Research Project: Trust and Reliance on AI Confidence Evaluation in Breast Cancer Clinical Care (TRACE-BC)
General Information & Study Invitation
Dear Colleague / Dear Participant,
We would like to invite you to participate in a scientific research study. This participant information provides all essential details regarding the study. Please read this information carefully. If you have any questions, we are at your disposal.
Healthy individuals involved in the care of breast cancer patients are eligible to take part in this study.
The study was planned by Universitätsklinikum Aachen and is conducted in cooperation with the Else Kröner Fresenius Zentrum Dresden in conjunction with the organisers of the St. Gallen Breast Conference. Our institution finances the study independently (self-funded). The study has been submitted to the responsible ethics committee, which raised no objections.
Participation in the study is voluntary. If you choose not to participate or if you later withdraw your consent, you will not experience any disadvantages. You do not need to state reasons for your decision. All data is collected and processed anonymously.
Why is this study being conducted?
AI-based clinical decision support systems are increasingly being integrated into oncological practice. However, the mechanisms by which stated AI confidence and model accuracy influence the clinical judgment of physicians remain insufficiently characterized. In particular, it is unclear to what extent clinicians are susceptible to automation bias — the tendency to adopt an erroneous AI recommendation when it is presented with high confidence. While this phenomenon has been documented in other high-risk domains (e.g., aviation, radiology), it has not yet been systematically investigated for guideline-based decisions in breast cancer care. A robust understanding of this human-AI interaction is a prerequisite for calibrating and implementing clinical AI systems so that they genuinely improve care quality rather than creating new sources of error.
In a within-subject factorial experiment, we investigate how correctness, stated confidence, and overall accuracy of a simulated AI (BCAI) influence the decisions of clinicians across 16 guideline-based clinical vignettes in breast cancer care. Participants first make an unassisted decision, then receive an AI recommendation with a defined confidence score, and can subsequently confirm or revise their answer. Comparing initial and final responses allows us to quantify the extent and direction of response shifts and their dependence on AI confidence and correctness, providing insights into automation bias and trust calibration toward AI decision-support aids.
How is the study conducted?
Participating in this study involves a single online session lasting approximately 15–30 minutes. The procedure is structured as follows:
- Pre-Study Questionnaire: At the beginning, we ask you to complete a short demographic and professional questionnaire (specialty, stage of training, and experience with AI-supported systems; approx. 2–3 minutes).
- Clinical Case Vignettes: In the main part of the study, you will evaluate 16 clinical case vignettes on breast cancer management based on the current St. Gallen International Breast Cancer Consensus (SGBCC 2025). For each vignette, you will go through two steps:
- Initial Decision: You read the clinical scenario and make your own unassisted decision. Additionally, you indicate how certain you feel about this decision.
- AI Recommendation & Final Decision: You then receive a recommendation from our simulated AI system (BCAI) including a stated confidence score. You can then confirm or change your original answer and re-evaluate your certainty.
Evaluating all 16 vignettes typically takes approximately 10–20 minutes.
- Post-Study Questionnaire: Finally, we ask you to complete a brief questionnaire regarding your trust in AI-supported decision aids and your general attitude toward AI in clinical practice (approx. 5 minutes).
Participation takes place in a single session; no follow-up appointments or further contact are planned. You may terminate your participation at any time without giving reasons simply by closing the browser window.
Is there a personal benefit from participating?
No direct personal benefit from your participation is to be expected. It is possible that you will derive no immediate advantage from taking part. However, the findings may help make future AI-based decision aids in breast cancer care more reliable — ultimately benefiting other physicians and patients in the long term.
Are there any costs? Do I receive compensation?
Participation in this study does not entail any costs for you. You will not receive monetary remuneration or expense compensation. However, among all participants, we will raffle two free registrations for the 20th St. Gallen International Breast Cancer Conference 2027 (SGBCC 2027). If you wish to enter the raffle, please voluntarily provide your email address at the end of the survey. Entering the raffle is completely independent of your answers and has no influence on the evaluation of your responses.
Who can I contact if I have further questions?
If you have further questions, please contact:
Mark Enrik Geißler
Medizinische Klinik III, Universitätsklinikum Aachen
Pauwelsstraße 30, 52074 Aachen, Germany
Email: egeissler@ukaachen.de
Joint Controllers (Art. 26 GDPR)
In this study, the study teams at Universitätsklinikum Aachen and Else Kröner Fresenius Zentrum Dresden are jointly responsible for data processing. The legal basis for data processing is your personal consent (Art. 6(1)(a) and Art. 9(2)(a) GDPR). All data is treated confidentially at all times and is collected and processed anonymously.
Data is collected exclusively for the purpose of the study described above and will be used solely within this framework.
