Impact of the new dossier requirements

Since 2011, pharmaceutical companies in Germany have been required to submit an AMNOG benefit dossier to the G-BA when entering the market or expanding the approval of a new drug. On the basis of the IQWiG benefit assessment, the G-BA carries out an assessment of the additional benefit of the drug and largely follows the assessment by IQWiG. In April 2020, a revised, currently valid dossier template for Module 4 came into force, which drastically increases the already large presentation effort for pharmaceutical companies. However, given the increased presentation effort, a significant proportion does not appear to be relevant to the conclusions for the additional benefit awarded by the G-BA. In an efficient AMNOG process, only evaluations with actual added value should be made mandatory in the future.

>> In April 2020, a revised dossier template for Module 4 came into force (Joint Federal Committee, 2019), which placed a number of additional requirements on data preparation and significantly increased the scope of this module compared to the previous dossier template (Joint Federal Committee, 2013) (Bartol et al., 2022). The current requirements include, among other things: the presentation of additional data cut-offs for all endpoints (EP), detailed evaluations of EP on adverse events (AEs) as well as subgroup (SG) analyzes for all EP (Association of Researching Drug Manufacturers, 2021). These adjustments should increase the traceability of the available data and align the quality of the assigned information between the benefit dossiers (Joint Federal Committee, 2019).
The sharp increase in requirements led to an increasing scope of Module 4 and is associated with considerable additional effort and additional resource consumption for companies, IQWiG and G-BA (Association of Research-Based Drug Manufacturers, 2021). A large part of the information presented is not taken into account in the benefit assessment process, as an analysis of ten procedures with current dossier templates from the Association of Research-Based Drug Manufacturers (vfa) showed in 2021 (Association of Research-Based Drug Manufacturers, 2021).
It is clear that the IQWiG already presented less for a large number of endpoints on adverse events and subgroups than was presented by the pharmaceutical companies. This discrepancy increases with the current dossier template. Given the increased presentation effort for pharmaceutical companies, a significant proportion of the additionally presented endpoints on adverse events and subgroups do not appear to be relevant to the additional benefit. Only serious adverse events and serious adverse events as well as the subgroups age, gender and disease severity appear to be relevant.
In general, it is not clear which objective criteria IQWiG and G-BA use to select the endpoints on adverse events and subgroups relevant for evaluation. In an AMNOG process that is efficient for everyone involved, only evaluations with actual added value should be presented in the future. It seems urgently necessary to adapt the dossier template accordingly.
Question
In view of the greatly increased requirements within the context of the benefit dossier, the fundamental question arises as to which data are necessary and sufficient for an efficiently designed benefit assessment that, on the one hand, requires a high degree of transparency on the part of the company and, on the other hand, a high quality of the assessment by IQWiG and the decision ensured by the G‑BA. Specifically, the present analysis therefore asks the question of how the additional data required in the current dossier templates in Module 4 are actually taken into account in the assessment.
As part of this question, we systematically examine what proportion of the data evaluations required by the companies are presented in the IQWiG benefit assessment. In addition, it is analyzed how this proportion has changed with the current compared to the old dossier template and to what extent the required evaluations are explicitly relevant to the determination of the ZN by the G‑BA. The focus of the analyzes is on the evaluations of UE and SG, as the most far-reaching changes were made in these two areas.
methods
To answer the questions, the company's benefit dossiers, the associated benefit assessments of the IQWiG as well as resolution texts and supporting reasons (TrG) of the G‑BA were examined with regard to the presented EP on AEs and SG characteristics. For this purpose, procedures were selected based on old and current dossier templates. Subsequently, data relevant to the investigation of the question on the primary data cut taken into account in the benefit assessment process was extracted from the documents described above from the company, IQWiG and G‑BA.
