Hi, our topic contributed (TC) session at the JSM needs a volunteer to be session chair.
I am a TC presenter and per ASA rules I am not allowed to also be chair.
This TC session is about data privacy and HIPAA considerations. . Completely separately, planning is underway for at least one webinar for the section about data privacy. I will have more details to share about the webinar at the section business meeting in Toronto.
Session title
"Protected Health Information Privacy, and Statistical Disclosure Considerations for Projects using HIPAA"
link to the session abstract and presenters here:
https://ww2.aievolution.com/JSMAnnual/index.cfm?do=ev.viewEv&ev=2268
link and a cut/paste of the session abstract
https://ww2.aievolution.com/JSMAnnual/index.cfm?do=abs.viewAbs&abs=1691
Session Description:
Statisticians working with data from humans, such as clinical trials, social media, voters, retail customers, etc. need to be aware of the possibility for adversaries to de identify subjects in the database. Privacy in the 21st century may no longer exist. The plethora of both public and private databases and the availability of sophisticated algorithms enabling adversaries to link public and private databases with many covariates simplifies the task to de-identify anonymized data and obtain actual name, address etc.. This session speakers address privacy considerations when using anonymized data. Statisticians working with data from humans always must consider privacy of the patients or volunteers providing the data. The speakers discuss privacy from several perspectives, appropriate to protection of privacy of data bases with data from individuals, such as voters, patients , etc. The consequences of privacy risk breaches are not benign but not necessarily always or easily quantifiable. An example of a publicly documented privacy breaches is the NETFLIX-IMDB de-anonymization. Two computer scientists had devised and published an algorithm using the IMDB database to de-anonymize individual NEFTLIX customers to determine customer name and address. NETFLIX management confirmed that the scientists had de-anonymized and determined the correct information about the customers, name and address and other information. A 2nd example is re-identification of Massachusetts Governor William Weld's medical data with an insurance data set which had been stripped of direct identifiers Barker (2022) recently reported a proof of concept of the unauthorized use of clinicaltrials.gov in combination with SEER to de-anonymize patients oncology patients enrolled in pharmaceutical industry oncology drug clinical trials.
The speakers /paper presenters with affiliations (where I have them) are
1. Jack Fitzsimons - Oblivious AI
2. Jimmy Efird - Boston VA Cooperative Studies Program Coordinating Cente - Abstract title: De-identification and privacy: The role of the statistician as a data guardian
3. Marcia Levenstein
4. Lu Tang - U. Pittsburgh
5. Chris Barker - moi! Consultant Statistical Planning and Analysis Services Inc.
------------------------------
Chris Barker, Ph.D.
2023 Chair Statistical Consulting Section
Consultant and
Adjunct Associate Professor of Biostatistics
www.barkerstats.com---
"In composition you have all the time you want to decide what to say in 15 seconds, in improvisation you have 15 seconds."
-Steve Lacy
------------------------------