@PWGTennant @otavio_ranzani @melb4886 @uk_biobank @ObserverUK @shanti_das @venkmurthy It also doesn't mention a whisper abt how actigraphy can easily facilitate de-ID. "Feasibility of Re-ID of Individuals in Large National Physical Activity Data Sets From
@venkmurthy @ObserverUK @shanti_das @NHSEngland @PalantirTech @PWGTennant @mikeysmith @DailyMirror @TheDA_UK @SteveBarclay Additional concerns expressed by Emma @emmakstapp and Madhav @MadhavThambiset in the use/misuse of personal pt data https://t.co/XK0j
RT @MadhavThambiset: @emmakstapp @uk_biobank @ObserverUK @PWGTennant @pash22 @BBCNews Indeed. I was waiting for someone to point this out a…
@Prof_Marciniak @HelenRSalisbury
You never walk alone, Rob. @ProfRobHoward
@emmakstapp @uk_biobank @ObserverUK @PWGTennant @pash22 @BBCNews Indeed. I was waiting for someone to point this out and not be the only bearer of bad news myself. Here is a paper describing 95% accuracy in identifying individual adults from their “deident
"The random forest algorithm successfully #reidentified the demographic and 20-minute aggregated PA [physical activity] data of 4478 adults..." https://t.co/T0y5iGdCnb #PrivacyMatters #DeidentifiedData #AggregatedData #HealthDataPrivacy #SensitiveInforma
RT @HKalodimos: Current de-identification techniques are unreliable and tech startups are not known for their thoughtfulness or ethics. htt…
RT @HKalodimos: Current de-identification techniques are unreliable and tech startups are not known for their thoughtfulness or ethics. htt…
Current de-identification techniques are unreliable and tech startups are not known for their thoughtfulness or ethics. https://t.co/Abtkq40FnG https://t.co/speUwTead8
https://t.co/fWO68sQXJO @KoYiKkoDaNs @chennithala @MullappallyR @Oommen_Chandy
Read these stories back to back last night 1: @MayoClinic researchers rebuild faces from MRIs & show they can be used to reidentify patients 2: Clearview AI takes in pictures of people & finds them in a database of 3bn images scrapped from the we
Today the @MayoClinic announced that their platform will leverage federated learning to let partners learn from patient data w/o having to it. Similar initiative announced by @KingsCollegeLon few weeks back. Some day all research could be done this way.
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @Healthunchaind: It's extremely important that people understand that the current layers of protection (de-identification) of data in #h…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
RT @bertcmiller: Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individ…
It's extremely important that people understand that the current layers of protection (de-identification) of data in #healthcare is not enough to actually keep it private. #federated #learning is a solution
Deidentified data is used to protect ppl from privacy violations, but this isn't a real fix bcuz AI can reidentify individuals: https://t.co/PHbHyVsQIm Future studies could use federated learning, enabling real deidentification by removing the need to sha
@rvaughnmd Going to make me think about Strava and Garmin Connect. https://t.co/1QYp3LwTVH
12/ In January 2019, a study was released in #JAMA discussing reidentification via health trackers. Here is an overview- https://t.co/O2uxv6Xvl6; here is the study https://t.co/jbaJUtcUMZ
@bobjherman @zackcooperYale And, of course, "de-identified" data usually isn't really. https://t.co/DoEJuBeOle
Anyway, it’s been clear, and even more so, that the promise of basic pseudonymity is not enough to protect people’s personal info Reminds me of this: AI being used to de-anonymise physical activity data that have had protected health information removed
nising a system of computers housing all the data, etc. I see how AI could work, but AI can also be used to de-anonymise information (https://t.co/t5Kgj3CEQu), so the promise of basic pseudonymity doesn’t seem enough to combat the mishandling of data, i.e
RT @ngranati: @tariqkrim @EinsteinsAttic @carolinejmolloy Below the research paper on re-identifying 'de-anonymised' individual data and im…
RT @ngranati: @tariqkrim @EinsteinsAttic @carolinejmolloy Below the research paper on re-identifying 'de-anonymised' individual data and im…
@tariqkrim @EinsteinsAttic @carolinejmolloy Below the research paper on re-identifying 'de-anonymised' individual data and implication for health data regulation in @_lucyingha's @VerdictUK article https://t.co/RnU3AhizQ3 https://t.co/iYlgTnHb1n