Selection of procedures
The selection of procedures based on the current dossier template included all procedures with a procedure start from April 1, 2020 (publication of the IQWiG benefit assessment from July 1, 2020 to June 15, 2021, G-BA decision text and TrG from September 17, 2020 to September 15, 2021). The selection of procedures based on the old dossier template included procedures up to April 1, 2020. In order to improve the comparability of the procedures, care was taken to assign each procedure with a current dossier template to the most up-to-date counterpart of a procedure with an old dossier template in a comparable indication area or comparable question. In order to fully examine the requirements of the dossier template, orphan drug procedures, procedures with single-arm studies, and procedures in which no data were regularly submitted (particularly with regard to AE and SG analyses) or evaluated by the G‑BA were excluded . If it became clear during the data extraction that there was no usable data for a procedure, an attempt was made to find a suitable alternative according to the same criteria. For two procedures based on the current dossier template, no suitable procedure based on the old dossier template could be identified.
data extraction and
quality assurance
For data extraction, a reviewer extracted the information on evaluations of UE-EP or information on SG in the company's benefit dossiers as well as in the benefit assessments or addenda of the IQWiG and the resolutions or TrG of the G-BA. The quality control of the extraction was carried out by a second independent reviewer. Discrepancies were consolidated jointly by both reviewers. In addition to higher-level, general information such as product name, date of the G‑BA decision or whether the old or current dossier template was used, data on the UE and the SG were primarily collected. With regard to AEs, the number of AEs-EP was recorded differentiated according to different categories and severity as well as SOC, PT and the number of adverse events of special interest (UESI). For IQWiG benefit assessments, the EPs for AEs were extracted from the results section including the appendix tables. For the SG, the representation of the mandatory SG characteristics of age, gender, ethnicity/region and disease severity as well as any other SG characteristics presented were recorded. In addition, it was examined whether the data recorded for UE and SG were also assessed as relevant for the determination of the ZN.
If the company presented data on different study phases, the data representing the entire study period were extracted. If the ZN derivation of the G-BA was based on a meta-analytic summary of several studies or an indirect comparison, only the data from the meta-analysis or indirect comparison were extracted. If several non-meta-analytically summarized studies were presented in one procedure, if only one study was presented in the partner procedure, only the information from the study that was most similar to the study in the assigned partner procedure in terms of design, patient numbers, etc. was extracted.
Statistical analyses
The data was analyzed descriptively. For categorical data, frequencies and percentages of the respective categories were reported. For continuous data, descriptive parameters such as median, arithmetic mean, minimum and maximum were presented. The analyzes were carried out separately according to the respective institutions pU, IQWiG and G‑BA as well as separately according to the old and current dossier template.
For the descriptive analysis of the total presentation effort for UE and SG, the variables “All endpoints for UE (including SOC/PT)” and “Total number of SG characteristics” were created. The first variable included a summation of the number of all AEs of any severity, SAEs, severe AEs, non-severe AEs, discontinuation due to AEs after SOC and PT and all non-severe, severe and serious AEs. The second variable contained a summation of all SG characteristics presented.
Results
The present analysis examined what proportion of the data evaluations required by the companies were presented in the benefit assessment of the IQWiG (perspective of the IQWiG) and to what extent the required evaluations were explicitly relevant to the conclusion of the ZN (perspective of the G‑BA). Furthermore, it was analyzed how the proportion of analyzes taken into account changed for all participants (company, IQWiG and G-BA) with the current compared to the old dossier template.

IQWiG perspective
A total of 94 procedures were selected for the comparison of dossier requirements. Of these, 46 were based on the old dossier template (Joint Federal Committee, 2013) and 48 on the current dossier template (Joint Federal Committee, 2019). The indication areas included oncological diseases (47 procedures in total), diseases of the respiratory system (n = 19), the musculoskeletal system (n = 7), the nervous system (n = 3) as well as mental diseases (n = 2), heart -Circulatory diseases (n = 2), infectious diseases (n = 5), skin diseases (n = 4), metabolic diseases (n = 4) and eye diseases
(n = 1). Within the indication areas, the distribution between procedures based on the old vs. the current dossier template was comparable.
Table 1 gives an overview of the number of EPs on UE presented by the company and IQWiG for the old and the current dossier template. Compared to the old dossier template, the median number of representations in the current dossier template has increased many times over for almost every AE endpoint category. There was also a discrepancy between the company's presentation effort and that of IQWiG, especially for the current dossier template. Only for the outcome categories SUE after SOC/PT and non-severe AE after SOC/PT was the median ratio of representations mathematically balanced, but the latter category was generally not presented by IQWiG. In contrast, the company found a maximum of up to 91 non-severe AEs according to SOC/PT under the current dossier template.
endpoints to undesirable ones
events
Table 1 shows a summary overview of the presentation of the EP on AEs in the company's benefit dossiers, the IQWiG benefit assessments and the difference between the two. A distinction is made between old and current dossier templates.
