scholarly journals DPCube: Releasing Differentially Private Data Cubes for Health Information

2012 ◽  
Author(s):  
Yonghui Xiao ◽  
James Gardner ◽  
Li Xiong
Keyword(s):  
Health Information ◽  
Data Cubes ◽  
Private Data

scholarly journals Privacy-Preserving Deep Learning for the Detection of Protected Health Information in Real-World Data: Comparative Evaluation (Preprint)

2019 ◽  
Author(s):  
Sven Festag ◽  
Cord Spreckelsen
Keyword(s):  
Deep Learning ◽  
Health Information ◽  
Clinical Data ◽  
Real World ◽  
Privacy Preserving ◽  
Stochastic Gradient Descent ◽  
Data Sets ◽  
Learning Approaches ◽  
Protected Health Information ◽  
Private Data

BACKGROUND Collaborative privacy-preserving training methods allow for the integration of locally stored private data sets into machine learning approaches while ensuring confidentiality and nondisclosure. OBJECTIVE In this work we assess the performance of a state-of-the-art neural network approach for the detection of protected health information in texts trained in a collaborative privacy-preserving way. METHODS The training adopts distributed selective stochastic gradient descent (ie, it works by exchanging local learning results achieved on private data sets). Five networks were trained on separated real-world clinical data sets by using the privacy-protecting protocol. In total, the data sets contain 1304 real longitudinal patient records for 296 patients. RESULTS These networks reached a mean F1 value of 0.955. The gold standard centralized training that is based on the union of all sets and does not take data security into consideration reaches a final value of 0.962. CONCLUSIONS Using real-world clinical data, our study shows that detection of protected health information can be secured by collaborative privacy-preserving training. In general, the approach shows the feasibility of deep learning on distributed and confidential clinical data while ensuring data protection.

scholarly journals 1657. Notes From the Field: A Survey of Mobile Device Usage Among Individuals in KwaZulu-Natal, South Africa

2019 ◽  
Vol 6 (Supplement_2) ◽  
pp. S606-S606
Author(s):  
Breanna R Campbell ◽  
Koeun Choi ◽  
Megan Gray ◽  
Chelsea Canan ◽  
Anthony Moll ◽  
...  
Keyword(s):  
South Africa ◽  
Health Information ◽  
Cell Phone ◽  
Data Access ◽  
People Living With Hiv ◽  
Medical Male Circumcision ◽  
Condomless Sex ◽  
Private Data ◽  
Kwazulu Natal ◽  
To Receive

Abstract Background mHealth (mobile health) is a promising tool to deliver healthcare interventions to underserved populations. Across low- and middle-income countries (LMIC), the prevalence of smartphones has risen to 42%. Effective mHealth deployment in LMIC requires an understanding of how LMIC populations use mobile technology. We characterized the use of mobile devices in rural KwaZulu-Natal, South Africa to tailor mHealth interventions for people living with HIV and at risk for acquiring HIV. Methods We surveyed participants in community settings and offered free HIV counseling and testing. Participants self-reported their gender, age, relationship status, living distance from preferred clinic, receipt of monthly grant, condomless sex frequency, and circumcision status (if male). Outcomes included cell phone and smartphone ownership, private data access, health information seeking, and willingness to receive healthcare messages. We performed multivariable logistic regression to assess the relationship between demographic factors and outcomes. Results Among 788 individuals surveyed, the median age was 28 (IQR 22–40) years, 75% were male, and 86% owned personal cell phones, of which 43% were smartphones. The majority (59%) reported having condomless sex and most (59%) males reported being circumcised. Although only 10% used the phone to seek health information, 93% of cell phone owners were willing to receive healthcare messages. Being young, female, and in a relationship were associated with cell phone ownership. Smartphone owners were more likely to be young and female, less likely to live 10–30 minutes from preferred clinic, and less likely to receive a monthly grant. Those reporting condomless sex or lack of circumcision were significantly less likely to have private data access. Conclusion Most participants were willing to receive healthcare messages via phone, indicating that mHealth interventions may be feasible in rural KwaZulu-Natal. Smartphone-based mHealth interventions specifically geared to prevent or support the care of HIV in young women in KwaZulu-Natal may be feasible. mHealth interventions encouraging condom use and medical male circumcision should consider the use of non-smartphone SMS and be attuned to mobile data limitations. Disclosures All authors: No reported disclosures.

scholarly journals Privacy-Preserving Deep Learning for the Detection of Protected Health Information in Real-World Data: Comparative Evaluation

10.2196/14064 ◽  
2020 ◽  
Vol 4 (5) ◽  
pp. e14064 ◽  
Author(s):  
Sven Festag ◽  
Cord Spreckelsen
Keyword(s):  
Deep Learning ◽  
Health Information ◽  
Clinical Data ◽  
Real World ◽  
Privacy Preserving ◽  
Stochastic Gradient Descent ◽  
Data Sets ◽  
Learning Approaches ◽  
Protected Health Information ◽  
Private Data

Background Collaborative privacy-preserving training methods allow for the integration of locally stored private data sets into machine learning approaches while ensuring confidentiality and nondisclosure. Objective In this work we assess the performance of a state-of-the-art neural network approach for the detection of protected health information in texts trained in a collaborative privacy-preserving way. Methods The training adopts distributed selective stochastic gradient descent (ie, it works by exchanging local learning results achieved on private data sets). Five networks were trained on separated real-world clinical data sets by using the privacy-protecting protocol. In total, the data sets contain 1304 real longitudinal patient records for 296 patients. Results These networks reached a mean F1 value of 0.955. The gold standard centralized training that is based on the union of all sets and does not take data security into consideration reaches a final value of 0.962. Conclusions Using real-world clinical data, our study shows that detection of protected health information can be secured by collaborative privacy-preserving training. In general, the approach shows the feasibility of deep learning on distributed and confidential clinical data while ensuring data protection.

scholarly journals Public Goods, Private Data: HIV and the History, Ethics, and Uses of Identifiable Public Health Information

2007 ◽  
Vol 122 (1_suppl) ◽  
pp. 7-15 ◽  
Author(s):  
Amy L. Fairchild ◽  
Lance Gable ◽  
Lawrence O. Gostin ◽  
Ronald Bayer ◽  
Patricia Sweeney ◽  
...  
Keyword(s):  
Public Health ◽  
Public Goods ◽  
Health Information ◽  
Private Data ◽  
Public Health Information
Sign in / Sign up

Export Citation Format

Share Document