RT @calamitysmith: semi-related article in @JAMANetworkOpen a while back, which found that de-identified physical activity datasets from accelerometers could be re-identified by machine learning if they also had demographic data #wellnessprograms #corpo…
semi-related article in @JAMANetworkOpen a while back, which found that de-identified physical activity datasets from accelerometers could be re-identified by machine learning if they also had demographic data #wellnessprograms #corporatewellness https://
UC Berkeley study underscores the importance of privacy-preserving tech for health data: "Linear support vector machine and random forests reidentified the 20-minute-level physical activity data of approximately 80% of children and 95% of adults" https:/
@hypervisible @Arrianna_Planey Given this research article, seems a recipe for a health disparity privacy nightmare: https://t.co/3wLmjeUGo8
RT @DrKrisSchneider: After @BehavioralMed's webinar on a Digital Health Decision-Making Checklist w/@cnebeker & @DrBartlettEllis, reading a…
RT @DrKrisSchneider: After @BehavioralMed's webinar on a Digital Health Decision-Making Checklist w/@cnebeker & @DrBartlettEllis, reading a…
After @BehavioralMed's webinar on a Digital Health Decision-Making Checklist w/@cnebeker & @DrBartlettEllis, reading about how it is possible to re-identify participants from "de-identified" accelerometer data highlights the importance of learning more
Is it possible to reidentify physical activity data that have had protected health information removed by using machine learning? https://t.co/kUCFFNbrUE
Feasibility of Reidentifying Individuals by Their Protected Health Information via #AI https://t.co/Y0e6h9Iqwc #HIPAA #infosec
RT @calimagna: I got asked a lot about the implications of a recent JAMA Network Open paper "Feasibility of Reidentifying Individuals by Th…
Worried about if that wearable data of yours is truly "de-identified" in a public dataset? @calimagna from @evidation walks through the scenarios. This problem isn't going away any time soon.
I got asked a lot about the implications of a recent JAMA Network Open paper "Feasibility of Reidentifying Individuals by Their Protected Health Information" https://t.co/Y8EfEmgfWW so I wrote a post about it ---> https://t.co/J2uEQtvi1H <--- feedb
Concerning: Supposedly de-identified (physical activity) #health #data easily reidentifiable https://t.co/GUKf6Urf2T
@PANAXEA Advances in artificial intelligence have created new threats to the privacy of health data. A study suggests current laws and regulations are nowhere near sufficient to keep an individual's health status private in the fac…https://t.co/ne6FjLQbxO
@PANAXEA Advances in artificial intelligence have created new threats to the privacy of health data. A study suggests current laws and regulations are nowhere near sufficient to keep an individual's health status private in the face of AI development. http
RT @experiencia_T: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Infor…
Thanks for the link! I can't wait to read this. #hcldr
#hcldr T2a In 2019, I think #privacy will become a hotter topic in #healthcare ⤵️ As more research on health - & fitness - wearables shoes that "deidentified physical activity (PA) data collected from wearable devices can be reidentified" https://t.co/
RT @experiencia_T: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Infor…
Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning https://t.co/RIMoPvlBix
Study on fitness & health wearables & #privacy ↔️ "Deidentified physical activity (PA) data collected from wearable devices can be reidentified" https://t.co/eXnZeKE6zZ @JAMA_current #bioethics #medicine https://t.co/OFCIiOgD2Y
interesting paper on the reidentification of anonymized fitness data https://t.co/pnDY63BBjS
Feasibility of Reidentifying Individuals by Their Protected Health Information https://t.co/Z2wDUvuXYp @JAMA_current @JAMANetworkOpen @JAMANetwork
Careful with anonymized data. Machine learning was used to re-identify health data that has been anonymized. Linear support vector machine and random forests re-identified the 20-minute-level physical activity data of approximatel…https://t.co/NdXfO9404q
RT @B_Lejeune: Artificial intelligence advances threaten privacy of health data https://t.co/ZBKz9IpQOn Study finds current laws and regul…
