Auditing medical AI algorithms under the the lens of medical applicability, bias & fairness, robustness, transparency and trustworthiness

Publications

• Fehr, J., Gunda, R., Siedner, M. J., Hanekom, W., Ndung´u, T., Grant, A., Lippert, C. & Wong, E. B. (2023). CAD4TB software updates: different triaging thresholds require caution by users and regulation by authorities. The International Journal of Tuberculosis and Lung Disease, 27(2), 157–160. https://doi.org/10.5588/IJTLD.22.0437

• Fehr, J., Konigorski, S., Olivier, S., Gunda, R., Surujdeen, A., et al. Computer-aided interpretation of chest radiography reveals the spectrum of tuberculosis in rural South Africa. npj Digit. Med.4, 106 (2021). https://doi.org/10.1038/s41746-021-00471-y   and medRxiv, doi: https://doi.org/10.1101/2020.09.04.20188045.

• Fehr, J., Jaramillo-Gutierrez, G., Oala, L., Gröschel, M. I., Bierwirth, M., Balachandran, P., Werneck-Leite, A., & Lippert, C. Piloting A Survey-Based Assessment of Transparency and Trustworthiness with Three Medical AI Tools. Healthcare, 10(10) (2022). https://doi.org/10.3390/healthcare10101923

• Fehr, J., Piccinnini, M., Kurth, T., Konigorski, S. Assessing the transportability of clinical prediction models for cognitive impairment using causal models medRxiv, doi: https://doi.org/10.1101/2022.03.01.22271617 (2022). BMC Med Res Meth (2023). https://doi.org/10.1186/s12874-023-02003-6

• Oala, L., Fehr, J., Gilli, L., Calderon-Ramirez, S., Li, D. X., et al. ML4H Auditing : From Paper to Practice. Proc. Mach. Learn. Res. for ML4H Workshop at NeurIPS136, 281–317, (2020). Link

• Oala, L., Murchison, A. G., Balachandran, P., Choudhary, S., Fehr, J., et al. Machine Learning for Health : Algorithm Auditing & Quality Control. J. Med. Syst.45, 1–8, doi: 10.1007/s10916-021-01783-y (2021).

• Fehr, J., Konigorski, S., and Lippert, C.. Data Science für Digitale Medizin - Buchkapitel in: Digitale Medizin – Kompendium für Studium und Praxis, Medizinisch Wissenschaftliche Verlagsgesellschaft Berlin, (2020); ISBN-10 3-95466-538-7, ISBN-13 978-3-95466-538-9

• Analysing the applicability of automated chest x-ray reading with CAD4TB to detect tuberculosis in a population screening program in rual South Africa. (Fehr et al. 2021 in npJ Digital Medicine, Fehr et al. IJTLD 2022)
• Demonstrating a quality assessment cycle for trustworthy AI with three exemplary medical AI algorithms: (See Oala et al. 2020&2021)
• Assessing the degree of compliance to transparent reporting and trustworthy AI guidelines (Fehr et al. 2022 in Healthcare)
• Assessing the robustness of machine learning algorithms
  • using a causal framework (Fehr et al. 2022, medRxiv, accepted at BMC Med Res Meth)
  • using synthetic images from generative models. This project is part of the Syreal project (manuscript under preparation)

• 'Syreal' a BMBF-funded consortium project with 7 project partners. Aims to synthesize realistic medical images to mitigate shortcomings in medical AI applications.
• 'Empower', a citizen-centered App to display personal health risks: MSc Project SS2019
• 'Model Zoo', to store and retrieve trained deep learning models for medical applications: MSc Project SS2020

• Tutor & Administration in 'Introduction to Deep Learning' at HPI (Summer semesters 2019-2021)
• Lectures in 'Python for Data Science in Digital Health' block-seminar at HPI on Data visualization and statistics (Winter semesters 2020-ongoing)
• Three lectures in 'Biostatistics II' seminar at Berlin Public school of Health on missing data, and regression models (Summer semester 2022)