Size matters: how population size influences genotype-phenotype association studies in anonymized data

Statistical biases due to anonymization evaluated in an open clinical dataset from COVID-19 patients

Anonymization has the potential to foster the sharing of medical data. State-of-the-art methods use mathematical models to modify data to reduce privacy risks. However, the degree of protection must be balanced against the impact on statistical properties. We studied an extreme case of this trade-off: the statistical validity of an open medical dataset based on the German National Pandemic Cohort Network (NAPKON), which was prepared for publication using a strong anonymization procedure. Descriptive statistics and results of regression analyses were compared before and after anonymization of multiple variants of the original dataset. Despite significant differences in value distributions, the statistical bias was found to be small in all cases. In the regression analyses, the median absolute deviations of the estimated adjusted odds ratios for different sample sizes ranged from 0.01 [minimum = 0, maximum = 0.58] to 0.52 [minimum = 0.25, maximum = 0.91]. Disproportionate impact on the statistical properties of data is a common argument against the use of anonymization. Our analysis demonstrates that anonymization can actually preserve validity of statistical results in relatively low-dimensional data.

Design and evaluation of a data anonymization pipeline to promote Open Science on COVID-19

The Lean European Open Survey on SARS-CoV-2 Infected Patients (LEOSS) is a European registry for studying the epidemiology and clinical course of COVID-19. To support evidence-generation at the rapid pace required in a pandemic, LEOSS follows an Open Science approach, making data available to the public in real-time. To protect patient privacy, quantitative anonymization procedures are used to protect the continuously published data stream consisting of 16 variables on the course and therapy of COVID-19 from singling out, inference and linkage attacks. We investigated the bias introduced by this process and found that it has very little impact on the quality of output data. Current laws do not specify requirements for the application of formal anonymization methods, there is a lack of guidelines with clear recommendations and few real-world applications of quantitative anonymization procedures have been described in the literature. We therefore believe that our work can help others with developing urgently needed anonymization pipelines for their projects.

Detecting the Presence of an Individual in Phenotypic Summary Data

As the quantity and detail of association studies between clinical phenotypes and genotypes grows, there is a push to make summary statistics widely available. Genome wide summary statistics have been shown to be vulnerable to the inference of a targeted individual's presence. In this paper, we show that presence attacks are feasible with phenome wide summary statistics as well. We use data from three healthcare organizations and an online resource that publishes summary statistics. We introduce a novel attack that achieves over 80% recall and precision within a population of 16,346, where 8,173 individuals are targets. However, the feasibility of the attack is dependent on the attacker's knowledge about 1) the targeted individual and 2) the reference dataset. Within a population of over 2 million, where 8,173 individuals are targets, our attack achieves 31% recall and 17% precision. As a result, it is plausible that sharing of phenomic summary statistics may be accomplished with an acceptable level of privacy risk.

Enhancing Reuse of Data and Biological Material in Medical Research: From FAIR to FAIR-Health

The known challenge of underutilization of data and biological material from biorepositories as potential resources for medical research has been the focus of discussion for over a decade. Recently developed guidelines for improved data availability and reusability-entitled FAIR Principles (Findability, Accessibility, Interoperability, and Reusability)-are likely to address only parts of the problem. In this article, we argue that biological material and data should be viewed as a unified resource. This approach would facilitate access to complete provenance information, which is a prerequisite for reproducibility and meaningful integration of the data. A unified view also allows for optimization of long-term storage strategies, as demonstrated in the case of biobanks. We propose an extension of the FAIR Principles to include the following additional components: (1) quality aspects related to research reproducibility and meaningful reuse of the data, (2) incentives to stimulate effective enrichment of data sets and biological material collections and its reuse on all levels, and (3) privacy-respecting approaches for working with the human material and data. These FAIR-Health principles should then be applied to both the biological material and data. We also propose the development of common guidelines for cloud architectures, due to the unprecedented growth of volume and breadth of medical data generation, as well as the associated need to process the data efficiently.

A multi-institution evaluation of clinical profile anonymization

BACKGROUND AND OBJECTIVE

There is an increasing desire to share de-identified electronic health records (EHRs) for secondary uses, but there are concerns that clinical terms can be exploited to compromise patient identities. Anonymization algorithms mitigate such threats while enabling novel discoveries, but their evaluation has been limited to single institutions. Here, we study how an existing clinical profile anonymization fares at multiple medical centers.

METHODS

We apply a state-of-the-artk-anonymization algorithm, withkset to the standard value 5, to the International Classification of Disease, ninth edition codes for patients in a hypothyroidism association study at three medical centers: Marshfield Clinic, Northwestern University, and Vanderbilt University. We assess utility when anonymizing at three population levels: all patients in 1) the EHR system; 2) the biorepository; and 3) a hypothyroidism study. We evaluate utility using 1) changes to the number included in the dataset, 2) number of codes included, and 3) regions generalization and suppression were required.

RESULTS

Our findings yield several notable results. First, we show that anonymizing in the context of the entire EHR yields a significantly greater quantity of data by reducing the amount of generalized regions from ∼15% to ∼0.5%. Second, ∼70% of codes that needed generalization only generalized two or three codes in the largest anonymization.

CONCLUSIONS

Sharing large volumes of clinical data in support of phenome-wide association studies is possible while safeguarding privacy to the underlying individuals.