It is possible that we may wish to contact you again after the study has concluded. If you agree, you may indicate your consent in the consent declaration.
Complete Anonymity & No Collection of Personal Data
We do not collect any personal identifying data. All data is recorded anonymously. Consequently, nobody — including the study investigators — can determine to whom the data belongs.
Hosting & Technical Infrastructure
Data is collected via self-developed software hosted on European servers provided by Render (Render Services, Inc., CA, USA) and stored on servers of Universitätsklinikum Aachen and the Else Kröner Fresenius Zentrum für Digitale Gesundheit.
The platform is designed as an event-driven web application (Node.js/Express backend, PostgreSQL database). It serves exclusively to conduct a cognitive-science behavioral experiment on human-AI interaction in clinical oncology.
- No Registration: No clear names, contact details, or user accounts are created.
- Pseudonymous Session Generation: Upon accessing the platform, the server cryptographically generates a random session ID (UUID v4). All inputs during the session are linked exclusively to this random ID.
- No IP Address Storage in Database: IP addresses are processed exclusively in volatile memory (in-memory) to protect the system against Denial-of-Service attacks and automated spam requests (rate limiting). No permanent logging or storage of IP addresses takes place in the database.
- Security Standards: Connections are fully encrypted (TLS/HTTPS), secured by robust security headers (Helmet, strict Content Security Policy, HSTS), and operate without third-party trackers, advertising networks, or analytics cookies.
Data Transfers
We do not transfer your data to other institutions in Germany, within the European Union, to third countries outside the EU, or to any international organization.
Scientific Publication & Research Data Sharing
Research data may be used for scientific publications and/or made available indefinitely to other researchers in scientific repositories. Data is shared only in a format that does not allow any inferences to be drawn regarding individual study participants (fully anonymized).
Voluntariness & Withdrawal of Consent
Consent to the processing of your data is entirely voluntary. You may withdraw your consent at any time without giving reasons and without any disadvantages to you. Following withdrawal, no further data will be collected. The lawfulness of processing carried out on the basis of consent prior to withdrawal remains unaffected.
Your Rights as a Data Subject
You have the right to obtain access to your data, including in the form of a free copy. In addition, you may request rectification, blocking, restriction of processing, or erasure, as well as data portability where applicable.
In such cases, please contact:
Mark Enrik Geißler
Medizinische Klinik III, Universitätsklinikum Aachen
Pauwelsstraße 30, 52074 Aachen, Germany
Email: egeissler@ukaachen.de
Important Note on Anonymity: Because the survey is completely anonymous, responses cannot be associated with any individual person. Once an answer has been submitted, it is therefore no longer technically possible to view, rectify, block, or delete that specific data.
Data Protection Officers
For inquiries regarding data processing and compliance with data protection laws, the following Data Protection Officers are available:
Data Protection Office of Uniklinik RWTH Aachen:
ISiCO GmbH, Cologne Branch
Tel: +49 (0)241 80 89051
Email: datenschutzbeauftragter@ukaachen.de
Technische Universität Dresden:
Data Protection Officer: Jens Syckor
01062 Dresden, Germany
Universitätsklinikum Carl Gustav Carus Dresden
Fetscherstraße 74, attn. Data Protection Officer
Right to Lodge a Complaint with a Supervisory Authority
You also have the right to lodge a complaint with any data protection supervisory authority. A directory of supervisory authorities in Germany can be found at: BfDI Directory of Supervisory Authorities.
The competent supervisory authorities for this study are:
- State Commissioner for Data Protection and Freedom of Information of North Rhine-Westphalia (LDI NRW):
Postfach 20 04 44, 40102 Düsseldorf, Germany • Tel: +49 (0)211 38424-0
- Saxon Commissioner for Data Protection and Transparency:
Postfach 11 01 32, 01330 Dresden, Germany • Tel: +49 (0)351 85471-101
Consent to Participate
"I have received and read the written participant information and declaration of consent for the above-mentioned study. I have been informed in detail in writing about the purpose and procedure of the study, the opportunities and risks of participation, and my rights and obligations."
"My consent to participate in the study is voluntary. I have the right to withdraw my consent at any time without giving reasons and without incurring any disadvantages."
"I hereby consent to participate in the aforementioned study and agree to the described processing of my personal data."
End of Study & Conference Raffle
We thank you for your participation. As part of our study, we are raffling two free registrations for the 20th St. Gallen International Breast Cancer Conference 2027 (SGBCC 2027). If you wish to register for the raffle, please provide your email address at the end of the survey. Your email address will not be stored together with your survey responses, and we will not contact you beyond participation in our survey. The processing and use of personal data for the above-mentioned study will take place exclusively as described in the study information.