Working with rare diseases in Germany? A single statistical insight might reshape how you think about data requirements: patients with suspected rare diseases generate healthcare costs 7.6 times higher than matched controls, leaving trails long before diagnosis. But does InGef’s 10.5 million patient database capture enough of that signal for your research question?
Key Takeaways
- The InGef research database’s approximately 10.5 million insured individuals (with 7.8 million actively insured in 2025) represent about 9% of Germany’s total population and a meaningful slice of the statutory health insurance (SHI) population – sufficient for many rare disease research questions, especially for conditions that are not ultra-rare.
- Germany’s FDZ Gesundheit grants access for approved public-interest purposes, such as medical research, health system improvement, and the development of drugs, medical devices, or digital health applications. Projects driven purely by commercial interests – such as market access research without a demonstrable public health benefit – may face eligibility barriers.
- Statistical metrics like confidence intervals, coefficients of variation, and effect sizes – not raw headcount alone – determine whether a dataset is sufficient for a given study.
- Operational flexibility (iterative design, faster timelines, follow-up analyses) is an underrated advantage of commercial databases that can directly affect evidence quality.
- The right data source depends on the specific scientific question – understanding that distinction is where the real analytical work begins, and MEDDDICAL’s coverage of German health data providers and claims data under the GDNG maps out the full landscape.
Picking a data source for rare disease research in Germany should start with the science, not with which dataset is bigger. The InGef database’s approximately 10.5 million insured individuals represent a meaningful slice of Germany’s statutory health insurance (SHI) population – and for a wide range of research questions, that slice is more than enough. Knowing when it’s enough, and when it isn’t, requires looking closely at both statistical thresholds and operational realities.
InGef’s Patient Base: A Commercial Snapshot, Not a National Census
The InGef research database draws from Germany’s statutory health insurance system, which covers roughly 74 million people – about 90% of the population. InGef’s approximately 10.5 million insured individuals (7.8 million actively insured in 2025) represent around 9% of Germany’s total population. That is not a small convenience sample. It is a longitudinal, pseudonymized claims dataset capturing diagnoses, medications, hospitalizations, and healthcare costs across years of routine clinical practice.
What it is not is a national census. The database reflects a specific mix of statutory health insurance funds, which means its demographic and social structure may differ slightly from the broader SHI population. That limitation matters – but for many research questions, it does not disqualify the database. Researchers simply need to be deliberate about what they are estimating and for whom.
Why Rare Disease Research Strains Any Dataset
Approximately four million people in Germany live with a rare disease. But the term covers thousands of individual conditions, most of which affect only a few hundred or a few thousand patients nationwide. That concentration problem is what makes data source selection so consequential.
The Diagnostic Odyssey Leaves a Data Trail
One of the most clinically significant phenomena in rare disease research is the diagnostic odyssey – the prolonged path patients travel before receiving a confirmed diagnosis. A study using claims data from five German statutory health insurance organizations found that patients with suspected rare diseases generate healthcare utilization and costs that are 7.6 times higher than matched controls. That signal is visible in claims data long before a definitive diagnosis is coded. Hospitalizations, specialist referrals, repeat lab orders, and prescription sequences all accumulate in the record, making administrative data a uniquely valuable lens for studying pre-diagnosis burden.
Where Claims Data Fill the RCT Gap
Randomized controlled trials are often not feasible for rare diseases. Patient populations are too small, too geographically dispersed, and too heterogeneous to support conventional trial designs. Real-world data (RWD) and real-world evidence (RWE) fill that gap by capturing what actually happens in routine clinical care – treatment patterns, healthcare resource utilization, disease progression, and long-term outcomes. Claims data are particularly well-suited to this role because they cover the entire care continuum, not just the clinical trial window.
FDZ vs. InGef: The Real Tradeoff
What FDZ Gesundheit Actually Covers
The FDZ Gesundheit (Research Data Center for Health) provides access to pseudonymized billing data from all statutory health insurers in Germany – approximately 95 funds – representing the full SHI population. The longitudinal data go back to 2009, and future integration with electronic patient records (ePA) is planned for October 2026. For a researcher studying an ultra-rare condition with 300 known cases nationwide, the FDZ is the logical starting point.
Access is granted for approved public-interest purposes, including medical research, health system improvement, and the development of drugs, medical devices, or digital health applications. Both public and private researchers may be eligible. However, projects driven purely by commercial interests – such as market access research without a demonstrable public health benefit – may face eligibility barriers. For a significant portion of HEOR and pharmaceutical evidence generation work, that distinction is a practical constraint worth assessing early.
What the InGef Commercial Database Offers Instead
InGef’s commercial database trades population breadth for accessibility and flexibility. Studies do not require the same statutory application process. Timelines are more predictable. Analytical environments can be adapted as research questions evolve. For organizations working within drug development or market access timelines, these are not minor conveniences – they are fundamental to whether a study can be completed at all.
Commercial claims databases like InGef have already demonstrated their utility in rare disease research. Published analyses have used the InGef database for cross-sectional studies of Huntington’s disease, Beta-Thalassemia, and Spinal Muscular Atrophy type 1 – all conditions that qualify as rare under the EU definition. Those studies produced robust population characterizations from the available patient base.