Using the current dossier template, the companies presented a median of 20% more analyzes for any AE according to SOC/PT than IQWiG. In comparison to the company, IQWiG presented a median of 87% fewer evaluations of severe AEs after SOC/PT. With the old dossier template, a maximum of 101 serious AEs after SOC/PT were submitted; with the current dossier template, up to 574 evaluations were submitted for this EP category submitted. The IQWiG, on the other hand, presented a maximum of less than 50 severe AEs according to SOC/PT under both the old and current dossier template. In the category of discontinuation due to AEs after SOC/PT, the IQWiG presented a median of only two thirds of analyzes compared to the company. A particularly drastic picture emerged for the UESI, where the presentation effort of the IQWiG compared to the companies even fell to less than a third of the median. When considering all EP on UE including all SOC and PT, the number of representations for the company almost tripled when comparing the old dossier template with the current one. On the IQWiG side, however, an increase of only 69% can be seen. Thus, compared to the company, IQWiG presented a median of only 67% of the number of evaluations in this category. The range of representations also made clear the discrepancy between the two parties. While the company presented a maximum of 2590 EP for the category All EP on UE including SOC/PT, the maximum representation of the IQWiG was only 822 EP.

Subgroups
The dossier template calls for a mandatory presentation of SG analyzes for the SG age, gender, ethnicity/region and disease severity. These SGs should therefore be regularly presented by the companies in the benefit dossiers. In the 46 evaluated procedures with old dossier templates and the 48 evaluated procedures with current dossier templates, representations of these SGs were found in more than 95% of the benefit dossiers.
Table 2 shows the total number of SG features presented. With the switch to the current dossier template, the median number of SG characteristics presented increased slightly for the company, but decreased slightly for IQWiG. For the current dossier template, it was shown that the IQWiG only represents 27% of SG characteristics compared to the company. The different presentation effort was also evident from the maximum SG features represented, which were 24 for the company and only 6 features for IQWiG. In addition, the number of SG features must be interpreted in the context of the large increase in EP (on which SG analyzes are calculated). Thus, the slight increase in the number of SG features is potentiated by the large increase in represented EP, leading to a greatly increased number of SG analyzes overall.
For the current dossier template, the SG age was taken into account in the IQWiG benefit assessment in 81% cases. The SG characteristic gender was taken into account in 73% of the cases. While the SG ethnicity/region presented by the company was still considered in 61% of the cases using the old dossier template from IQWiG, the relative number of representations taken into account decreased here to 23% using the current dossier template. For the current dossier template, three quarters of the representations on this SG were not taken into account in the IQWiG benefit assessment. The SG disease severity was taken into account for the current dossier template by IQWiG in almost half of the cases (48%).
In addition to the SG mentioned above, the companies are required to present results of SG analyzes planned a priori and specified in the study protocol for patient-relevant EP (Joint Federal Committee, 2021). Further analyzes showed that outside of the median SG mentioned above, no other SG were taken into account by IQWiG in the benefit assessment. This applies to both the old and the current dossier template. In some cases, evaluations of disease-specific SG were taken into account in the other SG characteristics.
Table 2 shows the total number of SG features represented. The SG analyzes presented by the companies in the benefit dossier and by IQWiG in the benefit assessment as well as the difference between the two are compared. A distinction is made between old and current dossier templates.

Perspective of the G‑BA
Following the IQWiG benefit assessment, the G‑BA makes a decision on the ZN of the zbAM based on the available data and explains this in more detail in the TrG (Joint Federal Committee, 2022).
The analysis of all 94 selected assessment procedures showed that the G-BA in the TrG follows the IQWiG with regard to the selection of the UE-EP presented in most cases (76%). When viewed separately, comparable values result for procedures with old and current dossier templates. Isolated AEs that have a statistically significant effect are occasionally presented to support the assessment and derivation of the ZN. The comparison showed that using the current dossier template increased the proportion of these supportively presented AEs that were included in the consideration of the ZN (43% vs. 56%).
