SynD Framework logoSynthetic Health Data Governance Framework (SHDGF)
Home
ResourcesAbout SynD
1
Use Case
2
Source Data
3
Generate
4
Re-ID Risk
5
Safety
6
Final

Step 3 - Generate Synthetic Health Data

Apply approved synthesis techniques to generate synthetic data while maintaining data utility and privacy. Document processes for transparency.

DPDSDR

Governance Intent

Balance data utility with privacy during generation.

Decisions

Does the synthetic dataset achieve adequate utility?

Has privacy risk increased due to overfitting?

Why This Step

Generate synthetic data using validated methods while maintaining the balance between data utility and privacy protection.

Prerequisites

  • Completed Steps 1 and 2
  • Approved synthesis method
  • Quality source data ready

Time Estimate

Variable (depends on data size and method)

Balance utility and privacy
Monitor for overfitting during generation. If privacy risk increases, apply additional controls before proceeding.

Synthesis Checklist

Complete all steps to ensure proper synthetic data generation.

Required Evidence

  • Model parameters + logs
  • Utility metrics
  • DP oversight record

Step completion requirements

0 / 5 complete

Finish these before marking the step complete:

  • Select synthesis method appropriate to data type and intended use.
  • Validate algorithms for statistical fidelity and privacy protection.
  • Record synthesis parameters and version of generator used.
  • Conduct internal testing for quality, representativeness and bias.
  • Store synthesis scripts securely and maintain provenance logs.

Resources

  • Decision Tree for Complex Cases (Appendix 8)
  • Lawful Pathways Guide (Appendix 9)
  • Evaluating Fidelity and Utility (Appendix 12)
Previous StepOverviewNext Step