Dataset Information

Global healthcare fairness: We should be sharing more, not less, data.

ABSTRACT: The availability of large, deidentified health datasets has enabled significant innovation in using machine learning (ML) to better understand patients and their diseases. However, questions remain regarding the true privacy of this data, patient control over their data, and how we regulate data sharing in a way that that does not encumber progress or further potentiate biases for underrepresented populations. After reviewing the literature on potential reidentifications of patients in publicly available datasets, we argue that the cost-measured in terms of access to future medical innovations and clinical software-of slowing ML progress is too great to limit sharing data through large publicly available databases for concerns of imperfect data anonymization. This cost is especially great for developing countries where the barriers preventing inclusion in such databases will continue to rise, further excluding these populations and increasing existing biases that favor high-income countries. Preventing artificial intelligence's progress towards precision medicine and sliding back to clinical practice dogma may pose a larger threat than concerns of potential patient reidentification within publicly available datasets. While the risk to patient privacy should be minimized, we believe this risk will never be zero, and society has to determine an acceptable risk threshold below which data sharing can occur-for the benefit of a global medical knowledge system.

SUBMITTER: Seastedt KP

PROVIDER: S-EPMC9931202 | biostudies-literature | 2022 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Global healthcare fairness: We should be sharing more, not less, data.

Seastedt Kenneth P KP Schwab Patrick P O'Brien Zach Z Wakida Edith E Herrera Karen K Marcelo Portia Grace F PGF Agha-Mir-Salim Louis L Frigola Xavier Borrat XB Ndulue Emily Boardman EB Marcelo Alvin A Celi Leo Anthony LA

PLOS digital health 20221006 10

The availability of large, deidentified health datasets has enabled significant innovation in using machine learning (ML) to better understand patients and their diseases. However, questions remain regarding the true privacy of this data, patient control over their data, and how we regulate data sharing in a way that that does not encumber progress or further potentiate biases for underrepresented populations. After reviewing the literature on potential reidentifications of patients in publicly ...[more]

PMID: 36812599

Similar Datasets

Project description:Introduction: Health research is gradually embracing a more collectivist approach, fueled by a new movement of open science, data sharing and collaborative partnerships. However, the existence of systemic contradictions hinders the sharing of health data and such collectivist endeavor. Therefore, this qualitative study explores these systemic barriers to a fair sharing of health data from the perspectives of Swiss stakeholders. Methods: Purposive and snowball sampling were used to recruit 48 experts active in the Swiss healthcare domain, from the research/policy-making field and those having a high position in a health data enterprise (e.g., health register, hospital IT data infrastructure or a national health data initiative). Semi-structured interviews were then conducted, audio-recorded, verbatim transcribed with identifying information removed to guarantee the anonymity of participants. A theoretical thematic analysis was then carried out to identify themes and subthemes related to the topic of systemic fairness for sharing health data. Results: Two themes related to the topic of systemic fairness for sharing health data were identified, namely (i) the hypercompetitive environment and (ii) the legal uncertainty blocking data sharing. The theme, hypercompetitive environment was further divided into two subthemes, (i) systemic contradictions to fair data sharing and the (ii) need of fair systemic attribution mechanisms. Discussion: From the perspectives of Swiss stakeholders, hypercompetition in the Swiss academic system is hindering the sharing of health data for secondary research purposes, with the downside effect of influencing researchers to embrace individualism for career opportunities, thereby opposing the data sharing movement. In addition, there was a perceived sense of legal uncertainty from legislations governing the sharing of health data, which adds unreasonable burdens on individual researchers, who are often unequipped to deal with such facets of their data sharing activities.

Project description:Fairness violations elicit powerful behavioral and affective responses. Indeed, people are willing to incur costs to sanction unfair behavior. Here we study the possible impact of long-term mental training in socio-affective capacities such as compassion on altruistic punishment and compensatory behavior in economic games. To this end we recruited a group of long-term meditation practitioners (LTPs) who had engaged in an average of 40 K h of mental training exercises including compassion-related meditation, along with a group of meditation-naïve controls. Participants played several adaptations of the dictator game in which they had the opportunity to punish the dictator both when they were the recipients of the dictator's offer and when they were third-party witnesses to the dictator's treatment of an anonymous second player. Compared to controls, LTPs were less likely to punish when they were the victims of fairness violations. However, both groups punished equivalently when they witnessed others receiving unfair treatment. In post-task questionnaires, controls reported significantly more anger in response to unfair offers than LTPs, although fairness judgments did not differ between groups. These data suggest that because the LTPs were less angered by unfair treatment of themselves, they punished that behavior less. However, when they witnessed the unfair treatment of others, they engaged in norm-reinforcing punishment. Finally, when participants played an additional game which included the opportunity to recompense victims, LTPs were more likely to do so. Together these data point to differential approaches to justice whereby LTPs engaged less in vengeful, retributive justice and focused more on norm reinforcement and the restoration of equity. These differences suggest that social preferences are plastic and that altruistic responses to unfairness may be shaped by the prolonged cultivation of prosocial motivation, altruism, and compassion.

Dataset Information

Global healthcare fairness: We should be sharing more, not less, data.

Publications

Global healthcare fairness: We should be sharing more, not less, data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets