Machine Learning for Health: Algorithm Auditing & Quality Control

Luis Oala; Andrew G Murchison; Pradeep Balachandran; Shruti Choudhary; Jana Fehr; Alixandro Werneck Leite; Peter G Goldschmidt; Christian Johner; Elora D M Schörverth; Rose Nakasi; Martin Meyer; Federico Cabitza; Pat Baird; Carolin Prabhu; Eva Weicken; Xiaoxuan Liu; Markus Wenzel; Steffen Vogler; Darlington Akogo; Shada Alsalamah; Emre Kazim; Adriano Koshiyama; Sven Piechottka; Sheena Macpherson; Ian Shadforth; Regina Geierhofer; Christian Matek; Joachim Krois; Bruno Sanguinetti; Matthew Arentz; Pavol Bielik; Saul Calderon-Ramirez; Auss Abbood; Nicolas Langer; Stefan Haufe; Ferath Kherif; Sameer Pujari; Wojciech Samek; Thomas Wiegand

doi:10.1007/s10916-021-01783-y

Machine Learning for Health: Algorithm Auditing & Quality Control

J Med Syst. 2021 Nov 2;45(12):105. doi: 10.1007/s10916-021-01783-y.

Authors

Luis Oala¹, Andrew G Murchison², Pradeep Balachandran³, Shruti Choudhary⁴, Jana Fehr⁵, Alixandro Werneck Leite⁶, Peter G Goldschmidt⁷, Christian Johner⁸, Elora D M Schörverth⁹, Rose Nakasi¹⁰, Martin Meyer¹¹, Federico Cabitza¹², Pat Baird¹³, Carolin Prabhu¹⁴, Eva Weicken⁹, Xiaoxuan Liu¹⁵, Markus Wenzel⁹, Steffen Vogler¹⁶, Darlington Akogo¹⁷, Shada Alsalamah^{18

19}, Emre Kazim²⁰, Adriano Koshiyama²⁰, Sven Piechottka²¹, Sheena Macpherson²², Ian Shadforth²², Regina Geierhofer²³, Christian Matek²⁴, Joachim Krois²⁵, Bruno Sanguinetti²⁶, Matthew Arentz²⁷, Pavol Bielik²⁸, Saul Calderon-Ramirez²⁹, Auss Abbood³⁰, Nicolas Langer³¹, Stefan Haufe³², Ferath Kherif³³, Sameer Pujari¹⁹, Wojciech Samek⁹, Thomas Wiegand⁹

Affiliations

¹ Fraunhofer HHI, Berlin, Germany. luis.oala@hhi.fraunhofer.de.
² Oxford University Hospitals NHS Foundation Trust, Oxford, United Kingdom.
³ Technical Consultant (Digital Health), Thiruvananthapuram, India.
⁴ University of Oxford, Oxford, United Kingdom.
⁵ Hasso-Plattner-Institute of Digital Engineering, Potsdam, Germany.
⁶ Machine Learning Laboratory in Finance and Organizations, Universidade de Brasília, Brasília, Brazil.
⁷ World Development Group Inc, Bethesda, MD, USA.
⁸ Johner Institute, Konstanz, Germany.
⁹ Fraunhofer HHI, Berlin, Germany.
¹⁰ Makerere University, Kampala, Uganda.
¹¹ Siemens Healthineers, Erlangen, Germany.
¹² University of Milano-Bicocca, Milan, Italy.
¹³ Philips, New Kensington, USA.
¹⁴ Office of the Auditor General of Norway, Oslo, Norway.
¹⁵ University Hospitals Birmingham NHS Foundation Trust & Academic Unit of Ophthalmology, Institute of Inflammation and Ageing, College of Medical and Dental Sciences, University of Birmingham, Birmingham, United Kingdom.
¹⁶ Bayer AG, Berlin, Germany.
¹⁷ minoHealth AI Labs, Accra, Ghana.
¹⁸ Information Systems Department, College of Computer and Information Sciences, King Saud University, Riyadh, Saudi Arabia.
¹⁹ Digital Health and Innovation Department, Science Division, World Health Organization, Winterthur, Switzerland.
²⁰ University College London, London, United Kingdom.
²¹ Open Regulatory, Bonn, Germany.
²² MIOTIFY LTD, London, United Kingdom.
²³ IEC TC62 and Siemens Healthineers, Erlangen, Germany.
²⁴ Helmholtz Zentrum München, Neuherberg, Germany.
²⁵ Oral Diagnostics Digital Health Health Services Research, Charité-Universitätsmedizin, Berlin, Germany.
²⁶ Dotphoton AG, Zug, Switzerland.
²⁷ Department of Global Health, University of Washington, Washington, USA.
²⁸ LatticeFlow & ETH Zurich, Zürich, Switzerland.
²⁹ De Montfort University & Instituto Tecnologico de Costa Rica, Cartago, Costa Rica.
³⁰ Robert Koch Institut, Berlin, Germany.
³¹ Department of Psychology, University of Zurich, Zürich, Switzerland.
³² Technische Universität Berlin, Berlin, Germany.
³³ Laboratory for Research in Neuroimaging, Department of Clinical Neuroscience, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland.

Abstract

Developers proposing new machine learning for health (ML4H) tools often pledge to match or even surpass the performance of existing tools, yet the reality is usually more complicated. Reliable deployment of ML4H to the real world is challenging as examples from diabetic retinopathy or Covid-19 screening show. We envision an integrated framework of algorithm auditing and quality control that provides a path towards the effective and reliable application of ML systems in healthcare. In this editorial, we give a summary of ongoing work towards that vision and announce a call for participation to the special issue Machine Learning for Health: Algorithm Auditing & Quality Control in this journal to advance the practice of ML4H auditing.

Keywords: Algorithm; Artificial intelligence; Auditing; Health; Machine learning; Quality control.

Publication types

Editorial

MeSH terms

Algorithms*
Humans
Machine Learning*
Quality Control*

Grants and funding

001/WHO_/World Health Organization/International