Subject De-Identification

The ImmPort team follows the Safe Harbor Method Principles in data capture templates and in data submission by:

  • Rounding any “ages” submitted > 89 to 90 and add a comment indicating that this was done
  • Not recording any dates within ImmPort templates; all time points and schedules are referenced to study day 0 (or a designated day in the study as determined by the data provider)
  • Not recording any geographic location subdivisions information below the state level in ImmPort templates
  • Not capturing any subject identifying information in ImmPort templates itemized in the Safe Harbor Method (Name, accounts, phone or fax numbers, email addresses, etc)

The ImmPort team removes the connection to subject identifiers in the source data management system by:

  • If possible, discuss with the data providers to determine that local subject identifiers that were used in the submission are deidentified
  • Not displaying subject source identifiers, sample source identifiers, or experiment sample source identifiers in the ImmPort application to cut linkage with the source system
  • Not exporting the subject source identifiers, sample source identifiers, or experiment sample source identifiers in any ImmPort generated data packages delivered online, at sFTP sites, or in the future, via Amazon AWS S3 storage
  • Not providing original, unstructured result files submitted by sites if they (1) haven’t stated concurrence with the disconnect with source identifiers; and/or (2) if the data can be parsed into standard ImmPort templates and source identifiers hidden

The ImmPort team does not normally accept or share GWAS level SNP human genotyping data nor Next Gen human genome sequencing data, and recommends depositing such data in dbGAP or SRA

For data provided in more “native” format (i.e., not in ImmPort template format), such as result files from closed clinical trials (assessments, lab tests, adverse events, etc):

  • If the data source is ITN TrialShare, the dates are already obfuscated and are not identifying as are the subject identifiers; no special treatment is required
  • For other data sources with closed clinical trial data (or the equivalent), data files are reviewed by the ImmPort Curation team who will 1) Remove identifying information such as names, dates, locations, etc...2) Normalize dates to study day 0; 3) Comment and Description fields are removed unless determined to be completely devoid of identifying information