Introduction
Cell and gene therapies comprise the majority of the products that the European Medicines Agency designates Advanced Therapy Medicinal Products (ATMPs) [
1]. While they offer patients potentially transformational gains in health, these therapies also pose issues for healthcare payers in all jurisdictions. Much has already been written about the distinctive characteristics of these new therapies, and the need (or otherwise) for new health technology assessment (HTA) methods, [
2‐
9]. There has also been discussion of how these characteristics might shape price and reimbursement negotiations [
10,
11] and managed entry agreements [
12,
13]. Finally, the literature has also explored the cost-effectiveness profile of cell and gene therapies [
14‐
16] and the variation in the reimbursement or coverage decisions made in different jurisdictions [
17‐
19].
Many of the assessment challenges relating to cell and gene therapies apply to rare disease treatments more generally, and most of the current cell and gene therapies are for rare conditions. However, in the case of cell and gene therapies the challenges are compounded by the potentially transformational nature of the health gain and the potential long-term nature of the ‘cure’, which is subject to considerable uncertainty.
The issue of how HTA bodies are responding to the assessment challenges posed by cell and gene therapies has received relatively lower attention. HTA methods are partly conditioned by the evidence currently available and new methods often imply the need for new types of evidence. For example, the clinical evidence for cell and gene therapies may be sparse in relation to that available for most pharmaceuticals, the economic benefits produced by cell and gene therapies may require additional evidence to demonstrate them and many of the innovative payment models being proposed may require further evidence generation post product launch.
In addition, although some of the papers in the current literature argue for a different approach to the HTA of cell and gene therapies, most HTA bodies prefer to use a standardized approach for all health technologies. In this context, the paper by Drummond et al. [
4] argues that HTA bodies could apply their standardized approach, but should give specific attention to some of the particular challenges posed by cell and gene therapies.
Therefore, the aims of this research were: i). to provide a better understanding of the specific evidence needs for assessment of the clinical and cost-effectiveness of cell and gene therapies, ii) to explore the extent that the relevant categories of clinical evidence, economic evidence and evidence post-launch are considered in the HTA processes in several jurisdictions, iii) to identify the issues in the generation and use of evidence that require further discussion and consideration.
Methods
To achieve the aims above, we undertook i) a targeted review of the literature on the clinical and economic evidence needs for these therapies, and ii) an in-depth analysis of HTA reports from 8 major jurisdictions for 9 cell and gene therapies in 10 indications, together with any associated publicly available documents on managed entry agreements and post-launch evidence requirements.
Targeted literature review
Our starting point for the targeted literature review was the paper by Drummond et al. [
4]. We selected this paper, which discussed the arguments for a new reference case for the HTA of gene therapies, because it contained the most extensive discussion to date of the implied clinical and economic evidence needs. However, the paper is essentially expert opinion based on a limited review of the literature available at the time. Therefore, rather than accept the suggestions made in that paper without question, we felt a further literature review was necessary to determine whether any new issues had arisen, or whether any relevant issues had been overlooked.
Given that there is already a substantial literature on the assessment challenges posed by cell and gene therapies, including several targeted or systematic reviews [
15,
20‐
22], we first reviewed the references in the most recently published systematic review [
22] to identify any papers focussing on evidence requirements. We selected 18 papers for detailed review from these sources. Then we followed a single-round ‘pearl growing’, or ‘snowballing’ approach, via PubMed, starting from the paper by Drummond et al. 2019 until the end of February 2022 [
4]. We excluded abstracts and conference proceedings and included peer-reviewed papers if they discussed existing or new challenges related to assessment, together with recommendations about the evidence needed to perform this assessment. In total there were 36 citations of the Drummond et al. paper during this period. Of these, 25 were retrieved as peer-reviewed papers, and 20 were used in the review. Five were excluded as they did not deal with the challenges of assessing cell and gene therapies. Additional references to post-launch requirements were also retrieved from the grey literature. The papers identified were critically appraised by two co-authors independently, and referenced and summarised narratively according to the framework proposed in the Drummond et al. checklist.
Analysis of HTA reports
For the selection and inclusion of HTA reports in our analysis, we identified HTA evaluation reports for cell and gene therapies from 7 HTA bodies, and one regulatory agency that also performs HTAs, published prior to end of July 2021. The agencies were based in the 5 largest European countries, plus the US and Canada: Agencia Española de Medicamentos y Productos Sanitarios (AEMPS) in Spain, Agenzia Italiana del Farmaco (AIFA) in Italy, Canadian Agency for Drugs and Technologies in Health (CADTH) in Canada, Gemeinsamer Bundesausschuss (G-BA) in Germany, Haute Autorité de Santé (HAS) in France, Institute for Clinical and Economic Review (ICER) in USA, National Institute for Health and Care Excellence (NICE) in England, and Scottish Medicines Consortium (SMC) in Scotland. The agencies selected were also representative of different approaches to conducting HTAs, some using the cost per QALY approach and others relying on an interpretation of the clinical data alone (see Table
1).
