to be successful. The pricing of such therapies merits special attention but falls outside the scope of this article.

Enforcement difficulties

Ensuring that postapproval studies are actually carried out, and in a timely manner, is another factor. Even though conducting postmarketing studies is an important condition of ‘conditional’ approval, such studies are often not enforced in practice. According to a study comparing the US, EU and Japanese systems of expedited approval of regenerative medicine products, obligations to collect and report postmarketing data on approved products are frequently bypassed [26]. An analysis of conditional approvals granted by the EMA also highlighted inconsistencies and delays with regard to the fulfilment of the criteria for authorization and the specific obligations imposed at the time of approval [30]. A notice of compliance with conditions, issued by Health Canada in the process of drug approval, and stipulating that the developer will undertake further studies to confirm benefit, is not legally binding [34].

Questionable quality of postmarketing studies

The questionable quality of postmarketing studies also contributes to insufficient evidence on the therapeutic efficacy of conditionally approved treatments. One recent British study comparing marketing applications for cancer drugs of uncertain therapeutic value reviewed by the FDA and EMA raised a concern that confirmatory studies imposed in the postmarketing period through special approval pathways (FDA's AA and EMA's CMA) may not be suited to deliver timely evidence to confirm remaining questions of efficacy and safety [35]. This is due to shortcomings in study design and delays in conducting required studies with due diligence [35]. For patient safety, it is important that postmarketing commitments are performed in a timely manner [36]. Poor quality of postmarketing studies is not unique to conditionally approved therapeutic products. In other product categories, compliance with phase IV study requirements has also been reported as poor, suffering from small sample sizes and delays in reaching protocol agreements among other issues, which together make the clinical usefulness of such studies questionable [37].

A particular type of postapproval study has especially raised doubts about their ability to generate efficacy evidence of high quality. It is observational studies using ‘real-world data’, required by Japan's CEA (unlike confirmatory trials with robust end points, demanded by USA's AA and EU's CMA) [23,38,39]. The Japanese system for conditional approval of regenerative medicine products has been criticized as allowing “products of undetermined efficacy to enter the Japanese market” [40]. With the clinical benefit of those products “underprioritized with respect to their safety”, patients and healthcare systems may end up burdened “without any clear benefits” [41]. This, in turn, may “undermine the credibility of the regenerative medicine field” or affect public trust in science [41]. Observational research methods carry a risk of confounding effects and introducing bias and must be carefully designed to reduce these effects [19]. Another important concern is that ‘real-world data’, if gathered instead of robust late-stage clinical trials, would affect incentives for companies to spend money conducting such trials [19], thus reducing the likelihood of generating robust efficacy evidence.

Long-term effects: what we do not know

Some treatments, where the introduced therapeutic agent can be active over a longer period, necessitate a long-term follow-up to fully estimate their efficacy. This is the case in, for example, gene or cell therapies. In the latter, the duration of follow-up would usually depend on the type of cells used, the disease treated and other factors. For various reasons, including funding issues, clinical trials often last up to 5 years, which may be too short a time to establish efficacy with certainty in cell or gene therapies. Thus, efficacy reports published after the completion of cell or gene therapy trials often cannot reflect long-term efficacy/safety information. Such reports should be read with attentiveness to the context in which this therapy is used. One should then evaluate whether there is a risk of important knowledge gaps concerning efficacy if long-term follow-up is missing. This does not imply that efficacy in all published cell or gene therapy trials is questionable, but that in some cases knowledge gaps may not be negligible.

Future RCTs: to be or not to be?

Weakened evidence standards in the approval of regenerative medicine products is a general problem, to which conditional approvals can contribute. Weakened evidence standards can negatively impact future research. For instance, access to therapy prior to registration has the potential to limit the number of patients available for trials and may impede further data collection [19]. In the country granting conditional approval, it would hardly be possible to recruit participants to placebo-controlled trials, since patients can access the same therapy outside a clinical trial (albeit not free of charge) and do not have to risk being allotted into a placebo arm. In the Japanese context, concerns have been raised that their system of conditional approval could encourage developers to design cheaper and shorter research and to prefer development processes that provide less convincing evidence of the therapeutic value of a new treatment [42]. This would signal that such evidence is sufficient to achieve market access [42], thus implying that no further research is needed.

Another example of accelerated access programs contributing to weakened evidence standards and making research in the form of RCTs difficult is the increasing acceptance of single-arm studies as proof of efficacy, as was shown in the context of cancer treatment studies [43]. With RCTs sometimes considered unfeasible due to, for example, low recruitment rates of trial participants [44], or where the use of placebo control would not be ethical [45], alternative methods such as single-arm studies can be considered [46], although they “cannot directly measure if a given treatment improves outcomes compared with existing standards of care” [43]. In recent years, “a large increase in submissions to the EMA and FDA using nonrandomized study designs involving comparisons with external controls” has been reported [45]. The use of single-arm studies has been criticized because it is not always “based on empirical evidence of infeasibility of RCTs”, which would imply questionable application of this method, possibly driven by the economic incentives of a manufacturer [43].

