Abstract
Data sharing within psychiatric and behavioral research represents a novel application of ethical principles in practice; however, it suffers from a dearth of practical experience and established ethical norms. In this study, we comprehensively examined the ethical considerations surrounding the acquisition, management, sharing, and utilization of such data. We graded sensitive data and suggest ethical standards for privacy protection based on varying levels of data sensitivity. The objective of this study is to foster orderly and standardized open sharing of psychiatric and behavioral research data, thereby advancing the development and progress of related academic disciplines in China. This Chinese expert consensus has been registered on the International Guide Registration platform (Registration Number: PREPARE-2024CN412).
Introduction
The global movement toward open science has catalyzed a significant increase in research data sharing. In the realm of psychiatry and behavioral science, this trend offers immense potential for accelerating discoveries and improving mental health outcomes. However, these data types present unique challenges: they are inherently subjective, highly diverse, and deeply sensitive.
Unlike general health data, psychiatric records often carry a high risk of social stigma and discrimination. A data breach in this field can lead to severe psychological distress for participants, potentially exacerbating the very conditions being studied. Despite the proliferation of open databases—such as the Psychology and Behavior Investigation of Chinese Residents (PBICR)—the field has lacked a standardized, ethical framework to govern data handling. This paper presents an expert consensus designed to bridge this gap, aligning international ethical standards with domestic legal requirements.
Methodology and Framework Design
This guideline was developed through a comprehensive synthesis of legal and ethical principles. The authors integrated international frameworks, including the General Data Protection Regulation (GDPR), the Health Insurance Portability and Accountability Act (HIPAA), and the 1978 Belmont Report, with Chinese domestic laws such as the Civil Code, the Personal Information Protection Law (PIPL), and the Data Security Law.
Formulated on behalf of the Behavioral Epidemiology Education Working Commission of the CMEA, the consensus centers on a novel Three-Tier Data Sensitivity Hierarchy. This hierarchy maps specific technical protections—such as AES/RSA encryption and multi-factor authentication—to the risk level of the data.
The Three-Tier Sensitivity Hierarchy
The core of the consensus is the classification of data into three distinct categories, each requiring specific handling protocols:
- Highly Sensitive Data: Includes personally identifiable information (PII), detailed mental health assessments, and precise geolocation. These data require the highest level of protection, including rigorous de-identification, multi-factor access control, and formal ethical review prior to any sharing.
- Moderately Sensitive Data: Encompasses general demographics and behavioral observation data (e.g., reaction times). Handling requires standard anonymization, strict data quality assurance, and the execution of binding Data Use Agreements (DUAs).
- Minimally Sensitive Data: Refers to aggregated research findings and processed statistical results. These require clear rules for public disclosure and ongoing monitoring to ensure compliant use.
Ethical Accountability and Stakeholder Responsibilities
A key insight of this framework is the division of ethical responsibility between data collectors and data users.
- Data Collectors are tasked with data minimization (collecting only what is necessary), ensuring transparent informed consent, and implementing robust encryption standards.
- Data Users must strictly adhere to DUAs, prevent "scope creep" (using data for unauthorized research purposes), and undergo mandatory data ethics training.
The authors emphasize that ethical data sharing is a dynamic process. There is a constant trade-off between anonymity and usability; excessive anonymization can strip behavioral data of its scientific value. Therefore, protection strategies must be adaptive, balancing the removal of identifiers with the preservation of statistical utility.
Conclusion and Future Directions
This expert consensus, registered as PREPARE-2024CN412, provides a vital blueprint for institutions involved in psychiatric research. While it establishes a strong foundation, the authors acknowledge that anonymization is not a permanent solution. As computational power and data linkage techniques advance, the risk of re-identification grows.
Future work must focus on the continuous evolution of de-identification technologies and the establishment of robust, inter-departmental regulatory frameworks. By integrating data ethics governance at a national level, the scientific community can ensure that the pursuit of open science does not come at the cost of individual privacy and dignity.
Summary updated