Robust
analytics: our
strategic approach

Many statistical consultancies begin an engagement by asking: What methods can we run on this data?

We take a different approach.

We don’t start with the method. We start with the decision.

We believe this is the right place to start.

We begin with the specific decision that the evidence needs to support and the requirements against which it will be assessed.

Advanced methodologies such as matching-adjusted indirect comparisons (MAIC), fractional polynomial models or the integration of real-world evidence (RWE) are therefore not the starting point. They are selected according to the decision context, the statistical challenge and the uncertainty that needs to be resolved.

Our three-step approach

Assessment context

Understand the context

before defining the

analysis.

We first establish the decision context in which the evidence will be assessed.

➝ The assessment framework

➝ Population and comparator requirements

➝ Key evidence questions

➝ Regulatory and HTA expectations

➝ Market-specific requirements

The same evidence can be interpreted differently across markets and assessment contexts.

What does the analysis need to demonstrate — and under which assessment requirements?

Statistical focus

Identify

areas of bias and

uncertainty.

We then identify the statistical questions that frame the key evidence issues.

➝ Study and population comparability

➝ Treatment effect modifiers and endpoint alignment

➝ Missing data

➝ Heterogeneity

➝ Modelling assumptions

We make uncertainty explicit, showing what the evidence supports and where conclusions remain conditional.

Where do bias and uncertainty enter the evidence — and how can they be characterised and reduced?

Method selection

Select the

method that fits the

evidence question.

We then select the methodology that best addresses the evidence question.

➝ MAIC

➝ ML-NMR

➝ STC

➝ Network meta-analysis

➝ Survival modelling

➝ Real-world evidence

➝ Sensitivity analyses

➝ Quantitative Bias Analysis (QBA)

➝ Other statistical approaches

Methods are tools, not strategies.
The appropriate method is the one that best addresses the underlying evidence question while making assumptions, limitations and remaining uncertainty transparent.

Which analytical approach provides the most defensible answer to the decision question?