Skip to content

Aetion has entered into an agreement to be acquired by Datavant, a leading health data platform company.   Learn more → 

Natalie Schibell, MPHMay 14, 20253 min read

New Release: Aetion Generate 3.7 Enhances Privacy-First Real-World Data Transformation

At Aetion®, we believe privacy should never hinder evidence generation. With version 3.7 of Aetion® Generate, we’ve expanded the platform’s ability to handle complex, multi-table datasets – enabling researchers, data scientists, and compliance leads to apply advanced transformations and risk assessments across broader, more sophisticated data environments.

Purpose-built for those working with longitudinal data, linked health records, or multi-source repositories, Generate 3.7 ensures that data transformed for de-identification remains scientifically valid and ready for analysis. With this release, organizations can confidently scale real-world data (RWD) preparation for high-stakes use cases—whether regulatory, payer-facing, or publication-driven.

What’s New in Generate 3.7

Smarter Transformations. Safer Data. Seamless Research.

Aetion® Generate 3.7 includes:

Unified Identifier Transformation Across Linked Datasets

Generate 3.7 enhances multi-table workflows by enabling privacy-preserving transformations across variables linked between multiple tables. Privacy officers, compliance analysts, privacy engineers, data scientists, and clinical researchers can consistently apply transformations—such as hashing, suppression, and grouping—across all related tables. This ensures referential integrity, strengthens privacy protections, and streamlines multi-table preparation for coherent, analysis-ready outputs.

Expanded Multi-Table Workflows

With identifier consistency across tables, Generate 3.7 expands the range of secure research applications:

  • Multi-Table Re-Identification Risk Assessment: Apply risk-mitigating transformations across linked datasets to proactively reduce re-identification risk while preserving analytic utility.
  • Partial Longitudinal Synthesis (PLS) Across Multi-Table Data: Generate synthetic versions of multi-table datasets while maintaining consistent identifier treatment, enabling secure collaboration and innovation.
  • Uniform Identifier Encryption Across Projects: Encrypt patient or subject IDs consistently across all associated tables to support defensible risk mitigation and maintain structural relationships for scalable research.

 

Expanded Transformation Techniques for Precision Privacy

Generate 3.7 adds new transformation methods that help better tailor de-identification strategies. These enhancements equip privacy officers, clinical researchers, and data engineers to meet stringent privacy requirements while preserving analytical value.

Table 1: Overview of New Data Transformation Methods in Generate 3.7

Transformation Method

Description

Use Cases

Examples

Conditional Suppression

Suppresses values based on logical conditions tied to other variables.

Protecting data in stratified cohorts or conditional privacy settings.

Suppressing rare diagnosis codes (e.g., shark attack) in medical claims data.

Conditional Hashing

Applies hashing only when specific criteria are met.

Targeted de-identification of sensitive subgroups.

Hashing patient identifiers only for patients with sensitive conditions—e.g., individuals diagnosed with HIV or receiving mental health services.

Quantile Transformation

Converts continuous variables into quantile-based segments.

Segmentation for outcomes research and stratified analysis.

Converting raw income into income quartiles to protect privacy.

Custom Binning

Buckets values into defined ranges with custom labels.

Tailored privacy protection while retaining analytical meaning.

Grouping ages into custom ranges (e.g., 18–25, 26–35) for analysis.

Grouping for Categorical Variables

Merges and relabels categories to reduce identifiability.

Improving dataset generalizability and reducing disclosure risk.

Combining rare racial/ethnic groups into an “Other” category.

Truncation

Removes parts of strings or numbers based on position or pattern.

Cleaning semi-structured fields like zip codes, names, or IDs.

Truncating ZIP codes to 3 digits (e.g., 123XX) for geographic de-identification.

 

Optimized for High-Volume, Privacy-Safe Data Transformation

Behind the scenes, Generate 3.7 includes performance enhancements that reduce processing time, support larger workloads, and maintain the highest levels of data protection. These optimizations ensure that workflows remain efficient and secure, even as datasets grow in size, complexity, and regulatory importance.

Defining the Next Standard in Privacy-Safe, Research-Ready Evidence Generation

With Generate 3.7, Aetion expands its leadership in privacy-preserving real-world data transformation through de-identification and synthesis. Generate empowers organizations to operate at scale without sacrificing compliance, scientific validity, or operational efficiency. Privacy protections are integrated across all tables and transformation types, enabling teams to turn raw data into compliant, research-ready datasets with speed, confidence, and control.

Ready to Transform Your Workflow?

Whether you're preparing data for regulators, payers, or internal research, Aetion® Generate 3.7 ensures your data is ready—scientifically rigorous, privacy-aligned, and built for impact.

Connect with us to explore how Generate 3.7 can help unlock new value from your real-world data.

RELATED ARTICLES