The selection of the SG age, gender, ethnicity/region and disease severity from the IQWiG was adopted by the G‑BA in at least half of the cases in both procedures with old and current dossier templates and was usually not commented on in the TrG (93% vs. 85%). In most cases, the selection of the SG was not relevant to the conclusion of the ZN awarded by the G‑BA (72%); in 18% of the cases, an evaluation of the data was not possible. When viewed separately, comparable values emerged for procedures with old and procedures with current dossier templates (76% vs. 69%).
discussion
Discrepancies between presented and assessed analyses
The additional evaluations required in the current dossier template result in an increased analysis and presentation effort for UE-EP and SG for the company. The dossier assessment is also becoming more complex for the IQWiG and partly for the G-BA. Particularly in the EP categories severe AE, UESI and all EP to AE, each according to SOC/PT, there is a sometimes significantly increased analysis and presentation effort. The present analysis clearly illustrates a large discrepancy between the number of EPs presented by the companies and the number of EPs that are assessed in the IQWiG benefit assessment. These results correspond to the results of a study carried out in 2021 (Association of Researching Drug Manufacturers, 2021) and, in addition to a significantly larger number of procedures, also provide current findings in the comparison of old and current dossier templates. Even in the old dossier template, IQWiG often did not take the AE analyzes presented into account. This gap has widened further with the introduction of the current dossier template. With regard to the presentation of SG characteristics, the present analysis also shows that, with both the old and the current dossier template, the company presents more SGs than are presented in the IQWiG benefit assessment. Since the G‑BA mostly follows the assessment by the IQWiG, there were already non-conclusion-relevant evaluations in the old dossier template, the number of which has now increased further with the current template.

Conceivable approaches for a more efficient Module 4 as part of a “learning system”
Even before the introduction of the current dossier template, the German benefit assessment process was considered to be incomparably extensive (Ivandic, 2014) and those responsible have described the German benefit assessment process since its launch as a “learning system” (Bokemeyer C., 2017), which has repeatedly been modified and modified in detail is further developed (Joint Federal Committee, 2019).
As can be seen from the present analysis, the current AMNOG process is overall more inefficient than before the change in April 2020, so an adjustment to the dossier template seems necessary.
For the evaluation and derivation of the ZN, isolated AEs that have a statistically significant effect are occasionally presented to support the assessment. However, mandatory required AE-EP, such as non-severe AEs and any AEs after SOC/PT, do not appear relevant for ZN derivation. It therefore seems logical to lift the general obligation to present this UE-EP. Even with regard to the scope of the UESI to be presented by the companies, it remains unclear which objective criteria or methodology the IQWiG uses to select the UESI for the benefit assessment based on the event frequencies that occurred in the studies. The IQWiG approach is not entirely based on medical or biometric rationale. While the obligation to present pre-specified, i.e. medically or biologically based, UESI is correct and understandable, the obligation to present UESI not predefined in the study program based on severity should be eliminated entirely. Mere signal detection does not make sense in the context of benefit assessment.
In addition to the SG characteristics of gender, age, disease severity, severity or stage of the disease as well as country and study center effects (Joint Federal Committee, 2022), additional, pre-specified SGs are added individually for each procedure.
For all of these SGs, analyzes for all safety EPs including UESI, SOC/PT (in the case of significant SOC or PT) must be submitted according to the dossier template, which multiplies the necessary UE analyzes for each additional SG. A division of decisions according to SG characteristics is extremely rare and subgroup effects in individual AEs according to SOC and/or PT did not play a relevant role in the decision-making process. The present analysis emphasizes that a general obligation to present all prespecified SG analyzes for all endpoints should be avoided. The SG ethnic/region shows particularly clearly that with both the old and the current dossier template the company presented more SGs than were presented in the benefit assessment procedure by IQWiG. The importance of this SG has decreased significantly with the current dossier template. However, the SG age, gender and, in some cases, disease severity seem to be particularly relevant. Other SG characteristics only play a role in isolated cases. A central point for the definition of SG in the context of benefit assessment is external evidence, i.e. the biological or sociological rationale (DGHO, 2016). However, precisely because of their lack of biological and biometric rationale, the SGs presented are usually not included in the discussion about the derivation of the ZN.
