Opt-In and Opt-Out Consent Procedures for the Reuse of Routinely Recorded Health Data in Scientific Research and Their Consequences for Consent Rate and Consent Bias: Systematic Review

Simulated Misuse of Large Language Models and Clinical Credit Systems

Large language models (LLMs) have been proposed to support many healthcare tasks, including disease diagnostics and treatment personalization. While AI may be applied to assist or enhance the delivery of healthcare, there is also a risk of misuse. LLMs could be used to allocate resources based on unfair, inaccurate, or unjust criteria. For example, a social credit system uses big data to assess "trustworthiness" in society, punishing those who score poorly based on evaluation metrics defined only by a power structure (corporate entity, governing body). Such a system may be amplified by powerful LLMs which can rate individuals based on multimodal data - financial transactions, internet activity, and other behavioural inputs. Healthcare data is perhaps the most sensitive information which can be collected and could potentially be used to violate civil liberty via a "clinical credit system", which may include limiting or rationing access to standard care. This report simulates how clinical datasets might be exploited and proposes strategies to mitigate the risks inherent to the development of AI models for healthcare.

From opt-out to opt-in consent for secondary use of medical data and residual biomaterial: An evaluation using the RE-AIM framework

BACKGROUND

Patient records, imaging, and residual biomaterial from clinical procedures are crucial resources for medical research. In the Netherlands, consent for secondary research has historically relied on opt-out consent. For ethical-legal experts who purport passive consent undermines patient autonomy, opt-in consent (wherein affirmative action is required) is seen as the preferred standard. To date, there is little empirical research exploring patient feasibility, organizational consequences, and the potential risks for research based on secondary data. Thus, we applied the RE-AIM framework to evaluate the impact of migrating from an opt-out to an opt-in consent process.

METHODS

This evaluation was carried out in Radboud University Medical Center, a large tertiary hospital located in the southeast of the Netherlands. All non-acute, mentally competent patients ≥16 years of age registered between January 13, 2020 and June 30, 2023 were targeted (N = 101,437). In line with the RE-AIM framework, individual and organizational consequences were evaluated across five domains: reach, efficacy, adoption, implementation, and maintenance.

RESULTS

101,437 eligible patients were approached of whom 66,214 (65.3%) consented, 8,059 (7.9%) refused consent and 27,164 (26.8%) had no response. Of the 74,273 patients with a response, 89.1% consented to secondary use. The migration to an opt-in consent system was modestly successful; yet notably, differential response patterns by key sociodemographic characteristics were observed. Adaptions to the process flow improved its effectiveness and resulted in a reasonable response over time. Implementation was most affected by budgetary restraints, thus impeding the iterative approach which could have further improved domain outcomes.

CONCLUSION

This evaluation provides an overview of logistical and pragmatic issues encountered when migrating from opt-out to opt-in consent. Response bias remains a major concern. Though not always directly transferable, these lessons can be broadly used to inform other health care organizations of the potential advantages and pitfalls of an opt-in consent system.

Strategies for Implementing Machine Learning Algorithms in the Clinical Practice of Radiology

Despite recent advancements in machine learning (ML) applications in health care, there have been few benefits and improvements to clinical medicine in the hospital setting. To facilitate clinical adaptation of methods in ML, this review proposes a standardized framework for the step-by-step implementation of artificial intelligence into the clinical practice of radiology that focuses on three key components: problem identification, stakeholder alignment, and pipeline integration. A review of the recent literature and empirical evidence in radiologic imaging applications justifies this approach and offers a discussion on structuring implementation efforts to help other hospital practices leverage ML to improve patient care. Clinical trial registration no. 04242667 © RSNA, 2024 Supplemental material is available for this article.