Table 1
HTA reports analyzed
Kymriah DLBCL | X | | | X | X | X | X | X |
Kymriah ALL | X | X | | | X | X | X | X |
Yescarta | X | X | | X | X | X | X | X |
Luxturna | X | X | X | X | | X | X | X |
Strimvelis | X | | | | | | | |
Imlygic | X | | | | | | X | X |
Alofisel | X | | | | | X | X | X |
Provenge | X | | | | | | X | |
Glybera | | | | | | X | X | |
Zolgensma | X | X | X | X | X | X | X | |
A data extraction form was developed based on the template first presented by Drummond et al. [
4] following verification of updates needed through the targeted review (see Table
2). The data extraction form included three sections: the first relates to clinical effectiveness assessments, including items such as sample size and duration of pivotal clinical trials used to approve the therapies, and whether trials were single-arm or uncontrolled; the second section relates to the valuation of benefits for gene therapies, including the treatment of severe disease, value to caregivers, scientific spillovers, and substantial improvements in life expectancy; the third section relates to additional considerations, such as discounting or consideration of alternative payment models.
Table 2
Checklist for assessing gene therapies
Clinical Effectiveness |
Surrogate endpoint used | □ | □ | Validation given? |
Rare disease | □ | □ | Prevalence _____ |
Serious condition | □ | □ | |
Single-armed trial | □ | □ | Matched historical cohort used? |
Pediatric population | □ | □ | Age range _____ |
Reporting of adverse consequences and risks | □ | □ | |
Size of clinical trial | _____ number of patients | |
Length of clinical trial | _____ duration in months | |
Extrapolation to long-term outcomes | _____ duration in months | |
Elements of Value | | | Quantification |
Severe disease | □ | □ | |
Value to caregivers | □ | □ | |
Insurance value | □ | □ | |
Scientific spillovers | □ | □ | |
Lack of alternatives | □ | □ | |
Substantial improvement in life expectancy | □ | □ | |
Other Considerations | | | Notes |
Discounting | | | |
Different discount rates explored | □ | □ | |
Uncertainty | | | |
Alternative payment models explored | □ | □ | |
The checklist has been previously used by Huygens et al. [
16] in a smaller study of a single gene therapy in 3 jurisdictions. We applied it by recording whether each item was considered to be relevant in the jurisdictions concerned (Yes/No/Not available), together with illustrative quotes from the HTA reports analysed. Data extraction was performed by one researcher and a random sample of the reports (
n = 10) was checked for accuracy of data extraction by another member of the research team.
We also conducted a more detailed analysis of the comments made in cases where a particular item was considered, by classifying the comments as ‘positive’, ‘neutral’ or ‘negative’. Positive assessments were those where the HTA body considered that the manufacturer had made a reasonable attempt at dealing with the issue, or where the HTA body recognised the importance of the item in considering the technology. Negative assessments were those where the HTA body felt that the issue had not been adequately addressed, or where it did not recognise the importance of the item in considering the technology. Neutral assessments were those where the committee made no comments either way. These assessments were made by two members of the research team, with a consensus being reached in cases where there were any differences of opinion.
The findings were discussed by a small group consisting of academic researchers, payers, patient representatives and pharmaceutical industry personnel, prior to drafting a paper for submission to a peer-reviewed journal.
Discussion
Based on the findings of the literature review and analysis of HTA, it is possible to identify several issues in the generation and use of evidence for cell and gene therapies that require further discussion and consideration However, in discussing these issues it is important to recognise that there are differences between HTA bodies in different jurisdictions in how the clinical and economic evidence should be used. Therefore, any suggestions for improvements need to take account of these country-level differences.
To date, most HTA bodies, including those studied here, have not instituted a separate assessment programme or approach for cell and gene therapies, or advanced therapy medicinal products (ATMPs) more generally, although some have supplemental processes for rare disease treatments that are more suited to their assessment [
61,
62]. Given the preference of most HTA bodies to use the same standardized approach for all health technologies, it may be more realistic to consider ways in which bodies can tackle the various challenges within their existing approach to HTA. Ultimately, any changes in how cell and gene therapies are assessed will reflect the views of the various HTA bodies and the constituencies they serve. However, a few suggestions are given below.
