File
M. Six Dijkstra1,2, E. Siebrand3, S. Dorrestijn4, E.L. Salomons5, M.F. Reneman2, F.G. Oosterveld1, R. Soer1,6, D.P. Gross7, H.J. Bieleman1
1Saxion University of Applied Sciences, School of Health, Enschede, Netherlands, 2University of Groningen, Department of Rehabilitation Medicine, University Medical Center Groningen, Groningen, Netherlands, 3Saxion University of Applied Sciences, Research Group Ethics and Technology, Enschede, Netherlands, 4Saxion Univeristy of Applied Sciences, Research Group Ethics and Technology, Enschede, Netherlands, 5Saxion University of Applied Sciences, School of Ambient Intelligence, Enschede, Netherlands, 6University of Groningen, University Medical Centre Groningen, Pain Centre, Groningen, Netherlands, 7University of Alberta, Department of Physical Therapy, Alberta, Canada
Background: Computer algorithms and Machine Learning (ML) will be integrated into clinical decision support within physical therapy. This will change the interaction between therapists and their clients, with unknown consequences.
Purpose: The aim of this study was to explore ethical considerations and potential consequences of using ML based decision support tools (DSTs). We used an example in the context of occupational physical therapy.
Methods: We conducted an ethical deliberation. This was supported by a narrative literature review of publications about ML and DSTs in occupational health and by an assessment of the potential impact of ML-DSTs according to frameworks from medical ethics and philosophy of technology. We introduce a hypothetical clinical scenario in occupational physical therapy to reflect on biomedical ethical principles: respect for autonomy, beneficence, non-maleficence and justice. The reflection was guided by the Product Impact Tool, an assessment tool in ethics and philosophy of technology.
Results: Respect for autonomy is affected by uncertainty about what future consequences the worker is consenting to as a result of the fluctuating nature of ML-DSTs and validity evidence used to inform the worker. A beneficent advisory process is influenced because the three elements of evidence based practice are affected through use of a ML-DST. The principle of non-maleficence is challenged by the balance between group-level benefits and individual harm, the vulnerability of the worker in the occupational context, and the possibility of function creep (the gradual widening of the use of a technology or system beyond the purpose for which it was originally intended). Justice might be empowered when the ML-DST is valid, but profiling and discrimination are potential risks.
Conclusion(s): Implications of ethical considerations have been described for the socially responsible design of ML-DSTs. Three recommendations were provided to minimize undesirable adverse effects of the development and implementation of ML-DSTs. The recommendations include:
1) encourage multidisciplinary involvement in developing and training ML-DSTs for responsible and valid design with minimized risks;
2) formal assessment by Health Research Ethical Committees for potential risks regarding privacy and custodian issues as well as discrimination and function creep; and
3) training physiotherapists in methods, validity and scope of ML-DSTs to support evidence based practice.
1) encourage multidisciplinary involvement in developing and training ML-DSTs for responsible and valid design with minimized risks;
2) formal assessment by Health Research Ethical Committees for potential risks regarding privacy and custodian issues as well as discrimination and function creep; and
3) training physiotherapists in methods, validity and scope of ML-DSTs to support evidence based practice.
Implications: On a micro-level, ethical considerations and knowledge of ML-DTSs has been implemented into the educational program of physical therapists and information technology students in Enschede, Netherlands. On the meso-level, the implications are being considered by a multidisciplinary team developing an ML-DST in occupational health. On the macro-level, the study is internationally published to emphasize the importance of ethical aspects for the responsible design, education of health care providers, and use of ML-DSTs in physical therapy and other disciplines.
Funding, acknowledgements: Funding: Netherlands Organisation for Scientific Research (NWO) (023.011.076) and Saxion University of Applied Sciences in The Netherlands.
Keywords: Algorithms, Ethics, Evidence based practice
Topic: Innovative technology: information management, big data and artificial intelligence
Did this work require ethics approval? No
Institution: University Medical Center Groningen
Committee: Ethics Board at the University Medical Center Groningen in The Netherlands
Reason: This study is part of a PhD project. In this particular study no human subjects were involved.
All authors, affiliations and abstracts have been published as submitted.
