Strategies
Successfully integrating AI into physiotherapy practice requires thoughtful approaches that maximise the benefits while mitigating potential risks. The strategies suggested here focus on developing practitioner competencies, establishing ethical frameworks, maintaining patient-centred care, and creating appropriate regulatory guidance. By implementing comprehensive approaches to the adoption of generative AI, the profession can make use of the advantages of this technology while preserving the human expertise and therapeutic relationships that remain central to effective care.
The following strategies provide an early foundation for responsibly incorporating generative AI tools into clinical practice while aiming to address the challenges they introduce.
Comprehensive AI education
Implementing AI literacy programmes is essential for ensuring physiotherapists can effectively and safely use AI tools in practice (Annapureddy et al., 2024).
Develop comprehensive training programmes covering both technical competencies and critical evaluation skills for AI outputs
Structure regular policy and guidance updates to keep practitioners current with evolving AI technologies and their clinical applications
Incorporate practical case studies demonstrating appropriate AI use in clinical decision-making scenarios
Maintain ongoing education about relevant legal frameworks and professional regulations around AI use
We should note that the term "AI literacy" is itself contested and comes laden with it's own challenges, which are beyond the scope of this document to address. It is therefore incumbent on us as a community to determine what "AI literacy" even means in the context of the profession, as well as how to implement and evaluate it among stakeholders.
Adaptive regulation
Regulatory frameworks must evolve alongside AI technology to ensure safe and effective implementation.
Establish 'regulatory sandboxes' for testing new AI applications in controlled clinical settings
Create mechanisms for regular dialogue between clinicians, regulators, service-users, and AI developers
Develop flexible policy and regulatory frameworks that can adapt to rapid technological changes
Implement systematic monitoring and evaluation of AI implementation outcomes
We have positioned this section on regulation early on, to highlight the importance of engagement with regulators, whose structures, policies, and guidance typically inform practice. We think the biggest obstacle in this context are the relatively slow institutional mechanisms we rely on to create policy and regulation.
Ethical guidelines and best practices
Clear frameworks for ethical AI use in physiotherapy are needed to ensure responsible implementation.
Establish comprehensive guidelines addressing patient privacy, data security, and informed consent
Create detailed protocols for balancing AI-assisted care with traditional therapeutic approaches
Implement regular review processes to update guidelines as AI technology evolves
Develop clear standards for documenting and reviewing AI-assisted clinical reasoning and decision-making processes
Promote responsible AI use that considers environmental and economic sustainability through resource efficiency and equitable access
Transparency and explainability
Ensuring that AI systems are transparent and understandable is crucial for building trust and effective use (Chaddad et al., 2023).
Collaborate with AI developers to create systems that provide clear rationales for their suggestions i.e. aim for explainable AI (however, this may not be achievable within existing generative AI paradigms)
Establish requirements for AI systems to explain their decision-making processes in clinical terms
Implement protocols for documenting and reviewing AI-assisted decision processes
Develop standards for communicating AI use and limitations to patients
Human-AI collaboration model
Establishing effective collaboration between practitioners and AI systems is crucial for optimal outcomes.
Train physiotherapists to leverage AI as a supportive tool while maintaining independent clinical judgement
Encourage systematic evaluation of AI recommendations against clinical expertise and patient context
Develop frameworks for integrating AI insights with traditional clinical reasoning processes
Create clear protocols for when to leverage AI assistance versus human expertise
Patient-centred approach
Maintaining focus on patient needs and preferences is essential, especially in the context of patients making use of their own personal generative AI models, to inform health choices.
Create structured protocols for discussing AI use with patients and obtaining informed consent
Develop clear communication strategies about how clinicians are using AI tools to support their clinical practice
Establish processes for incorporating patient preferences into AI-assisted care planning
Implement regular feedback mechanisms to assess patient comfort and their priorities are met with AI integration
Discussion questions on the strategies for integrating AI into clinical practice
What specific AI literacy skills do you think are most crucial for physiotherapists to develop?
How can we effectively balance the use of AI tools with the maintenance of essential human skills in physiotherapy practice?
What key points should be included in ethical guidelines for AI use in physiotherapy clinical practice?
How can we ensure that AI systems used in physiotherapy are sufficiently transparent and explainable?
In what ways can we involve patients in the process of integrating AI into physiotherapy practice?
How can we promote the sustainable use of AI in physiotherapy, considering both environmental impact and equitable access to technology?
How can physiotherapists best prepare for the regulatory requirements of AI implementation, particularly in light of frameworks like the EU AI Act?
What strategies can we develop to measure and evaluate the effectiveness of AI integration in physiotherapy practice?
How might the role of clinical mentorship evolve as AI becomes more integrated into physiotherapy practice?
What safeguards should be established to prevent over-reliance on AI tools while maintaining clinical reasoning skills?
Last updated