When InGef’s Patient Base Is Statistically Enough
Conditions Where Commercial Data Has Proven Sufficient
Published evidence suggests commercial databases perform well for conditions with prevalence in the range of roughly 1 in 2,000 or higher. At that prevalence, the InGef database yields case counts in the thousands – enough for meaningful subgroup analyses, treatment pathway mapping, and healthcare cost estimation. Conditions like Huntington’s disease, certain lysosomal storage disorders, and hereditary metabolic diseases have been successfully characterized using commercial SHI data in Germany.
The Metrics That Determine Sufficiency
Dataset size alone is a proxy, not a measure. The actual determinants of statistical sufficiency are:
- Confidence intervals – How wide is the uncertainty band around the estimate? Wider intervals signal underpowered analyses.
- Coefficients of variation (CV) – A CV above 30% (or more precisely, 33.3% as per some statistical guidelines) typically indicates unreliable estimates.
- Standard errors – Smaller standard errors reflect more precise parameter estimates.
- Effect size – Larger effects require smaller samples to detect reliably.
- Margin of error – For prevalence estimates, what level of precision is scientifically acceptable?
For many HEOR questions – burden of disease, healthcare resource utilization, treatment patterns – these metrics often land within acceptable bounds using InGef’s patient base. The practical difference between a 10.5 million-patient analysis and a 74 million-patient analysis is, in many cases, a narrower confidence interval, not a different conclusion.
When Nationwide Coverage Is Non-Negotiable
Ultra-Rare Conditions and the Extrapolation Problem
For conditions with a prevalence below roughly 1 in 10,000 – true ultra-rare diseases – the arithmetic becomes unforgiving. At those case counts, confidence intervals widen dramatically, subgroup analyses become unreliable, and any finding requires extrapolation to the national population, adding a layer of uncertainty that may not be scientifically defensible.
This is where the FDZ’s nationwide coverage changes the evidence quality meaningfully. When every additional patient in the dataset materially improves statistical certainty, maximizing case ascertainment is a scientific requirement. Federated data networks (FDNs), which combine harmonized data from multiple statutory sources, are also being developed specifically to address this challenge for geographically dispersed rare disease populations.
This makes the choice between the FDZ Gesundheit and commercial claims databases particularly important for rare disease studies. See FDZ Gesundheit vs Commercial Claims Data Providers for a detailed comparison.
Operational Flexibility: The Underrated Variable
Rare disease studies rarely unfold in a straight line. As analysis progresses, researchers often identify new subgroups worth examining, encounter unexpected coding patterns, or need to refine inclusion criteria based on early results. Commercial databases accommodate this iterative process – researchers can adjust study designs, run additional analyses, and revisit datasets for follow-up publications without restarting a formal access application.
The FDZ, by design, works differently. Access is granted to a predefined dataset within a predefined analytical environment, established before analysis begins. That structure serves important data governance purposes, but it can limit the flexibility that rare disease research often demands. When a study needs to evolve, the operational model of the data source becomes as important as its statistical power.
Six Questions to Ask Before Choosing a Data Source
Before committing to either pathway, research teams should work through these questions:
- How rare is the condition? – Conditions close to the 1-in-10,000 threshold require careful power calculations before assuming commercial data will suffice.
- Is nationwide coverage scientifically necessary? – Or would a large, representative commercial sample answer the research question equally well?
- Will the study design need to evolve? – If iterative analysis is likely, operational flexibility belongs in the decision framework.
- What are the timeline constraints? – FDZ applications follow a statutory process with defined review periods; commercial databases offer faster initiation.
- Does the project qualify for statutory access? – Projects without a demonstrable public health benefit may not meet FDZ eligibility criteria.
- Are follow-up studies or updated analyses anticipated? – Ongoing access and update cycles favor commercial providers.
For Most HEOR Questions, InGef Is the Pragmatic Starting Point
For rare disease researchers and HEOR professionals working in Germany, the InGef research database represents a pragmatic and statistically defensible starting point for most evidence generation tasks. It covers a large, longitudinal cohort of insured patients, carries a published track record in rare disease cross-sectional research, and operates under a commercial access model that supports the timelines and iterative workflows that pharmaceutical development requires.
Nationwide FDZ data remains the right choice when ultra-rare conditions demand maximum case ascertainment, when extrapolation is scientifically unacceptable, and when the research purpose meets statutory eligibility criteria. Those situations are real – but they are not the default case for every rare disease HEOR study.
For studies requiring the maximum possible number of patients, the nationwide coverage available through the FDZ may be particularly relevant. The implications of FDZ coverage for Data Model 3 should also be considered when designing longitudinal studies.
InGef’s patient base suffices when the statistical metrics confirm it, when the condition is not ultra-rare, and when operational flexibility matters. That covers a substantial portion of the rare disease research agenda in Germany.
For a deeper look at Germany’s evolving health data infrastructure – including how the GDNG shapes data access across both statutory and commercial pathways – MEDDDICAL provides ongoing analysis and guidance for research teams working through these decisions.