There is currently a lack of traceability as to why
half and under what conditions the required
Further analyzes of UE and SG by IQWiG and G‑BA are not taken into account. Since the G-BA does not comment on AEs and SGs that are not relevant to the conclusion as part of its decision, it also remains unclear to what extent these data are required or used for the assessment of the ZN. The lack of discussion of the SG presented also emerges from another study from 2021. Using the current dossier template, a total of only 23% of the evaluations presented by the companies were taken into account; on average, only a maximum of 14% of the SG analyzes presented by the companies were taken into account by IQWiG and/or G‑BA in the assessment (Vfa, 2021) . In general, if the analyzes presented by IQWiG and G‑BA are not taken into account, reasons should be given, as is the case with excluded analyses. All evaluations taken into account should be commented on, even if only briefly, in the IQWiG benefit assessment or G‑BA decision. Continuous commentary from IQWiG and G‑BA is also required with regard to results relevant to the assessment and conclusions.
In order to make decisions here understandable and their appropriateness assessable, a discussion between the G-BA and IQWiG on the analyzes required individually for each benefit assessment as part of a consultation appears to be a sensible solution. Here, the existing consultations could generally be used to determine in advance the analyzes to be presented for the respective indication or product class in an exchange between all those involved in the AMNOG benefit assessment. For example, it would be conceivable to have a “core set” of central EP defined by the G‑BA – such as overall survival or SAE – for each assessment procedure. Only for these individually relevant EP should it be mandatory to submit SG analyses, namely of pre-specified and medically-biologically justified SG. Given the findings from this analysis, a general presentation of the subgroup analyzes for AEs after SOC and PT should be completely avoided. With regard to the endpoints for AEs, the evaluations of non-severe AEs can be dispensed with and overall a discussion about adjusting the limit for mapping any AEs according to SOC and PT is necessary. The corresponding adjustment of the dossier template and the exchange between IQWiG and G‑BA with the companies in advance as part of a consultation appears more plausible and efficient than the current signal detection and offers the desired traceability of the methodology for everyone involved.
Conclusion
The German benefit assessment procedure is considered a learning system. However, since the current AMNOG process is overall more inefficient than before the change in April 2020, the dossier template needs to be adjusted again. Only evaluations with actual added value for the benefit assessment should be presented in a mandatory manner. However, the obligation to present further evaluations should no longer apply. The evaluations of UE and SG relevant to the respective benefit dossier should be specified in advance in order to ensure improved traceability of the procedures of IQWiG and G‑BA. <<

from:
Dr. rer. nat. Maria Katharina Schweitzer1 / Dr. rer. nat. Manuel Nico Dold2 /
Dr. rer. nat. Astrid Genet3 /
Klaus Gossens PhD4 /
Dipl.-Math. Thomas Klein-Hessling5 /
Nils Löffler MSc6 /
Dr. phil. Matthias Rabel7 /
Dr. PH Andrej Rasch8 /
Jessica Schmelcher MSc9 /
Dr. rer. nat. Sebastian Werner10 /
Natalia Wolfram11

background
Since the introduction of the Medicines Market Reorganization Act (AMNOG) in 2011, pharmaceutical companies (companies) in Germany have been obliged to submit a dossier for benefit assessment in accordance with Section 35a SGB V when placing new drugs on the market or expanding their approval (Bundestag, 2010). On the basis of this benefit dossier and on the basis of the dossier assessment by the Institute for Efficiency in Health Care (IQWiG), the Federal Joint Committee (G‑BA) decides on an additional benefit (ZN) of the drug to be evaluated (zbAM) compared to a comparator determined by the G‑BA appropriate comparative therapy (zVT). To create the benefit dossier, which is made up of five different modules, specified document templates must be used (Joint Federal Committee, 2022). These dossier templates request different information about the zbAM. Module 4 in particular requires extensive and very detailed preparation of study data and evaluations by the company.