The reference case and the base case are two distinct building blocks of a health-economic evaluation. The reference case describes methodological standards intended to make analyses comparable and quality-assured. The base case is the specific model calculation with defined, non-varied parameter values from which sensitivity analyses start.
The reference case as a methodological framework
IQWiG uses the reference case in particular for cost-benefit assessments in the context of Section 35b SGB V. It consolidates specifications on the decision problem, the population, the intervention and comparators, the outcome measure, the perspective, and the time horizon. Added to this are requirements for evidence, modelling, costs, presentation of results, and the handling of uncertainty. It is therefore an orientation framework, not the result calculation for a single product.
The value of a reference case is that different health-economic analyses can be planned and read according to comparable principles. It does not eliminate every methodological decision in individual cases. Rather, it makes transparent which decisions deviate from the standard assumption with justification and which information is required to assess a model.
The base case as the starting calculation
For a specific analysis, the base case defines values and model assumptions that are not varied initially. It provides the primary estimate for costs, benefits, and incremental outcomes. In a Markov model, these may include, for example, transition probabilities, utility values, resource use, and the time horizon. The base case must appropriately reflect the research question, but it is not a claim that its parameters are certain or without alternatives.
Starting from this initial calculation, sensitivity analyses examine how robust the result remains under changed assumptions. Parameter, structural, or scenario analyses can show whether a conclusion depends on a single input. A carefully described base case therefore clearly separates the central model configuration from deliberately deviating variants.
Distinction: base case and baseline
The base case of a health-economic model calculation is not the baseline value of a clinical study. The established term baseline usually refers to measurements of study participants before treatment begins, such as a symptom score or a baseline characteristic. Such values can influence model parameters, but they are not themselves the base case.
Likewise, the reference case is not the same as a comparator therapy. In the German benefit assessment process, the appropriate comparator therapy determines the clinical comparison. The reference case, by contrast, determines the methodological standards according to which a cost-benefit assessment should be constructed. Both levels can come together in an analysis, but they serve different functions.
A reference case also specifies how sources are searched for, selected, and transferred into an analysis. It is therefore not a checklist that could replace a clinical assessment by mechanically filling it in. If multiple data sources lead to different effect or cost values, the model needs a subject-matter justification for the selection. The reference case provides the language in which such justifications can be reviewed and compared across different analyses.
In the base case, every assumption should be traceable: table values, conversions, derived transitions, and simplifications made belong in technical documentation. This applies in particular when study endpoints do not directly correspond to model states. In that case, mapping from surrogates, time-dependent trajectories, or the aggregation of multiple events may be required. Sensitivity analyses only gain interpretive value when it is clear which precise base assumption they change.
A plausible base case does not necessarily have to show the most favourable scenario for an intervention. Its task is a justified primary presentation, not the selection of an advantageous parameter combination.
The conceptual separation also facilitates the review of a report: for the reference case, the question is whether the methodological standards are met or deviations are explained. For the base case, the review checks whether its specific figures, formulas, and model assumptions match the research question.
This distinction also prevents sensitivity analyses from becoming a silent substitute calculation. Deviations from the base case are only interpretable if each change is named individually and its impact on costs and benefits is made visible.
Relevance for clinical trials
Studies provide data that can feed into the base case, for example effect measures, adverse event rates, utility values, and resource use. Even when planning data collection, it should be clear which data are needed for model states or transitions and whether they apply to the relevant population. If parameter selection and data transformations are not documented transparently, the path from study result to base case can hardly be verified later.
Full-service CROs such as Mediconomics support this by linking clinical data management with health-economic model planning. They can document sources for base-case parameters, align analytical decisions with the study protocol, and coordinate reporting materials for scenario and sensitivity analyses so that the reference case and the specific model calculation are presented consistently.
Frequently asked questions (FAQ)
Is the reference case a single mandatory model formula?
No. It consists of methodological standards and requirements that structure the planning and assessment of health-economic evaluations.
Why is the base case varied?
Parameters and model assumptions are subject to uncertainty; variations show whether the result rests on robust or fragile premises.
Can a baseline value be part of a base case?
Yes. A baseline value can be used as a model parameter, but conceptually it remains distinct from the overall base-case configuration.
Regulatory references
- Section 35b SGB V – provides the legal context for cost-benefit assessment.
- IQWiG, General Methods Version 8.0 – defines the reference case and base case.
- Rules of Procedure of the G-BA, Chapter 5 – places medicinal product assessments within the G-BA process.