Consideration of clinical evidence
In terms of establishing the validity of the surrogate endpoints, current appraisals of cell and gene therapies are unlikely to meet the highest standards (e.g., link between treatment effects on surrogate and final outcomes established based on meta-analyses of RCTs in the same indication and treatment class). However, it is still recommended that appraisals of these therapies follow a structured approach to the validation of surrogate endpoints, and justify the adoption based on the prognostic value of the biomarker rather than on the surrogacy value, on relaxed requirements of indication- and treatment- specific evidence of validation, on full or partial exchangeability of evidence across treatment classes and populations. In addition, carefully designed post-launch evidence generation programmes may answer the question of surrogate validity and effectiveness on patient-relevant endpoints, which may be the preferred approach in some jurisdictions. Therefore, it would be useful for HTA bodies and payers to agree standards for long-term patient relevant outcome collection and for validating surrogate endpoints, while acknowledging the challenges posed by small patient populations. Ideally, post-launch evidence generation should be agreed at the time of the HTA, be based on rigorous protocols, and, if possible, enforced by management agreements that make the reimbursement and price status contingent on data collection and the relevant findings.
Recruitment to conventional randomized controlled trials may be difficult due to the small target population, and may be influenced by the expectation of sizeable benefits in high unmet-need areas, that strongly enhances the attractiveness of the experimental arm over the standard of care. In addition, there may be operational hurdles, in enrolling patients outside qualified centres, and in the choice of the comparator arm. Although methods for the design and analysis of clinical trials in small populations have been proposed [
63‐
65],they have not yet been widely accepted in the HTA community.
A high preponderance of pivotal single-arm uncontrolled trials means payers need to find, or assemble, a suitable contemporary historical cohort of patients treated according to standard practice, which can be used to determine the comparative effectiveness of the therapies. A common concern observed in the HTA reports was that ‘standard practice’ could vary across jurisdictions or change over time. Guiding principles for performing such treatment comparisons could be agreed in advance between manufacturers and decision-makers, starting from high-quality patient-level data from reliable and traceable sources; appropriate cohort selection based on matching inclusion/exclusion; suitability of real-world endpoints; fit-for-purpose analytical methodologies. The acceptability of synthetic cohorts or in silico modelling for this purpose should be established. Therefore, it would be useful to agree criteria for when a single-arm clinical study can be deemed acceptable, and guidelines for assembling a matching comparator cohort.
Due to the short timeframe of the pivotal clinical studies (typically one or two years), long-term benefit and adverse events of cell and gene therapies are not substantiated by trial data and remain largely uncertain at the time of the assessment. Therefore, there is a need to support the assumption of maintained improvement in health or treatment effect in the long-term. It is recommended that current biological knowledge on the pathophysiology of the disease complemented by additional external sources of evidence inform this step. The extrapolation of survival curves is more and more often based on mixture cure models in appraisals of cell and gene therapies, although the plausibility of presence and magnitude of a cure fraction must be discussed at length. Extensive scenario and sensitivity analyses are warranted, given that this parameter may have the largest impact on the cost-effectiveness analyses results.
Finally, quality of life measures, although important to patients, are not necessarily considered by those HTA bodies assessing added clinical value, as they are perceived as being subjective. Some countries may need to consider changing their assessment framework to encompass quality of life, although efforts to collect these data in clinical studies also need to improve, as quality of life evidence for cell and gene therapies, and rare diseases more generally, is scarce and often of poor quality [
66]. It would also help if there was a requirement from regulatory or HTA bodies to undertake quality of life data collection as part of clinical trials.
Consideration of economic evidence
In addressing the consideration of economic evidence, it is again important to recognise that HTA processes differ quite widely across countries. In those countries that assess the ‘added clinical value’ provided by therapies, economic evidence is not explicitly considered in the assessment process, although the acquisition cost of new therapies is considered as part of the price negotiation. In those countries assessing the ‘cost-effectiveness’ of new therapies, it is possible to discuss how economic value is characterised and whether the definition of ‘value’ should be broadened. It is very rare for countries to switch from one approach to HTA to the other, although France did add the consideration of cost-effectiveness as a supporting approach to its ‘clinical added value’ decision-making procedure. Therefore, in making recommendations for change, it is important to consider how these can be implemented within each of the contrasting approaches to HTA.
