dicentra published a new JALM article (Aug 2025) presenting a staged framework that integrates analytical validity, clinical validity, clinical utility, regulatory strategy, and integrated evidence generation.

Regulatory Approved Point-of-Care Diagnostics (FDA & Health Canada): A Comprehensive Framework for Analytical Validity, Clinical Validity, and Clinical Utility in Medical Devices | The Journal of Applied Laboratory Medicine | Oxford Academic

This holistic roadmap guides developers from lab validation through clinical trials to regulatory submission and real-world implementation. It explicitly links test performance metrics to regulatory and clinical goals so study design and submission planning are more deliberate and efficient.

Analytical validity: Can the test measure reliably?

Analytical validity is about measurement: does the assay produce accurate and precise results under controlled conditions? Typical analytical questions include:

• Bias (closeness to reference)
• Imprecision (coefficient of variation)
• Limit of detection (LOD)
• Linearity
• Interference
• Lot-to-lot consistency

Bench studies and contrived samples answer these questions. Relevant methods include:

• Method comparison (Passing–Bablok, Deming regression)
• Bland–Altman plots → visualize bias and limits of agreement
• CLSI-style LOD studies → dilution series

Note: when you see the term “sensitivity” in this context, it usually refers to analytical sensitivity (LOD), not clinical sensitivity in patients.

Clinical validity: Does the measured value mean anything in patients?

Clinical validity establishes whether the test result correctly classifies disease or clinical state in the intended use population. This is where clinical sensitivity/specificity, positive and negative percent agreement (PPA/NPA), predictive values, and ROC/AUC analyses are determined in patient samples.

Study designs can include:

• Retrospective archival specimen comparisons
• Prospective multicenter studies in intended settings (ED, primary care, home use)

Common statistical tools for clinical validation:

• ROC/AUC with DeLong’s test → discriminatory performance
• Paired t-tests → continuous method comparisons
• McNemar’s test → paired categorical outcomes
• Cohen’s κ → agreement beyond chance
• Logistic regression → adjusting for clinical covariates

Importantly, where no single “gold standard” exists, developers should report PPA/NPA and explain limitations.

Clinical utility: Does using the test improve care?

Clinical utility asks whether deploying the test changes clinical decisions and leads to better patient outcomes or system efficiency.

Utility is demonstrated by outcome-focused, pragmatic studies measuring endpoints such as:

• Time-to-treatment
• Length of stay
• Readmission rates
• Patient-centered metrics

Economic analyses (e.g., cost per QALY, budget-impact models) quantify value for health systems and payers.

Example: in emergency department evaluations, rapid molecular respiratory testing has shortened time to targeted therapy and reduced unnecessary admissions. Such trials can show both clinical and economic benefit.

From a methods standpoint, utility evidence often relies on:

• Randomized or pragmatic cohort designs
• Time-to-event analyses (Kaplan–Meier, Cox models)
• Decision-analytic modeling (sensitivity analysis, probabilistic simulation)

Integrated evidence development: Plan studies that serve multiple goals

A practical advantage of the staged framework is intentional overlap: design studies that answer analytical, clinical, and utility questions where possible.

• Bench work addresses analytical limits first.
• Clinical trials can (and should) collect secondary endpoints for utility (e.g., time to therapy, short-term clinical outcomes) and health-economic data.

This “parallel evidence” approach increases efficiency. For example, a single multicenter prospective accuracy study can include:

• Short-term outcome measures
• Usability assessments
• Health-economic endpoints

Together, these generate material that supports regulatory submissions and payer conversations simultaneously.

For every study, explicitly state:

• The primary question (analytical vs. clinical vs. utility)
• The specimen source (contrived vs. clinical)
• The statistical plan

This helps reviewers and regulators interpret results in the correct context.

Limitations and practical considerations

Integrated development is powerful but resource-intensive. It requires:

• Early multidisciplinary alignment (clinical operations, biostatistics, regulatory affairs, human-factors, health-economics)
• Careful protocol design to avoid bias (spectrum and verification bias)
• Pre-planned subgroup analyses

Manufacturers must balance breadth of evidence with feasibility. Staged milestones, adaptive protocols, and early meetings with regulators reduce the risk of late-stage gaps.

dicentra’s experience

dicentra brings hands-on CRO experience across the POC lifecycle:

• Analytical method development and CLSI-aligned bench validation
• Multicenter clinical programs
• Human-factors/usability testing
• Payer evidence generation

Our teams design efficient, multi-purpose studies that meet regulator expectations (FDA and Health Canada) and build health-economic models that speak to payers.

We specialize in aligning statistical rigor with practical operations — reducing development time and strengthening submissions so promising POC technologies reach patients faster and with clearer value propositions.

Bottom line

A staged, integrated evidence strategy de-risks POC development.

By:

• Distinguishing analytical from clinical measures
• Embedding utility endpoints into clinical studies
• Coordinating regulatory and statistical plans early

…developers can build robust dossiers that demonstrate accuracy, patient impact, and health-system value.

dicentra’s pragmatic, cross-functional approach helps teams implement this roadmap efficiently — from prototype to cleared, impactful diagnostics.