Will German registries be good enough to close AMNOG’s evidence gaps?

“We have to identify the evidence gaps that absolutely need to be closed.” This statement was made by Josef Hecken, the Chairman of Germany’s Gemeinsamer Bundesausschuss (GBA; Federal Joint Committee), at a conference on real-world evidence (RWE) on 18th February 2020. Recently enacted legislation has reaffirmed the GBA’s authority to require the collection of RWE from the time a new drug is launched.

Speaking at the same event, Thomas Kaiser, the Head of Drug Assessment at the Institut für Qualit?t und Wirtschaftlichkeit im Gesundheitswesen (IQWiG; Institute for Quality and Efficiency in Health Care), reported that data on patient-relevant endpoints were the most important deficiencies in submissions it reviewed. In addition, data on control groups were frequently missing, even when other drugs for a given indication were already on the market.

Mistrust of electronic patient records and claims data

Having identified the evidence gaps, the next challenge is to find suitable ways to close them. According to Hecken, electronic patient records simply are not widely enough available to fill this need, and health insurance funds’ claims data are invalid as a means of extracting data on patient-relevant endpoints such as morbidity, mortality, quality of life or side effects. He commented: “At best, I can see if the patient is alive. If not, however, I can barely tell what the cause of death was.”

IQWiG shares this scepticism about the potential utility of electronic patient records and claims data as sources of RWE. In a recently published study conducted on behalf of the GBA, IQWiG expressed the belief that “the collection and processing of routine practice data from electronic patient records and claims data from health insurance funds is currently not possible with regard to benefit assessments of drugs and will not be possible in the near future. This is mainly because the data quality in these sources is insufficient and important data are not collected. These problems cannot be solved in the short or medium term.”

Are registry-based studies the answer?

The IQWiG report recommended a role only for registry-based studies in the AMNOG process. The authors concluded that it would be feasible to use RWE from registry-based studies for early benefit assessment, provided that the quality of the data was high enough. Such studies could be either randomised or non-randomised, but randomisation would be necessary for a drug to be judged to have more than just a hint of an effect. The acceptability of retrospective study designs would depend on whether the available data sources contained the necessary data in the necessary quality, which would require use of the same data sources (e.g., disease-specific clinical registries) in both studies.

IQWiG believes that randomised “large simple trials” in routine clinical practice could readily be conducted, particularly after marketing authorisation. The institute suggests that randomised registry-based studies could be faster, and less complex and resource-intensive, than conventional trials.

For registry-based studies to be useful for the AMNOG process, Kaiser believes that they would need to include patient-relevant endpoints that go beyond the scope of the data usually collected in routine care. In addition, robust, standardised study protocols would be needed.

Cancer registries—a cautionary tale

The experience of cancer registries perhaps exemplifies the difficulties of building a robust national network. The first such registries were established five years ago with the objective of collecting comprehensive national data on the epidemiology, diagnosis and treatment of cancers, and thereby improving the quality of care.

A fundamental obstacle to this objective, however, is that Germany’s states have each built their own clinical cancer registries, based on individual organisational principles and using different software. Moreover, data are not readily available to researchers in a usable format and have to be requested on an ad hoc basis. A working group now meets every six weeks to try to find ways to harmonise data sets from the various regional cancer registries.

An initiative by the GBA to use data from the registries to assess the impact of cancer screening programmes on incidence, disease progression and mortality proved unsuccessful. The cancer registries are now in discussion with the GBA about substantially expanding the scope of their data sets to meet the committee’s requirements for the AMNOG process.

Need to improve the quality of German registries

Jürgen Windeler, IQWiG’s Director, believes the first data from high-quality registries will soon be available for use in the AMNOG process. However, he concedes that “the conditions for high-quality registries could be better. This concerns both funding and the fact that there are different requirements for data protection in different German federal states.”

On the issue of funding, Josef Hecken has proposed that the GBA should finance the new registries. In his view, “that would have the big advantage that we would also then be master of the data.”

Wary response from the pharmaceutical industry

The pharmaceutical industry has called on the GBA to accept a broad range of RWE, including data from electronic patient records and claims data from health insurance funds. Markus Frick, the Managing Director of the Verband Forschender Arzneimittelhersteller (VFA; Association of Research-Based Pharmaceutical Companies), recently remarked that “there is actually no reason why data that are generated anyway, and that are used in everyday care and are easier to access due to digitalisation, should not be used for additional benefit assessment.” This appeal is likely to fall on deaf ears.

Rather, manufacturers need to prepare for the prospect of registry-based studies. The industry argues that the GBA should mandate such research only in cases where the EMA has likewise requested similar studies on clearly defined grounds. Furthermore, the industry believes it would be essential to set specific endpoints, control costs (particularly for orphan drugs) and ensure that the administrative burden on healthcare professionals is not excessive.

Drug manufacturers will certainly want the GBA to ensure that a future network of registries generating data for the AMNOG process is more coherent than the current cancer registries: having to meet the disparate requirements of 16 regional registries would be too onerous and costly. Companies may also have some misgivings about the idea of GBA “ownership” of the registries. Alignment of regulatory and HTA requirements for post-marketing research would help to allay some of the industry’s concerns regarding the additional expense and workload of this activity, but it remains to be seen if the GBA will take its lead from the EMA.

The GBA’s strategy for incorporating RWE into the AMNOG process should become clearer in the coming months. It will be an anxious wait for the pharmaceutical industry.