Perception of integration of Artificial Intelligence into physiotherapy practice among Physical Therapists of Pakistan

The purpose of this study was to assess the knowledge and attitudes of PTs regarding the use of AI in rehabilitation, focusing on variables such as years of experience, subspecialties, work setting (academic vs. clinical), and exposure to AI technologies.

Data was collected via a four-part survey from 256 PTs with over six months of clinical experience. The survey included:

1. Demographics: Age, gender, years of experience, education, location, and work setting.
2. Knowledge of AI: Familiarity with AI, exposure to AI technologies, and training.
3. Attitudes toward AI: Views on AI’s potential to reduce workload, improve patient care, enhance diagnostics, and aid in rehabilitation.
4. Barriers and Ethical Concerns: Challenges such as lack of training, complexity, and ethical concerns.

Questions included a mix of Likert scales, multiple-choice, and open-ended formats, with a content validity index (CVI) of 0.8.

The findings revealed that 67% respondents either strongly agreed or agreed about positive impact of AI in rehabilitation, 18% remained neutral, while 15% disagreed.

Among those who agreed, the majority (78%) had only heard of AI tools like PhysioAI, Physitrack, while 22% had used these or some other tools in their clinical practice atleast once. 

In terms of experience, 72% PTs with less than five years of experience were aware of AI applications, compared to 58% of those with more than five years of experience. 

Additionally, 64% respondents acknowledged AI’s potential to reduce workload, while 70% agreed it could enhance patient care. However, 29% expressed concerns about understanding AI’s complexity, highlighting the need for further training and integration into curricula. Despite these concerns, a large majority (89.6%) expressed a willingness to receive more information about AI, indicating openness to learning.

While 67% respondents recognize AI's potential in enhancing rehabilitation, a concerning 29% reported inadequate training on its applications in their physiotherapy education. Additionally, 70% emphasized the urgent need for more research about AI usage in practice. These findings underscore a critical gap in education and training, highlighting the necessity for further studies to harness AI's potential in physical therapy.

While this study indicates a readiness among PTs to adapt AI and awareness of AI’s potential impact on physiotherapy, there are significant gaps between scientific evidence and its practical application in clinical settings. There is a need for robust clinical trials to validate AI tools in real-world scenarios, and the necessity for integrating AI technologies into physical therapy education and training programs. Future research should focus on addressing these gaps by establishing the necessary infrastructure and curricula and conducting longitudinal studies to assess the effectiveness of AI-driven interventions. Moreover, potential barriers, such as cost of AI technology, resistance to change, and the complexity of AI systems, need to be thoroughly examined and mitigated to facilitate smoother adoption in clinical practice.