Virtual data enclaves for sharing restricted data pre

Virtual data enclaves for sharing restricted data pre and post publication As Open as Possible, As Closed as Necessary: Empowering Transparency in Publications Based on Sensitive Research Data Margaret Levenstein November 20, 2020 1

Resources for Journals 2

Open. ICPSR: Data for reproducibility Any researcher can deposit data and receive DOI immediately Ø Available to share with journal prior to publication Ø Data are not curated, but they are checked for confidentiality Ø Data can be restricted or embargoed Dedicated journal repositories Ø Journal-branded Ø Journal-specific criteria Ø Work flow allows journal to approve deposit prior to data publication 3

I can’t share. My data’s private Why is the data private? Ø Privately owned by third party, governed by NDA and/or RDUA Ø Researcher promised privacy to respondent/participant and/or IRB Ø Data cover sensitive topic Solutions Ø Consent statements and RDUAs that permit sharing for reproducibility and research Ø Share code, process for obtaining access Ø Mask the data Ø Restrict access Ø IRBs will often allow data sharing with restricted access, even when that was not explicitly permitted in research protocol 4

Safe data Curation includes steps to create safe data Ø Consistent with fitness for use Criteria for evaluating data safety Ø Ø Ø Ø Living persons Vulnerable populations Expectations of privacy Data type and level Unit of analysis Sampling Longitudinal Availability Social relationships Geography Date specificity Sensitivity Small or distinct populations Methods for “anonymization” Ø Ø Aggregation Suppression Swapping Perturbation/noise infusion 5

Safe places Secure online analysis Ø Expensive to setup, but easy to access Ø Requires confidence in automatic disclosure checks Encrypted downloads Ø Ø Secure, local computing environments Training Restricted data use agreements Researchers do own disclosure review 6

Safe places Virtual Data Enclave Ø Secure, controlled computing environment Ø Training Ø Restricted data use agreements Ø Third party disclosure review Physical data enclave Ø Secure, controlled physical and computing environment Ø Training Ø Restricted data use agreements Ø Third party disclosure review 7

Sharing data safely Is possible Multiple solutions Increases impact of research Increases credibility of research 8

Data Jeff is happy to take your questions 9