Although conditional approvals require further confirmatory evidence from postapproval studies [45], one could argue that patients are denied access to an approved treatment if a placebo-controlled trial is launched to further test a therapy already under such approval. Opposition to placebo controls may arise due to confusion about whether a ‘best-proven therapy’ exists when similar therapeutic approaches are granted conditional approvals, be it in a home country or abroad. For instance, the use of placebos in the RETHRIM consortium, conducting the first Europe-wide placebo-controlled, randomized, phase III trials using MSC regenerative therapy for the treatment of steroid-resistant visceral GvHD [47] was questioned based on the conditional approvals of MSC-based therapies for GvHD in Japan, the concern being whether these therapies should be considered as a “best-proven” therapy.

Weakened requirements of evidence in the approval of regenerative medicine products can lead to situations where an approved product may become a standard of care, despite limited data that lead to this approval. An MSC product developed by JCR (partner of Mesoblast) in Japan, for instance, received a full (not even conditional) approval [33] as soon as about 1 year after filing for marketing approval. One of the consequences was that the development of similar MSC products was severely hampered, since the JCR product, having full approval, became a standard of care, in spite of the limited data that led to this approval. As a consequence, no placebo-controlled trial could be done.

Such an effect on potential placebo-controlled RCTs may be long-lasting. If, for instance, a conditional approval were lifted for one reason or another, it is not certain that a therapy previously under conditional or full approval would not continue to be referred to as a standard of care, at least for some time. Once a therapy is on the market and patients and clinicians become familiar with it, regulatory withdrawal and/or disinvestment are unlikely to occur [19]. If it does, it is likely to take time and, by the time products are removed from the market, large numbers of patients may have been exposed to an unsafe or ineffective therapy [19]. A possible explanation is that once used to provide a particular therapy, physicians continue to use it [48]. They may form a belief that the administered therapy plays a role in a patient's disease progression, but it is important that it is science, and not eagerness to help one's patient, that guides medical practice [49]. It is crucial that physicians are well informed about the scientific evidence upon which conditional approvals are granted and the importance of further generation of efficacy evidence in rigorous clinical trials. Clear information from, for example, EMA or FDA on whether the product in question is under conditional or full approval should be made available to medical professionals. For instance, the conditional approval status of a therapeutic product could be made more visible on the therapeutic product itself.

Approved de jure, unused de facto?

It must be borne in mind that conditional approval of a therapy does not mean that this therapy would necessarily be trusted and widely used in practice by physicians in the position to provide it to their patients. Empirical evidence illustrates the lack of trust among physicians considering the prospects of the practical application of therapies under conditional approval. An Australian study investigating attitudes of physicians toward accelerated access to medicines has found a diversity of clinical opinions in this area, including those reflecting a cautious approach among the interviewed physicians, advocating for the need “to know enough to feel confident that this patient in this circumstance, in this context, actually will benefit from the use of this drug” [18]. Another example relates to physicians' mistrust of the Alzheimer's treatment, aducanumab [50,51], which was conditionally approved by the FDA in 2021 on the basis that it reduced amyloid plaques in studied patients' brains, although it “did not meaningfully improve patients' ratings on a clinical dementia scale” [52].

There are some efforts, albeit lacking harmonization among countries, to address the issues just mentioned. One recent study found that available results from postmarketing studies were almost always used within health technology assessment, and had led to changes in conclusions about the relative effectiveness of treatments [53]. Also, some regulatory pathways (e.g., marketing approvals for regenerative medicine products in Japan) use time limitations and enable conditional approval to be automatically withdrawn if the applicant or sponsor cannot obtain standard approval during the allocated period [26]. But lack of harmonization among different regulatory agencies as to how much evidence is needed to allow the marketing of therapies contributes to complicating the estimation of the presence or absence of proven therapy. This problem is not unique to regenerative therapies. In the case of cancer treatment, for instance, a matched comparison of marketing applications for cancer drugs of uncertain therapeutic value reviewed by both FDA and EMA found frequent discordance between the two agencies on regulatory outcomes and the use of special regulatory pathways [35].

More consideration is needed as to how further evidence should be collected in transforming unproven stem cell-based interventions into more evidence-based ones [29]. In cases where the nature of experimental regenerative products would make the conduct of controlled studies unethical (e.g., due to the invasive nature of the delivery procedure), clear pathways to making such transformations are especially important. There can be cases where RCTs are not possible due to the limited number of patients or other reasons, and established alternatives to RCTs are lacking. Although submissions to the EMA and FDA using nonrandomized study designs involving comparisons with external controls have increased [54] and new products have been authorized based on nonrandomized efficacy and safety data [55], such applications cannot be called an established alternative to RCTs.

Finally, it should be noted that previously mentioned problems, if not dealt with effectively, may affect increasing numbers of therapeutic products and approaches. Although enacted for the rapid commercialization of innovative products (as it is in the Unites States and EU), legislation for expedited-approval pathways has been found to benefit less innovative products [26]. For instance, FDA review programs were expanded to include a larger proportion of products approved overall [36]. Such a trend would have an important implication for patient care if treatments begin to be approved at earlier stages based on less rigorous clinical testing [36].

