Every effort must be made to protect the privacy and confidentiality of participants regardless of the type of data being used as part of a research project. A key protective measure is to de-identify datasets.
Typically, this involves removing 18 key direct identifiers as described under HIPAA, and indirect identifiers, such as age, gender, location, etc from any part of the data.

The aim of de-identification is to maintain confidentiality through elimination of identifiers in a way that eliminates or greatly diminishes the risk of re-identification of an individual patient. Maintaining strong data security processes is an additional key measure. Sharing data should be done securely, after the information is de-identified and within a data management and sharing plan. This plan should delineate data security, restrictions on data which cannot be sufficiently de-identified, data use agreements, naming conventions and other clearly specified requirements. In most instances, sharing data should be possible without compromising the confidentiality of participants.

Fred Hutch Information Classification Standard

Data should be handled according to the Fred Hutch Information Classification Standard, which describes security and handling standards commensurate with the risk of information mishandling. These standards provide guidelines for managing data at Fred Hutch and for sharing information within Fred Hutch and between Fred Hutch and an outside party. Genomic data, outside of public reference data, is considered confidential (Level II) or strictly confidential information (Level III).

If you are sharing data that includes a third party’s proprietary data or has third-party commercial restrictions, please contact Fred Hutch Business Development & Strategy (BD&S). When research is funded by a commercial sponsor, restrictions on data sharing may apply because of agreements with the sponsor. Any such restrictions should be highlighted in the data management and sharing plan. In the event that you apply for or receive commercial funding for any part of research, you should advise BD&S of the situation without delay.

Data Classification Overview

Public - Level I

This level includes published research results, Fred Hutch publications and communications press announcements, public record documents, job postings, open source configuration list/code/recipes, reference genomes, released data sets, public cryptographic keys.

Confidential - Level II

This level includes pre–publication research information and analyses, medical expense information, invoices, legal instruments or agreements, transaction documents and reports, Fast file and Secure File server, building plans and information about the physical plant, de-identified research participant information, donor information, metadata, Human Resources data/Employee ID numbers, server names/IP addresses, corporate policies, DNS and LDAP information.

Strictly Confidential - Level III

This level includes Protected Health Information (PHI), Individually Identifiable Health Information (IIHI), Personally Identifiable Information (PII), passwords and encryption keys, proprietary information including that belonging to entities other than Fred Hutch, hardware and software authtorization or authentication keys, electronic communication and documents regarding personal or financial matters or other sensitive subjects.

Level Subject to FH Admin. Control Access Requires Authentication Logging/Audit Encryption at Rest (Encryption effective mid-2019) Encryption in Transit Email Paper-Based
Public - Level I No No No   No, On Premises No, Off-Premises (Cloud) No Restrictions No Restriction
Confidential - Level II Yes Yes Yes, if automated or manual system allows. No, On Premises No, Off-Premises (Cloud) No, On Premises Yes, Off-Premises (Cloud) FH Supported Systems Confidential Labeling
Strictly Confidential - Level III Yes, Both on and off Premises Yes Yes. Audited Semi-Annual Yes, On Premises Yes, Off-Premises (Cloud) Yes, On Premises Yes, Off-Premises (Cloud) Encrypted and FH Approved Systems Tamper-proof Envelope/Registered Mail/Signed Delivery

All research involving 1) human subject participants, 2) patient information, or 3) tissue samples derived from patients/human participants must include appropriate safeguards to protect the privacy of research participants. You must ensure the necessary patient consent:

  1. adheres to Human Subjects Protection by receiving IRB approval, and
  2. is signed by the patient prior to data sharing.

Requirements to adhere to relevant regulatory, ethical, or institutional policy should be met, data security measures established and all IRO and patient permissions should be in place prior to disclosing any data. Requirements may dictate sharing through a data use agreement.

Security

Common mechanisms for sharing datasets:

Note: Fred Hutch data should not be stored or shared via applications utilized for personal use.

Office 365 Email and OneDrive

The implementation of Office 365 provides the ability to secure email, storage and sharing. Outlook email can be secured within the Fred Hutch environment and between Fred Hutch and external entities. To trigger email encryption, type “secure” (without quotes) in the subject line of any email and additional instructions for securing email are provided here. For users of mobile devices, Outlook mobile app (iOS and Android) ca display encrypted email within the app without opening browser.

More information about using OneDrive to share data can be found on our Collaborative Data Storage page.

Final Notes

Generally, Investigators can more widely share their data to the scientific community by transferring it to a data archive facility to maintain documentation and meet reporting requirements. Data archives are particularly attractive for investigators concerned about managing a large volume of requests for data, vetting frivolous or inappropriate requests, or providing technical assistance for users seeking to help with analyses. Fred Hutch does not have a centralized data archive and does not have a preferred recommendation for one.

Datasets that cannot be distributed to the general public, due to confidentiality concerns or third-party licensing or use agreements that prohibit redistribution, can be accessed through a data enclave. Fred Hutch does not have a data enclave and does not have a preferred recommendation for one. A data enclave provides a controlled secure environment in which eligible researchers can perform analyses using restricted data resources.