Health 2.0 Presentation How BigPrivacy Enables Lawful Analytics

Presentation Transcript
Read the press article about Anonos' presentation at the Health 2.0 event: News Article
Health 2.0 Presentation How BigPrivacy Enables Lawful Analytics
Slide 1
Regulation as Competetive Advantage
Slide 2
My name is Gary LaFever and I will demonstrate how BigPrivacy technology enables lawful Big Data analytics.
BigPrivacy Demonstration
Slide 3
Static approaches to privacy fail to protect Big Data. However, complying with new dynamic data protection requirements actually enables even greater Big Data use.
Gartner: Variant Twins are the SOLUTION
Slide 4
When awarding Anonos BigPrivacy cool vendor status for Privacy Management, Gartner noted that common static privacy techniques fail to protect Big Data.
Unsustainable Conflict
Slide 5
Neither HIPAA de-identification techniques protect privacy for dynamic data use. FTC rules that apply to data collected from fitness devices similarly fail. On the other extreme, Common Rule and related IRB restrictions can be so rigid that BigData breakthroughs are impossible.
BigPrivacy Variant Twins Reconcile Conflict
Slide 6
Anonos BigPrivacy dynamically produces non-identifying Variant Twins of protected data to reconcile this conflict.
Data Protection for Static vs Dynamic Data Uses
Slide 7
Earlier laws protected static data use. If data was used for a single purpose, it was protected. But combining or sharing data broke down the protection. New laws - including here in California - now require protection for dynamic data use.
Variant Twins Maximize Lawful Data Value
Slide 8
BigPrivacy creates Variant Twins that maximize the lawful value of data in 3 steps: 1) De-Linking, 2) De-Risking and 3) Controlling the Linkability of data. Variant Twins reveal information that can be used for lawful Big Data analytics.
Protected Health Information
Slide 9
We start with protected health information which is static and exposes the individual data subject to potential privacy harms.
Step #1: De-Linking (Automated)
Slide 10
In the first step, data is automatically de-identified by de-linking it from identifying information. Each occurrence of a person's name is replaced with a different dynamically generated pseudonym that can be re-linked back to an individual's name but only under tightly controlled conditions.
Step #2: De-Risking (Automated)
Slide 11
The second step is de-risking. Here common data values are grouped into a common cohort class which is non-identifying and represented by a dynamically generated pseudonym.
Step #3: Controlled Re-Linking (Automated)
Slide 12
In the third step, BigPrivacy dynamically classifies values within cohort classes for later controlled re-linking. In this example, age, gender, and race are represented by cohorts that do not reveal underlying values to protect against discrimination.
Result = Variant Twin Data
Slide 13
This enables the production of Variant Twins that support lawful Big Data analytics. Here, you see how many people are in each of the different groups without revealing identities.
Controlled Re-Linking for Re-Identification
Slide 14
Most importantly, only if a user has the authority and a lawful reason to do so, can they re-link data back to identities.
Watch: TED Talk on BigPrivacy
Slide 15
We invite you to watch a Ted Talk six minutes in length at describing how BigPrivacy can enable maximum Big Data value. On the right-hand side is an article highlighting a heart-wrenching situation where the limitations of static privacy technologies make it impossible to identify children with similar symptoms as this young girl to help in her treat ent. But, with BigPrivacy technology, her doctors would not have to settle for this result
BigPrivacy Demonstration
Slide 16
Prior static approaches to data protection are no longer lawful. But complying with new legal requirements enables even greater Big Data use. This is Big. This is BigPrivacy.