Artificial intelligence advances threaten privacy of health data https://t.co/ZBKz9IpQOn Study finds current laws and regulations do not safeguard individuals’ confidential health information Advances in artificial intelligence have created new threat… h
Feasibility of Reidentifying Individuals by Their Protected Health Information https://t.co/P1UB5AwFcM
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
“Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning” spoiler: it’s feasible
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
RT @jpwarren: The cynic in me thinks that more evidence won't change the mind of people who have ignored the evidence so far. Hopefully my…
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
Researchers worked on a project to re-identify users in a national de-identified physical activity dataset using #MachineLearning. This algorithm accurately matched physical activity data and demographic information up to 95% of adults in a dataset https:
RT @Asher_Wolf: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Informat…
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
RT @Asher_Wolf: Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus …
RT @Asher_Wolf: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Informat…
Another nail hammering shut the coffin of current deidentification practices in Australia. Ping @jpwarren @VTeagueAus @bipr @chrisculnane @DRWaus @efa_oz @FutureWiseAU @trentyarwood https://t.co/uMvi9AFsqe
Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning. | Electronic Health Records | JAMA Network Open https://t.co/uMvi9AFsqe
RT @aregenberg: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Informat…
RT @aregenberg: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Informat…
Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning. | Electronic Health Records | JAMA Network Open | JAMA Network https://t.co/HhVEC3
RT @chrishendel: Advances in artificial intelligence have created new threats to the #privacy of people's health data, reports a new study…
RT @chrishendel: Advances in artificial intelligence have created new threats to the #privacy of people's health data, reports a new study…
Advances in artificial intelligence have created new threats to the #privacy of people's health data, reports a new study published recently in @JAMANetworkOpen https://t.co/JJHp1slhtC #AI #ArtificialIntelligence #smartphone #smartphones #HIPAA
@JAMA_current discusses activity trackers, and how proper deidentification should work to ensure privacy: https://t.co/9YIPaRckhE
RT @JAMANetworkOpen: Better privacy safeguards are needed for physical activity data because machine learning algorithms can correctly re-i…
Feasibility of Reidentifying Individuals by Their Protected Health Information https://t.co/RStInwxQE7
RT @DrvanTilburg: Think twice about uploading your Fitbit, applewatch or other fitnesstracker information. Once it is on your phone you do…
RT @DrvanTilburg: Think twice about uploading your Fitbit, applewatch or other fitnesstracker information. Once it is on your phone you do…
RT @cpontesg: Machine learning allows reidentifying physical activity records and demographics for 80% of individuals in a large deidentifi…
RT @DrDanaB_S: #Deidentification no longer works. #DataPrivacy https://t.co/brRVrquKqq
Machine learning allows reidentifying physical activity records and demographics for 80% of individuals in a large deidentified dataset from NHANES; better techniques needed for data protection https://t.co/xjERI8jGIm
RT @DrvanTilburg: Think twice about uploading your Fitbit, applewatch or other fitnesstracker information. Once it is on your phone you do…
RT @donaldhtaylorjr: Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Inf…
Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning. | Electronic Health Records | JAMA Network Open | JAMA Network https://t.co/hmrYFP
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @SteveJoffe: Reidentification of participants in "deidentified" data sets: not just a challenge for genomics any more. https://t.co/msQg…
Hey @MatthewFitz_1 This one too apparently 🤦♂️
RT @DrDanaB_S: #Deidentification no longer works. #DataPrivacy https://t.co/brRVrquKqq
RT @DrDanaB_S: #Deidentification no longer works. #DataPrivacy https://t.co/brRVrquKqq
#Deidentification no longer works. #DataPrivacy
RT @DrvanTilburg: Think twice about uploading your Fitbit, applewatch or other fitnesstracker information. Once it is on your phone you do…