Considerations underlying the use of mixed group validation.

Area of Science:

• Biostatistics
• Diagnostic Test Evaluation
• Clinical Research Methodology

Background:

• Mixed Group Validation (MGV) is an emerging method for estimating diagnostic accuracy.
• It presents an alternative to Known Groups Validation (KGV), notably not requiring a perfect gold standard.
• However, optimal research designs and the applicability of MGV assumptions in clinical settings require further investigation.

Purpose of the Study:

• To explore optimal research designs for Mixed Group Validation (MGV) studies.
• To assess the validity of MGV assumptions using clinical data.
• To identify potential violations of MGV assumptions in published research.

Main Methods:

• Described an ideal research design to minimize error in MGV studies.
• Tested MGV assumptions against clinical datasets.
• Evaluated published MGV studies for evidence of assumption violations.
• Provided practical guidance on MGV assumptions and an example of optimal use.

Main Results:

• MGV assumptions were tested for their applicability with clinical data.
• An assessment was made regarding the prevalence of assumption violations in existing MGV studies.
• An optimal use case for MGV was illustrated.

Conclusions:

• Mixed Group Validation (MGV) is not universally superior to Known Groups Validation (KGV).
• MGV can be a valuable tool when its assumptions and standard error are appropriately managed.
• Careful consideration of study design and assumption adherence is crucial for effective MGV implementation.