Robot-Assisted Broad Consent Collection

Initial Experience in Practice

As digital technologies are increasingly becoming integrated into medicine, the hope is that routine data can be used for research to an increasing degree. Since 2018 the Medical Informatics Initiative (MII)—funded by the German Federal Ministry of Education and Research (BMBF, since 6 May 2025 the Federal Ministry of Research, Technology and Space, BMFTR)—has worked on establishing decentralized data infrastructures in Germany, to make it possible for healthcare data to find their way into research faster and in a more standardized form (1). In the initial phase of the MII, the focus was on setting up data integration centers at all university hospitals in order to make routine clinical data available for research purposes in a structured and standardized form. In the current expansion phase of the MII, the focus is on the expanded data exchange between teaching hospitals and ne partners. One central prerequisite is patients’ consent to the collection and use of clinical data in medical research. To this end, standardized patient information and broad consent were established in the context of the MII, which enables the nationwide collection of pseudonymized clinical data in accordance with the EU General Data Protection Regulation (GDPR) (2).

Integrating broad consent into routine clinical practice does, however, often present departments with challenges in terms of staffing and organizational issues. In addition to oral information delivered by trained staff, further means may be employed in order to convey the sometimes complex topics in patient consent in a comprehensible form. In addition to printed information materials, a comprehensible information film was made (3). The department of General Internal Medicine and Psychosomatics at Heidelberg University Hospital was the first to implement robot-assisted broad consent collection. In this article we will explain the approach and present an evaluation of the pilot phase.

Methods

From October 2024 to January 2025, the Department of General Internal Medicine and Psychosomatics at Heidelberg University Hospital piloted the use of a service robot for collecting MII broad consent (3). The project used a telepresence and navigation robot (“Temi”, Robotemi Ltd, New York, USA, distributed in Germany by Roboterly GmbH, Olpe, Germany). A web-based management platform makes it possible to plan in advance sequences of actions (for example, autonomous movement, voice interaction, provision of visual media) for the service robot, which can be retrieved by clinical staff as needed. The cost of purchasing the robot is in the four-digit range. Data transfer is based on a WLAN connection; for reasons of data protection the service robot is not integrated into the Hospital Information System. The Figure displays in detail the process of collecting robot-assisted broad consent. In the setting of a naturalistic study design, some of the patients who presented to our department’s irritable bowel clinic as outpatients in the recruitment phase were consecutively included in the study. Patients were selected on the basis of staff availability and capacity among the responsible colleagues. In total, 78 patients who were able to give consent were included in the study; 60 of these additionally completed an anonymized satisfaction questionnaire. As the data collection was completely anonymized, and as we collected neither sociodemographic nor disease-related data, no ethics approval was required according to the current regulations.

Figure

Process of robot-assisted broad consent collection

Results

74 of 78 patients (97.4%) agreed to this procedure of collecting broad consent. The satisfaction survey showed that information provided by the service robot was entirely positively accepted by the patients. The mean satisfaction score was highest for comprehensibility (mean=9.1 of a maximum of 10, standard deviation [SD]=2) and for lowest for wellbeing during the interaction with the robot (mean=7.5; SD=2.5). Satisfaction received a mean score of 8.6 (SD=1.6) and participants’ willingness to recommend the approach to others received a mean score of 8.4 (SD=1.9). Patients’ suggested improvements concerned lacking subtitles during the explanation of the process by the service robot (N=2), difficulties when the volume changed (N=3), and the speed at which the robot spoke (N=1). None of the patients requested additional oral information from the clinic staff. The time taken by the hospital staff for providing information, which was captured on the basis of self-reports, was reduced from about 10 minutes per patient (4) to about 5 minutes, mainly for the purpose of documentation.

Discussion

The results of the pilot of robot-assisted broad consent collection indicate good feasibility, acceptance, and approval rates. Compared with earlier results from other clinical settings, participants’ degree of informedness and willingness to consent seem at least comparable if not even better (4). Participating patients on average rated the procedure very positively. The reduction in staff expenditure and time supports the use of service robots in collecting broad consent and could contribute to improving the nationwide implementation rate, which has been low to date. Because of the smaller group size and the sampling bias caused by the selection of patients (5), however, the results should be considered as preliminary. Our department is planning to continue the study and to include all patients in order to further evaluate the procedure.

The comprehensive use of routine clinical data for research purposes has the potential to gain new insights on how to optimize healthcare. The required framework conditions that are needed to this end are, however, tied not only to complex data protection regulations but also to organizational and staff-related demands that are often difficult to implement. Our preliminary results indicate that using AI-based service robots enables a reduction in staff-related expenditure and time while simultaneously ensuring high rates of approval and acceptance among patients. In the long term, using such technologies may contribute to improving efficiency in collecting data and to preserve resources for direct patient care.

Erklärfilm: Die Patienteneinwilligung der Medizininformatik-Initiative des BMBF

Joe J. Simon, Stephan C. Feder, Katrin Meyer, Hans-Christoph Friederich, Mechthild Hartmann

Conflict of interest statement
The authors declare that no conflict of interest exists.

Manuscript received on 11 April 2025, revised version accepted on 26 June 2025.

Translated from the original German by Birte Twisselmann, PhD.

Cite this as
Simon JJ, Feder SC, Meyer K, Friederich HC, Hartmann M: Robot-assisted broad consent collection: Initial experience in practice. Dtsch Arztebl Int 2025; 122: 564–5. DOI: 10.3238/arztebl.m2025.0120

References

1. Semler SC, Boeker M, Eils R, et al.: Die Medizininformatik-Initiative im Überblick–Aufbau einer Gesundheitsforschungsdateninfrastruktur in Deutschland. Bundesgesundheitsbl 2024; 67: 616–2

2. ZZenker S, Strech D, Ihrig K, et al.: Data protection-compliant broad ­consent for secondary use of health care data and human biosamples for (bio)medical research: Towards a new German national standard. J ­Biomed Inform 2022; 131: 104096

3. Medizininformatikinitiative – Mustertext zur Patienteneinwilligung 2025. Verfügbar unter: www.medizininformatik-initiative.de/de/mustertext-zur-patienteneinwilligung

4. Fischer-Rosinsky A, Eienbroker L, Mockel M, et al.: Broad consent in the emergency department: a cross sectional study. Arch Public Health 2025; 83: 44

5. Moser K, Bauch F, Richter M, et al.: Bias in obtaining broad consent in a German general practice? – Preliminary results from a cross-sectional study. J Family Med Prim Care 2024; 13: 4056–65