First, there may be arguments for expanding the HTA process to include the impacts on caregivers and the family. These are important for all therapies, but arguably particularly important in the case of therapies for very severe debilitating illnesses and those affecting children. Within the cost-effectiveness approach this suggests that impacts on caregivers’ quality of life could be assessed and possibly the additional costs imposed on families. If a broader societal perspective were adopted, costs falling on patients and families would be included [
67]. In those jurisdictions adopting a narrower health care perspective, it would still be possible to present these costs separately from the costs falling on the health care system, so that decision makers can consider them.
Within the clinical added value approach, consideration of impacts on caregivers and the family could be made an explicit item of value within the ‘clinical’ value, or the composition of assessment committees could be expanded in include patient representatives, who are likely to stress these issues.
Secondly, there is broad agreement that cell and gene therapies offer the potential for transformational gains in health. Some jurisdictions following the cost-effectiveness approach do apply a ‘modifier’ to the threshold of the maximum level of incremental cost-effectiveness they are willing to accept in cases where there is a ‘step-change’ in survival and/or quality of life. There could be further examination of whether and how these modifiers are applied in practice and whether they make a difference in the assessments that are performed. It is more difficult to determine how this issue would be explored in those jurisdictions using the ‘added clinical value ‘ approach, other than comparing the decisions across a range of technology appraisals. One study comparing the assessments made by NICE in England and HAS in France did find a correlation between the level of QALYs gained and level of ASMR awarded [
68]. Perhaps the notion of ‘added clinical value’ could be made more explicit, outlining the nature of the ‘added value’ that is considered relevant, both to patients and the general public, as well as to clinicians.
Thirdly, the application of discounting, within the ‘cost-effectiveness’ approach is critical in the assessment of cell and gene therapies, with their high ‘up-front’ costs and potential long-term benefits. It was noticeable in the analysis of HTA reports that where discounting was applied, the effect of differential discount rates was not explored in many cases. Given the sensitivity of the cost-effectiveness estimate of cell and gene therapies to the discount rate, a sensitivity analysis could be presented to decision-makers in all cases.
Fourthly, insurance value was not considered in any of the HTA assessments examined, probably because no evidence was presented. It is possible that this is a major element of the value of therapies for catastrophic disease and deserves more investigation. However, further research is required to determine whether the insurance value is substantial for these therapies and whether this justifies the application of a modifier within the HTA process.
Finally, a major consideration in all jurisdictions was the uncertainty surrounding cell and gene therapies. Conventional approaches for characterizing uncertainty, such as undertaking probabilistic assessment, are not very helpful when considering issues such as the long-term durability of these therapies, which is often completely unknown. Consideration could be given to the more extensive use of scenario analysis, suggested by Huygens et al. [
16], or the more extensive use of outcome-based managed entry agreements, suggested by Drummond et al. [
4].
Consideration of evidence post-launch
Cell and gene therapies, and advanced therapy medicinal products in general, are often characterized by a high level of uncertainty at market launch over their (long-term) clinical and economic impact. Consequently, ad hoc post launch studies to produce real-world evidence (RWE) could be implemented, if they are needed and feasible. These studies can be linked with outcome-based managed entry agreements in the form of coverage with evidence development: initiating these agreements without an appropriate protocol and data collection may represent missed opportunities.
If post-launch data collection through an ad hoc study is not feasible, a pragmatic approach to data collection is better, e.g., the one adopted by England through Managed Access Agreements that rely on existing data collection programs managed by the industry, and existing NHS registries/administrative databases.
Individual-based agreements rely on existing drugs registries or ‘ad hoc’ data collection on ‘clinical outcome’ indicators. In many cases they do not provide a complete data set to assess the real-world impact of these therapies, but they could serve performance-linked reimbursement agreements, in which payment is made when patients respond to treatment.
In general, more transparency on outcome-based agreements and RWE collected through these agreements is needed, provided that the most sensitive data of the agreement (e.g., the actual price if the therapy is subject to a discounts) are kept confidential. Transparency would provide better feedback to those who are tasked with data collection, and would enhance the replicability of the agreements negotiated, thereby benefiting the health care community as a whole. Guidance for implementing and conducting managed entry agreements has recently been developed for rare disease treatments, which could form a basis for discussions on MEAs in cell and gene therapy [
69].
Acknowledgements
This project was funded by an unrestricted sponsorship agreement between F. Hoffmann-La Roche and Bocconi University. This support is gratefully acknowledged. We are grateful to Thomas Kenny, Tina Taube, Entela Xoxi and 5 other individuals representing industry, patient and payer perspectives, for expressing their views during the discussion of the results of the study, and for comments on an earlier draft of this paper. However, the authors are solely responsible for the views expressed. Preliminary results of this research we presented as a poster at the ISPOR Annual Meeting, National Harbor (MD), May 2022.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.