8145 research outputs found
Sort by
Standardization and Accuracy of Race and Ethnicity Data: Equity Implications for Medical AI
No full textAbstract
The rapid integration of artificial intelligence (AI) into healthcare has raised many concerns about race bias in AI models. Yet, overlooked in this dialogue is the lack of quality control for the accuracy of patient race and ethnicity (r/e) data in electronic health records (EHR). This article critically examines the factors driving inaccurate and unrepresentative r/e datasets. These include conceptual uncertainties about how to categorize races and ethnicity, shortcomings in data collection practices, EHR standards, and the misclassification of patients’ race or ethnicity. To address these challenges, we propose a two-pronged action plan. First, we present a set of best practices for healthcare systems and medical AI researchers to improve r/e data accuracy. Second, we call for developers of medical AI models to transparently warrant the quality of their r/e data. Given the ethical and scientific imperatives of ensuring high-quality r/e data in AI-driven healthcare, we argue that these steps should be taken immediately.
Author summary
Healthcare systems are increasingly using artificial intelligence (AI) to improve clinical care in various settings such as hospitals and patient care facilities. In this paper, we discuss how these AI systems may be trained using inaccurate and incomplete patient race and ethnicity data. We identify several key issues underlying this data quality problem: the conceptual challenges in defining race and ethnicity categories, inconsistent data collection practices across healthcare facilities, and frequent errors in classifying patients. These problems create unreliable training data that undermines efforts to avoid and correct biases within these medical AI tools. To address these challenges, we propose two practical solutions. First, hospitals should adopt best practices for collecting race and ethnicity information, including patient self-reporting, staff training, and transparent processes. Second, developers of medical AI should be required to disclose the quality and sources of the demographic data used to train their models. Our work emphasizes that discussions about fairness in medical AI must include attention to the quality of race and ethnicity data. As these technologies become more widespread in healthcare, ensuring they work effectively for all patients requires addressing these fundamental data issues
Revolutionizing Brain Research Using Portable MRI in Field Settings: Public Perspectives on the Ethical and Legal Challenges
No full textIntroduction New, highly portable MRI (pMRI) technology promises to revolutionize brain research by facilitating field-based studies that can expand research to new settings beyond the traditional MRI suite in a medical center. At this early stage of development, understanding public knowledge and attitudes about pMRI research is crucial. Objective In this article we present the first empirical study of the general public’s willingness to participate in pMRI research, and their perceptions of expected benefits and concerns. Methods & Results We conducted a nationally representative online survey (N = 2,001) administered Aug. 15-31, 2022. We found that respondents were overwhelmingly willing to participate in pMRI research, with no significant differences between five key demographic sub-groups: rural residents, older adults (65+), Hispanics, non-Hispanic Blacks, and those economically disadvantaged. Respondents saw many potential benefits (e.g., follow-up information about the study’s results) and few concerns (e.g., insufficient payment) associated with participating. Conclusion Such high public interest in participating confirms the importance of developing ethical guidance for pMRI research now, before that research rapidly expands. The results speak to the importance of minimizing the therapeutic misconception in pMRI research, as the survey reveals gaps in participant knowledge about the capabilities and limitations of pMRI devices to provide clinically informative scans. Our data showed that a lack of trust in scientists can reduce likelihood of participation, and thus researchers will need to engage participant communities to fully realize the potential of pMRI research to reach remote and historically underrepresented populations
Liability, Property, and Inalienability Rules in Employee Data Regulation
No full textLegal protections for workers’ data have usually taken the form of privacy protections designed to deter data processing that is excessive or invasive. Such protections generally fall into the category of liability rules, under which rights can be infringed as long as compensation is provided for the violation. As Guido Calabresi and A. Douglas Melamed have described in “Property Rules, Liability Rules, and Inalienability: One View of the Cathedral,” liability rules are contrasted with either property rules that prevent the involuntary transfer of rights or inalienability rules that prohibit rights transfers altogether. This article explores how property rules and inalienability rules could provide better protections for employee data rights in certain contexts. Property rules would allow employees to maintain control over their data, requiring employers to negotiate for its use rather than unilaterally collecting and processing it. Inalienability rules could shield particularly sensitive categories of worker data — such as biometric information or private communications — by imposing strict limitations and severe penalties on their collection and use. By rethinking the rules governing employee data, this article advocates for a more equitable approach of mixed regulatory approaches, alternatively providing workers with compensation, greater economic power, or legal barriers to any potential processing
The Good, the Bad, and the Ugly: A Comparative Constitutional Analysis of Whistleblowing Speech, the Government\u27s Managerial Domain, and the Imperatives of Democratic Self-Government
No full textSince issuing its 1968 landmark decision in Pickering, which first recognized that the First Amendment protects government employees’ speech about matters of public concern, the U.S. Supreme Court has proceeded to whittle away First Amendment protections for government employees. The Justices have done so by adopting a series of categorical exclusions to Pickering that all strongly favor the government as an employer and manager. These subsequent decisions have created a jurisprudential obstacle course that government employees must successfully run in order to invoke the Free Speech Clause at all. The current U.S. approach is plainly bad. However, it could be even worse—it could be ugly. In Australia, the High Court has given the government a green light to censor any and all government employee speech under viewpoint-based speech regulations. Thus, in today’s Australia, it’s perfectly fine for a public servant to praise the government but not to criticize it.
By way of contrast, in Canada, no categorical exclusions exist on the scope of constitutionally protected government employee speech, and the government must always be prepared to justify disciplinary actions based on a government employee’s speech activity. Canada’s approach is good—and clearly better than either the U.S. or Australian doctrines. By taking context fully into account, Canada’s government employee speech doctrine allows for courts to consider carefully how to reconcile the three competing interests at stake (namely, the government’s interest as a manager of its workforce, government employees’ autonomy interests as would-be speakers, and the collective interest We the People possess in access to government employee speech in general and whistleblowing speech in particular). Canada has built a better mousetrap; the federal courts should seriously consider reforming the Pickering/Connick/Garcetti framework to more closely resemble the Supreme Court of Canada’s holistic approach
