Responsible Translation

Trustworthiness, Robustness, Usefulness and Transparency in Data- and Algorithm-driven Biomedical Research: A Meta-Research Approach

Background

The QUEST Center aims to develop and implement new approaches to ensure that biomedical research is conducted in a trustworthy manner, produces useful results, and meets ethical standards. This project addresses the area of data- and algorithm-driven biomedical research, which is primarily based on techniques of artificial intelligence (AI), especially machine learning (ML). There is currently a lack of work examining the topics of trustworthiness/robustness and usefulness/transparency in this research area applying meta-research techniques. Applying meta-research, we aim to develop overarching recommendations for optimizing data- and algorithm-driven biomedical research.

Objectives

The objective of this project is to do conceptual/methodological development work to determine the state of trustworthiness/robustness and usefulness/transparency in data- and algorithm-driven research, develop concrete research questions for international relevant meta-research projects, and conduct them. Additionally, an interview study with relevant stakeholders will be carried out to evaluate their views on trustworthiness/robustness and usefulness/transparency and the meta-research results. In the final step, specific recommendations (best practices) will be developed.

Methodology

The project will begin with narrative reviews to identify the gaps in the current research on trustworthiness/robustness and usefulness/transparency in data- and algorithm-driven biomedical research. Then, two meta-research projects (systematic reviews) will be conducted, with the first focusing on the area of trustworthiness/robustness and the other on usefulness/transparency. An interview study with relevant stakeholders will also be conducted, and their views on the results of the systematic reviews will be evaluated. Finally, specific recommendations (best practices) will be developed. The project will begin with narrative reviews to identify the gaps in the current research on trustworthiness/robustness and usefulness/transparency in data- and algorithm-driven biomedical research. Then, two meta-research projects (systematic reviews) will be conducted, with the first focusing on the area of trustworthiness/robustness or usefulness/transparency. An interview study with relevant stakeholders will also be conducted, and their views on the results of the systematic reviews will be evaluated. Finally, specific recommendations (best practices) will be developed.

Expected Results / Implications / Perspectives

This project is expected to generate concrete best practices for trustworthiness/robustness and usefulness/transparency in data- and algorithm-driven biomedical research. The recommendations developed in this project will enable biomedical researchers to conduct their research in a more trustworthy manner, produce useful results, and meet ethical standards. In addition, this project will also provide perspectives on how to optimize research practices in the context of data- and algorithm-driven biomedical research.