Diagnostic Assay Accuracy and Validation Specification

Role: Lead Diagnostic Biostatistician with expertise in clinical assay validation, CLSI guidelines, and FDA diagnostic submissions.

Context

• Biomarker assay name and modality: {{biomarker_name}}
• Reference standard or comparator test: {{gold_standard_test}}
• Target clinical sensitivity cutoff: {{target_sensitivity_cutoff}}
• Target clinical specificity cutoff: {{target_specificity_cutoff}}
• Target disease prevalence in intended use population: {{prevalence_rate}}
• Cohort enrichment and sampling strategy: {{sample_enrichment_strategy}}

Task

Generate a rigorous diagnostic accuracy statistical validation specification to demonstrate analytical performance and clinical utility against standard acceptance criteria.

Method

1. Define the 2x2 contingency matrix comparing {{biomarker_name}} against {{gold_standard_test}}.
2. Specify point estimation formulas for sensitivity, specificity, PPV, NPV, and likelihood ratios.
3. Incorporate {{prevalence_rate}} into predictive value derivations to reflect intended use.
4. Formulate the exact two-sided 95% confidence interval methodology (e.g., Clopper-Pearson or Wilson Score).
5. Establish non-inferiority or point-estimate acceptance criteria against {{target_sensitivity_cutoff}} and {{target_specificity_cutoff}}.
6. Outline Receiver Operating Characteristic (ROC) curve analysis and Area Under the Curve (AUC) calculation steps.
7. Detail subgroup stratification rules according to {{sample_enrichment_strategy}}.
8. Specify indeterminate result handling and repeat testing protocols.

Constraints

• MUST specify Clopper-Pearson (exact) or Wilson score methods for binomial confidence limits.
• MUST NOT calculate Positive Predictive Value (PPV) without adjusting for {{prevalence_rate}} if using enriched samples.
• The reference standard status must be blinded during assay index testing.
• Power requirements must account for both positive and negative disease sub-cohorts.

Output format

• Validation Scope and Diagnostic Claim (under 120 words)
• Performance Metric Specifications (Table with Metric, Target, Lower Bound CI, Formula)
• Sampling and Cohort Distribution Table (Diseased vs. Non-Diseased breakdown)
• ROC and Cutoff Determination Rules (under 150 words)
• Data Handling and Missing Data Policy (Numbered list of 4-6 items)

Self-review

• Ensure predictive value calculations correctly integrate {{prevalence_rate}}.
• Verify that acceptance criteria match or exceed {{target_sensitivity_cutoff}} and {{target_specificity_cutoff}}.
• Confirm that the confidence interval technique handles boundary conditions near 100% appropriately.