Conclusion

The ethics of using placebo controls have been discussed for decades. But in recent years, the shift toward programs allowing accelerated patient access to therapies on the basis of less robust evidence regarding their safety and efficacy necessitates a new ethical debate that would take this context into consideration. I am not suggesting reconsidering the existing ethical guidelines as to the ethics of using placebo controls. What I propose is, rather, to debate how they should be interpreted and applied in the context of changing requirements for evidence. This change is supported by programs allowing patients to access therapies on the basis of reduced safety and efficacy evidence than required before introducing these programs.

On the one hand, the demands we have for the quantity and quality of evidence needed for a given therapy to be considered a proven intervention seem to become less strict with the introduction of conditional approval programs. The quality of efficacy evidence necessary to allow patients to access a particular therapy has a bearing on the ethical justification of using placebo controls. This is particularly problematic when rigorous clinical trials are put in place to study the efficacy of therapeutic approaches similar to those already under conditional approval. When conditionally approved therapies are made available to patients in approving countries, this makes it difficult to conduct placebo-controlled RCTs aiming to study the efficacy of similar therapeutic approaches in other jurisdictions. If conditionally approved therapies are interpreted as ‘proven intervention’ or ‘established effective intervention’, the use of placebo controls (provided the absence of compelling methodological reasons that would justify disregarding this criterion) becomes ethically problematic.

On the other hand, the generation of further efficacy evidence after conditional approval is notoriously difficult. There are many factors that contribute to that, including the lack of enforcement of postapproval studies, their low quality and the interpretation that insufficient efficacy evidence is enough to achieve market access. The latter implies that no further research is needed. The difficulty of generating evidence about the efficacy and clinical utility of a given therapy in rigorous clinical trials leaves us in a vicious circle. Unable to perform rigorous clinical trials, we continue to assume that a therapy under conditional approval is efficacious, or maybe even consider it as a proven intervention. This, in turn, further precludes placebo-controlled RCTs to establish the efficacy of the treatment in question. It can even become more difficult to study the efficacy of therapeutic approaches similar in kind to those accessible under conditional approvals in the same or other jurisdictions. An example of such a therapeutic approach could be cell therapy involving the same type of cells and aiming to treat the same condition, whereas dosing, administration intervals and timing in the course of the disease may differ. The way to get out of this vicious circle is to create favorable conditions for the generation of evidence about the therapeutic efficacy of a given treatment even after its conditional approval and marketing.

The first step would be to acknowledge at the international level that efficacy evidence on which conditional approvals are granted, if not obtained through rigorously conducted late-stage RCTs, cannot be considered sufficient to provide continuous or long-term marketing of approved therapies, thus allowing ignorant patients to be treated with probably safe, but hardly effective products. Provided there is a scientific need for a placebo-controlled RCT to generate efficacy evidence and provided such evidence about this therapy is lacking, such RCTs should, in principle, be conducted even if a tested therapy is already conditionally marketed. Practical difficulties in recruiting participants into such trials would, however, remain, especially if the therapy tested in an RCT is the same as the one approved conditionally and available on the market. One could therefore consider testing the efficacy of therapeutic approaches similar in kind to the approved one, if relevant. Alternatively, other methodological approaches than placebo-controlled RCTs could be considered.

The second step would be to encourage public debate and public outreach about the rationale behind conditional approval pathways and the dangers of abusing them. Both patients, especially those desperate to try something rather than nothing, and medical professionals, eager to find treatment solutions for their patients with unmet medical needs, should be clearly informed about the efficacy status of therapies they are considering. This is important to enable informed decision-making, both among patients and medical professionals. I am not making a case against conditional approval pathways, but against them becoming a hindrance to the further generation of robust and trustworthy efficacy evidence. Without it, informed consent will reduce to a formality and treatment to guessing rather than evidence-based medicine.

Future perspective

Awareness is needed that the scope of concepts such as ‘proven intervention’ or ‘established effective intervention’ present in main ethical guidelines' requirements as to the use of placebo controls can become a matter of interpretation and in this way challenge the further generation of evidence about treatment efficacy, including using placebo-controlled RCTs. There is a need for an international debate and harmonization as to how these requirements should be interpreted and applied in light of programs that allow patients accelerated access to therapies approved only conditionally and on the basis of less robust evidence regarding their safety and efficacy than has been required before these programs emerged. Public outreach about these issues would help keep both patients and medical professionals informed about the efficacy status of therapeutic products they are using or prescribing. This, in turn, would enable their informed decision-making. There is likewise a need to reflect on methodological aspects and end points in other methodological approaches where RCTs cannot be performed due to, for example, a limited number of patients. In case other approaches are used, such as single-arm studies, guidance on how to understand the role of nonrandomized evidence is needed. Moving toward personalized medicine and precision medicine, where treatment may become available for individual patients, actualizes a need to consider novel methodologies.