This session provides guidance about methods and approaches to achieve de-identification in accordance with the HIPAA Privacy Rule, De-identification is a process that organizations can use to remove personal information from data that they collect, use, archive, and share with other organizations.
Today health information needs to be shared more than ever, but how can that be done most easily within the limits of HIPAA? One way is to de-identify the information. Once PHI has been de-identified, it is no longer protected under HIPAA and may be shared freely without limitation. The problem is that it is not easy to truly de-identify information and if it is not done correctly, the sharing of the information may be considered a breach that requires reporting to HHS and the potential for penalties and corrective action plans.
De-identification of Protected Health Information requires removing all eighteen of the listed identifiers, or anything else that might be used to identify the individual about whom the information exists. Or you can have an expert certify that the information is not identifiable. But neither of these is foolproof. You need to look more closely to be sure the data cannot be identified.
You may wish to communicate with another provider, or with an agency that is not covered under HIPAA, using plain e-mail, but you want to strip out the name and use a code that both parties understand. Is that sufficient to allow the use of plain e-mail? You need to run though some examples and some tests to make sure before you go ahead.
The necessity to consider the context of information is essential, especially when the information is unique. A staff member may think a photo of an injury has no identification on it and by itself is not PHI, but if the photo is posted on the staff member's Facebook page shortly after the incident and it's a small town and everyone knows whose injury it is, it's been identified by the context. Sometimes you may need information for research that does not require specific identification of the individual, but does need some information listed in the eighteen identifiers, such as Zip code, dates of birth or death, or dates of treatment. In those cases, often partially de-identified data, known as a Limited Data Set, will suffice, and such data can be used without obtaining an Authorization or approval by a review board. The information must still be protected with HIPAA-quality security, but it can be used for research under a Data Use Agreement.
There are specific steps that you must go through to ensure that if you want to de-identify PHI, you actually do so properly, and that the resulting information is truly de-identified and its use or disclosure will not result in a reportable breach under HIPAA. If you create a Limited Data Set, you need to ensure the proper agreements are in place and the information is transmitted securely. If de-identification or a Limited Data Set are not possible, the appropriate Authorizations or approvals must be in place before sharing the data. This session will explore the concepts and methods of de-identification and many of the typical questions that arise. Attendees will be able to go forward with de-identification with greater confidence, and better sharing of information will be possible.
Why should you Attend: Health information is proliferating and its sharing among health care providers and researchers is necessary for providing health care services and advancing essential health care research. But health information protected under the Health Insurance Portability and Accountability Act of 1996 (HIPAA) is limited in how it can be shared, and when it is shared, is required to be protected.
In some cases, it may be possible to share the needed information more easily once it has been properly de-identified. While a risk analysis would indicate the necessity to encrypt e-mail with Protected Health Information when communicating between providers over the Internet, for instance, if the information is not identifiable, encryption is no longer needed. While releasing information for research purposes may call for a HIPAA Authorization from each patient or approval by review boards and stringent controls on the information, if the research can be done without the identifying data, such Authorizations are reviews are not necessary. But truly de-identifying information is never as simple as it looks. Oftentimes the context of the information or the uniqueness of information can give away the identity. If you use patient initials in an e-mail to identify the individual to the recipient, how unique are those initials? What if the patient's name is Xavier Xanadu -how many individuals have the initials X. X.? If you see information about eight siblings with the same birth date and location, you know it's about the famous Octuplets, because that's the only set of eight you will find.
If information is not properly de-identified and released inappropriately as a result, it can result in fines and corrective action plans that can reach into the millions of dollars. The right process needs to be followed to ensure that data that is shared is shared appropriately, either as identifiable information, as a partially de-identified Limited Data Set, or as properly de-identified information. This session will review guidance from the HHS Office for Civil Rights (OCR) and from the National Institute of Standards and Technology (NIST) about how to properly de-identify health information. The various needs for de-identified information will be discussed and typical questions covered in the guidance will be discussed, in order to provide a sound, defensible basis for an organization's decisions and processes surrounding de-identification of PHI.
Areas Covered in the Session: