Exploring the tradeoff between data privacy and utility with a clinical data analysis use case

Data reusability has become a distinct characteristic of scientific, commercial, and administrative practices nowadays. However, an unlimited and careless reuse of data may lead to privacy breaches and unfair impacts on individuals and vulnerable groups. Data content adaption is a key aspect of preserving data privacy and fairness. Often, such content adaption affects data utility adversely. Further, the interaction between privacy protection and fairness protection can be subject to making trade-offs because mitigating privacy risks may adversely affect detecting unfairness and vice versa. Therefore, there is a need for research on understanding the interactions between data utility, privacy and fairness. To this end, in this contribution, we use concepts from causal reasoning and argue for adopting an integrated view on data content adaption for data driven decision support systems. This asks for considering the operation context wholistically. By means of two cases, we illustrate that, in some situations, local data content adaption may lead to low data quality and utility. An integrated wholistic approach, however, may result in reuse of the original data (i.e., without content adaption, thus in higher data utilization) without adversely affecting privacy and fairness. We discuss some implications of this approach and sketch a few directions for future research.

Vietnam: Data Privacy in a Communist ASEAN State

Personal big data pricing method based on differential privacy

Personal data privacy is a great concern by governments across the world as citizens generate huge amount of data continuously and industries using this for betterment of user centric services. There must be a reasonable balance between data privacy and utility of data. Differential privacy is a promise by data collector to the customer’s personal privacy. Centralised Differential Privacy (CDP) is performing output perturbation of user’s data by applying required privacy budget. This promises the inclusion or exclusion of individual’s data in data set not going to create significant change for a statistical query output and it offers -Differential privacy guarantee. CDP is holding a strong belief on trusted data collector and applying global sensitivity of the data. Local Differential Privacy (LDP) helps user to locally perturb his data and there by guaranteeing privacy even with untrusted data collector. Many differential privacy algorithms handles parameters like privacy budget, sensitivity and data utility in different ways and mostly trying to keep trade-off between privacy and utility of data. This paper evaluates differential privacy algorithms in regard to the privacy support it offers according to the sensitivity of the data. Generalized application of privacy budget is found ineffective in comparison to the sensitivity based usage of privacy budget.

Big Data, Big Expectations, and Big Judgements.

Public trust is paramount for the well functioning of data driven healthcare activities such as digital health interventions, contact tracing or the build-up of electronic health records. As the use of personal data is the common denominator for these healthcare activities, healthcare actors have an interest to ensure privacy and anonymity of the personal data they depend on. Maintaining privacy and anonymity of personal data contribute to the trustworthiness of these healthcare activities and are associated with the public willingness to trust these activities with their personal data. An analysis of online news readership comments about the failed care.data programme in England revealed that parts of the public have a false understanding of anonymity in the context of privacy protection of personal data as used for healthcare management and medical research. Some of those commenting demanded complete anonymity of their data to be willing to trust the process of data collection and analysis. As this demand is impossible to fulfil and trust is built on a false understanding of anonymity, the inability to meet this demand risks undermining public trust. Since public concerns about anonymity and privacy of personal data appear to be increasing, a large-scale information campaign about the limits and possibilities of anonymity with respect to the various uses of personal health data is urgently needed to help the public to make better informed choices about providing personal data.

Data anonymization is commonly utilized for the protection of an individual's identity when his personal or sensitive data is published. A well-known anonymization model to define the privacy of transactional data is the km-anonymity model. This model ensures that an adversary who knows up to m items of an individual cannot determine which record in the dataset corresponds to the individual with a probability greater than 1/k. However, the existing techniques generally rely on the presence of similarity between items in the dataset tuples to achieve km-anonymization and are not suitable when transactional data contains tuples without many common values. The authors refer to this type of transactional data as diverse transactional data and propose an algorithm, anonymization of diverse transactional data (ADT). ADT is based on slicing and generalization to achieve km-anonymity for diverse transactional data. ADT has been experimentally evaluated on two datasets, and it has been found that ADT yields higher privacy protection and causes a lower loss in data utility as compared to existing methods.

Efficient perturbation techniques for preserving privacy of multivariate sensitive data

Healthcare organizations are undergoing tremendous change in analytical computing. The vehicle for this change is implementing a Business Intelligence (BI) platform, which includes On-Line Analytical Processing (OLAP) Cubes, data visualization tools, and integrated data warehouse (DW). Extracting sensitive data (Personally Identifiable Information-PII, Personal Health Information-PHI) from operational databases, then saving them into a consolidated central repository to facilitate efficient analysis are the best solution. This solution demonstrates a good understanding of overall business operational performance and trends as well as improves the business processes through using statistical and data mining methods. Nonetheless, this huge data warehouse (DW) constitutes one of the most serious privacy breach threats that any healthcare organization might face when many internal users of different security levels have access to BI components. Data masking techniques are used to minimize the inadvertent disclosure risk of sensitive data as well as to preserve the basic quality of data analytics (data utility). However, the traditional masking methods fail to maintain the utility of data analysis for reporting and research purposes. In this paper a practical classification component for a built-in data masking framework (IMETU-Identify, Map, Execute, Test, and Utilize) is proposed and focuses on the first two modules. The analyzed health data attributes are based on the Discharge Abstract Database (DAD - Acute Inpatient in Canada). The proposed component is to identify the sensitive data, and select the best masking format to provide more data privacy and protection at rest (i.e., it can accurately be drilled down to a monthly aggregated level). This Module allows sensitive data attributes to be safely used and complies with the privacy regulatory requirements within the healthcare data warehouse by mapping them with the proper irreversible or reversible masking techniques. Native encryption techniques are avoided due to complex calculation and increase storage space.

The extensive prevalence of computing technologies in our daily lives and physical landscapes has led to the generation of vast datasets for analysis. Although, there is a growing uncertaintyregarding potential privacy breaches, as sensitive data could be exposed if not adequately protected during analysis. Many existing privacy-preserving methods encounter challenges such as inefficiency, scalability limitations, and the delicate balance between data utility and privacy conservation.
 In this study, we introduce an algorithm known as PABIDOT. PABIDOT employs optimal geometric transformations to safeguard privacy within the realm of big data. We assess the efficacy of PABIDOT through a series of experiments involving nine distinct datasets and five classification algorithms. Our results highlight PABIDOT's exceptional execution speed, scalability, resilience against potential attacks,andits accuracy in large-scale datasets.In addition, we delve into the practical implications of PABIDOT in real-world scenarios, emphasizing its adaptability across diverse industries and applications. The algorithm's robustness in safeguarding sensitive information while maintaining data integrity sets a new standard in privacy preservation. Moreover, the study sheds light on potential avenues for further advancements in the field of secure data analytics, paving the way for more comprehensive and effective privacy solutions in the era of burgeoning data generation and utilization.

ProblemMedication for opioid use disorder (MOUD) is recommended for persons with opioid use disorder (OUD) during pregnancy. However, knowledge gaps exist about best practices for management of OUD during pregnancy and these data are needed to guide clinical care.Period Covered2014–2021.Description of the SystemEstablished in 2019, the Maternal and Infant Network to Understand Outcomes Associated with Medication for Opioid Use Disorder During Pregnancy (MAT-LINK) is a surveillance network of seven clinical sites in the United States. Boston Medical Center, Kaiser Permanente Northwest, The Ohio State University, and the University of Utah were the initial clinical sites in 2019. In 2021, three clinical sites were added to the network (the University of New Mexico, the University of Rochester, and the University of South Florida). Persons receiving care at the seven clinical sites are diverse in terms of geography, urbanicity, race and ethnicity, insurance coverage, and type of MOUD received. The goal of MAT-LINK is to capture demographic and clinical information about persons with OUD during pregnancy to better understand the effect of MOUD on outcomes and, ultimately, provide information for clinical care and public health interventions for this population. MAT-LINK maintains strict confidentiality through robust information technology architecture. MAT-LINK surveillance methods, population characteristics, and evaluation findings are described in this inaugural surveillance report. This report is the first to describe the system, presenting detailed information on funding, structure, data elements, and methods as well as findings from a surveillance evaluation. The findings presented in this report are limited to selected demographic characteristics of pregnant persons overall and by MOUD treatment status. Clinical and outcome data are not included because data collection and cleaning have not been completed; initial analyses of clinical and outcome data will begin in 2023.ResultsThe MAT-LINK surveillance network gathered data on 5,541 reported pregnancies with a known pregnancy outcome during 2014–2021 among persons with OUD from seven clinical sites. The mean maternal age was 29.7 (SD = ±5.1) years. By race and ethnicity, 86.3% of pregnant persons were identified as White, 25.4% as Hispanic or Latino, and 5.8% as Black or African American. Among pregnant persons, 81.6% had public insurance, and 84.4% lived in urban areas. Compared with persons not receiving MOUD during pregnancy, those receiving MOUD during pregnancy were more likely to be older and White and to have public insurance. The evaluation of the surveillance system found that the initial four clinical sites were not representative of demographics of the South or Southwest regions of the United States and had low representation from certain racial and ethnic groups compared with the overall U.S. population; however, the addition of three clinical sites in 2021 made the surveillance network more representative. Automated extraction and processing improved the speed of data collection and analysis. The ability to add new clinical sites and variables demonstrated the flexibility of MAT-LINK.InterpretationMAT-LINK is the first surveillance system to collect comprehensive, longitudinal data on pregnant person–infant dyads with perinatal outcomes associated with MOUD during pregnancy from multiple clinical sites. Analyses of clinical site data demonstrated different sociodemographic characteristics between the MOUD and non-MOUD treatment groups.Public Health ActionsMAT-LINK is a timely and flexible surveillance system with data on approximately 5,500 pregnancies. Ongoing data collection and analyses of these data will provide information to support clinical and public health guidance to improve health outcomes among pregnant persons with OUD and their children.

Collection and management of clinical data for administration and analysis is a time-consuming and complex task, especially when multiple data providers been involved. Even if people are willing to take on the burden for it, there is still no mature solution to protect data privacy for distributed data providers. Distributed ledger is an emerging technology that supports decentralized data sharing and management. Based on this, we present a platform which enables distributed and truthful data collection and serves privacy-preserving needs in clinical data management. Our system, built on Hyperledger Fabric, used smart contract to execute data aggregation and provide basic analysis methods. The system used ledger and world status to record data access history and other metadata. This decentralized platform enables data providers to proactively share and protect their data, Thus can simplify clinical data collection procedure and promote efficient collaboration between providers.

Personal information and other types of private data are valuable for both data owners and institutions interested in providing targeted and customized services that require analyzing such data. In this context, privacy is sometimes seen as a commodity: institutions (data buyers) pay individuals (or data sellers) in exchange for private data. In this study, we examine the problem of designing such data contracts, through which a buyer aims to minimize his payment to the sellers for a desired level of data quality, while the latter aim to obtain adequate compensation for giving up a certain amount of privacy. Specifically, we use the concept of differential privacy and examine a model of linear and nonlinear queries on private data. We show that conventional algorithms that introduce differential privacy via zero-mean noise fall short for the purpose of such transactions as they do not provide a sufficient degree of freedom for the contract designer to negotiate between the competing interests of the buyer and the sellers. Instead, we propose a biased randomized algorithm to generate differentially private output and show that this algorithm allows us to customize the privacy-accuracy tradeoff for each individual. We use a contract design approach to find the optimal contracts when using this biased algorithm to provide privacy and show that under this combination the buyer can achieve the same level of accuracy with a lower payment as compared to using the conventional, unbiased algorithms, while at the same time incurring lower privacy loss for the sellers.

Background – Depression has a profound and complex association with coronary risk, both as an independent risk factor and indirectly through its effects on determinants of coronary risk such as BMI, diet, physical activity, smoking and medication treatment fidelity. Of particular interest is the impact of depression symptoms on initiation of and adherence to behavioural change to reduce coronary risk. Objective – To investigate the association of depression with dietary adherence and weight loss outcomes in a community-based weight loss intervention. Design – A group of 64 overweight (BMI >27), otherwise healthy adults, were recruited and randomised to follow either their usual diet, or an isocaloric diet in which saturated fat was replaced with MUFA to 50% by adding macadamia nuts to the diet. Subjects were assessed for depressive symptoms at baseline and at ten weeks using the Beck Depression Inventory (BDI-II). Both control and intervention groups received advice on National Guidelines for Physical Activity and adhered to the same protocol for food diaries and trial consultations. Anthropometric and clinical measurements were taken at baseline and 10 weeks. Outcomes – There was a significant correlation (R=-0.38, p<0.05) between BDI-II scores at baseline and duration of participation in the trial. Subjects with a baseline BDI≥10 (moderate to severe depression symptoms) were more likely to dropout of the trial before 10 weeks (p<0.001). The BDI-II scores in the intervention (monounsaturated fat) diet group decreased, but increased in the control group over the 10-week period. Univariate analysis of variance confirmed these observations (adjusted R 2 = 0.257, p = 0.01). Body weight remained static over the 10-week period in the intervention group, corresponding to a relative increase in the control group (adjusted R 2 = 0.097, p = 0.064). Conclusion – Depression symptoms have the potential to affect adherence to diet and physical activity-based risk reduction interventions, and may therefore be useful screening criteria.

In the current framework of human rights, data privacy is finding its home as an independent human right, separate from its historical home under the umbrella of general privacy. However, there is no consistent system or standard for defining this right, and different regions require substantially different levels of protection. This inconsistency has allowed for corporations, by way of their executive officers, to avoid or completely ignore the requirements imposed by many countries. Moreover, the penalties in many regions are not severe enough to incentivize corporations to change their behavior. The lack of a truly global system and standard for enforcing this right, and the specific lack of pressure on the officers that direct corporate policy, has allowed data privacy violations to go severely under-checked. This Note seeks to provide a novel solution for tackling corporate holdup in complying with data privacy laws. This Note examines the historical roots of data privacy as a human right, discusses its similarity to resources that have been considered public utilities, and provides examples of instances where the right to data privacy has been ignored by corporate officers. By modifying the United Nations’ existing sanction procedure and jurisdiction, this Note proposes that the United Nations would be able to target corporate officers individually for their roles in data privacy violations. By leveraging personal liability for noncompliance, the United Nations could generate a global sense of accountability to the modern, human right to privacy in one’s personal data.