Designing Value-Centered Consent Interfaces: A Mixed-Methods Approach to Support Patient Values in Data-Sharing Decisions

TL;DR

This paper tackles the ethical and practical shortcomings of existing consent interfaces in health data sharing, which often privilege data collectors’ values over patients’. Using an exploratory sequential mixed-methods design, the authors elicit patient values via participatory workshops, design a Value-Centered Consent Interface incorporating transparency, autonomy, and a reflection prompt, evaluate its effects in online experiments, and contextualize refinements through a co-creation workshop with domain experts before testing with hospital inpatients. The findings show that value-centered reflection can improve value congruence for privacy-valuing users and that knowledge about data-use options strongly supports deliberate, value-aligned decisions, though prompts alone may not universally enhance decision quality. The study demonstrates how a human-centered design process can ground consent interfaces in real patient values within a concrete health data donation context, offering actionable design guidelines for CSCW and digital health practitioners. Overall, the work advances ethical data practices by showing how reflection, multi-modal information, and grounded context can empower patients to exercise autonomy in data-sharing decisions with greater alignment to their values.

Abstract

In the digital health domain, ethical data collection practices are crucial for ensuring the availability of quality datasets that drive medical advancement. Data donation, allowing patients to share their medical data for secondary research purposes, presents a promising resource for such datasets. Yet, current consent user interfaces mediating data-sharing decisions are found to favor data collectors' values over those of data subjects. This raises ethical concerns about the use of data collected, as well as concerning the quality of the resulting datasets. Seeking to establish value-centered data collection practices in digital health, we investigate the design of consent user interfaces that support end-users in making value-congruent health data-sharing decisions. Focusing our research efforts on the situated context of health data donation at the psychosomatic unit of a university hospital, we demonstrate how a human-centered design can ground technology within the perspective of a vulnerable group. We employed an exploratory sequential mixed-method approach consisting of five phases: Participatory workshops elicit patient values, informing the design of a proposed Value-Centered Consent Interface. An online experiment demonstrates our interface element's effect, increasing value congruence in data-sharing decisions. Our proposed consent user interface design is then adapted to the research context through a co-creation workshop with domain experts and a user interface evaluation with patients. Our work contributes to recent discourse in CSCW concerning ethical implications of new data practices within their socio-technological context by exploring patient values on medical data-sharing, introducing a novel consent interface leveraging reflection to support value-congruent decision-making, and providing a situated evaluation of the proposed consent user interface with patients.

Figures (7)

• Figure 1: Overview of the five subsequent phases of the exploratory sequential mixed-methods approach for design and evaluation of the Value-Centered Consent Interface.
• Figure 2: Overview of the participatory value elicitation workshop procedure consisting of four sequential phases. Representations adapted from sorriesAdvocatingValuesMeaningful2024a.
• Figure 3: Consent interface including the Disclosure Options Ⓐ selector and the Reflection Prompt Ⓑ.
• Figure 4: Overview of the data donation consent interface incorporating the Value-Centered Reflection PromptⒶ accompanied by the four possible congruence messages with their accompanying visualization Ⓑ.
• Figure 5: Overview of the experimental study procedure and the three experimental groups, representing differing realizations of the (Value-Centered) Reflection-Prompt.