Leveraging Existing Clinical Data for Pediatric Medical Device Indications

What You Need to Know? 👇

What is pediatric extrapolation for medical devices and when can it be used?

Pediatric extrapolation is the process of leveraging existing clinical data from adults or other populations to support medical device approvals for pediatric patients. It can be used when the course of disease and device effects are sufficiently similar between populations, helping address the scarcity of pediatric-specific clinical evidence.

How does FDA determine if extrapolation is appropriate for a pediatric device indication?

FDA uses a decision tree framework that evaluates three main factors: similarity between adult and pediatric populations, quality of existing data, and whether the data constitutes valid scientific evidence. The process considers disease characteristics, device performance, patient factors, and endpoint relevance separately for safety and effectiveness.

What’s the difference between full and partial extrapolation in pediatric device development?

Full extrapolation uses existing adult data as a complete substitute for pediatric clinical data, requiring no prospective pediatric studies for that endpoint. Partial extrapolation combines adult data with pediatric data through statistical modeling to increase precision and account for population differences.

Can adult safety data be extrapolated to pediatric populations for medical devices?

Yes, but it’s expected to be rare. Safety extrapolation requires careful consideration of physiological differences between adults and children. The decision is made independently from effectiveness extrapolation, and additional pediatric safety studies are often needed due to unique pediatric factors like growth and development.

What statistical methods are recommended for leveraging adult data in pediatric device studies?

Bayesian hierarchical modeling is commonly used to “borrow strength” from adult studies while accounting for population differences. Other methods include propensity score approaches and power priors. The choice depends on data availability, population similarities, and the ability to model relevant covariates.

Does this guidance apply to all medical device regulatory pathways including 510(k) submissions?

No, this guidance specifically applies to PMA, de novo, and HDE submissions where pediatric indications are sought. It doesn’t apply to 510(k) submissions because the substantial equivalence standard differs from the reasonable assurance of safety and effectiveness standard addressed by the Pediatric Medical Device Safety and Improvement Act.

---

What You Need to Do 👇

Recommended Actions

1. Engage with FDA early in development process when considering extrapolation
2. Follow the Pediatric Extrapolation Decision Tree to determine if extrapolation is appropriate:
   • Assess relevancy of existing data
   • Evaluate similarities/differences between populations
   • Consider data quality
3. Document rationale for extrapolation approach including:
   • Justification for full vs partial extrapolation
   • Analysis of population similarities/differences
   • Statistical methodology
   • Data quality assessment
4. Consider separate evaluations for safety and effectiveness extrapolation
5. Plan appropriate supplemental pediatric studies if needed
6. Develop comprehensive labeling that reflects extrapolation approach and limitations
7. Consider post-market surveillance needs, especially for safety
8. Ensure statistical methodology is appropriate and agreed upon with FDA
9. Maintain focus on providing valid scientific evidence to support reasonable assurance of safety and effectiveness
10. Document all assumptions and limitations of extrapolation approach

Key Considerations

Clinical testing

• Extrapolation decisions for safety and effectiveness are made independently
• Full extrapolation of safety data is expected to occur rarely
• When prospective pediatric data are needed, study design and analysis should be appropriate for the circumstances
• Post-market surveillance may be required, particularly when full extrapolation of safety data is used

Human Factors

• Must consider differences in human factors between adult and pediatric populations that could impact device safety or effectiveness
• Special considerations needed for pediatric-specific challenges like growth during device use period

Labelling

• Must provide appropriate labeling to support safe and effective pediatric use
• Labeling should reflect any limitations of extrapolated data

Safety

• Safety extrapolation requires careful consideration of physiological differences between adults and children
• Full extrapolation of safety data is rare
• Additional safety studies in pediatric populations may be needed even when effectiveness data can be extrapolated

Other considerations

• Must consider similarities/differences in:
  • Disease characteristics between populations
  • Device characteristics and use between populations
  • Location and duration of device implantation/contact
  • Patient characteristics unique to pediatrics
• Statistical methodology must be appropriate for extrapolation approach
• Quality of existing data must be sufficient
• Early engagement with FDA recommended when considering extrapolation

---

Relevant Guidances 🔗

• Evaluation of Sex-Specific Data in Medical Device Clinical Studies
• Benefit-Risk Determinations for Medical Device Premarket Review
• Premarket Assessment of Medical Devices for Pediatric Use
• Statistical Guidance for Reporting Diagnostic Test Results with Qualitative Outcomes

Related references and norms 📂

• 21 CFR 860.7: Valid scientific evidence requirements
• 21 CFR 860.7(c)(1)&(2): Definition of valid scientific evidence

---

Original guidance

• Leveraging Existing Clinical Data for Pediatric Medical Device Indications
• HTML / PDF
• Issue date: 2016-06-21
• Last changed date: 2024-04-08
• Status: FINAL
• Official FDA topics: Medical Devices, Pediatric Product Development, Good Clinical Practice (GCP), Premarket Approval (PMA), Premarket, Biologics, HUD/HDE
• ReguVirta ID: 4271e4766f3c157420c5b6a2bd02efbe