BARRIERS TO THE IMPLEMENTATION OF ARTIFICIAL INTELLIGENCE IN REHABILITATION: A CROSS-SECTIONAL SURVEY AMONG PHYSICAL THERAPISTS IN PESHAWAR, PAKISTAN

Background: Artificial intelligence (AI) has emerged as a transformative tool in healthcare, offering significant opportunities for enhancing diagnosis, treatment planning, and patient monitoring in rehabilitation. By enabling predictive analytics, clinical decision support, and improved efficiency, AI holds promise in addressing rehabilitation challenges. However, despite these advantages, its adoption among physiotherapists remains minimal, particularly in low- and middle-income countries where infrastructure, education, and organizational readiness are limited. Understanding the current level of awareness, readiness, and barriers is essential to guide future implementation strategies. Objective: The study aimed to assess physiotherapists’ knowledge, attitudes, and perceived barriers toward AI adoption in rehabilitation in Peshawar, Pakistan. Methods: A descriptive cross-sectional survey was conducted between January and March 2025 among 200 practicing physiotherapists working in hospitals, clinics, and academic institutions in Peshawar. Eligibility criteria included at least one year of clinical experience. Data were collected using a structured, self-administered questionnaire distributed electronically through emails and WhatsApp groups. The instrument captured demographics, knowledge and awareness of AI, perceived barriers across five domains (educational, organizational, technical, ethical, and financial), and readiness to adopt AI. Data were analyzed using SPSS version 27, applying descriptive statistics including means, standard deviations, frequencies, and percentages. Chi-square and Fisher’s exact tests were performed to explore associations between barriers and demographics, with a significance threshold of p < 0.05. Results: Out of 200 participants, 120 (60%) were male and 80 (40%) female. Clinical experience ranged from 1–5 years (35%), 6–10 years (40%), and >10 years (25%). Regarding awareness, 48 (24%) had no knowledge of AI in rehabilitation, 74 (37%) reported theoretical knowledge only, 66 (33%) demonstrated basic awareness, and just 12 (6%) had hands-on use in practice. Barriers reported included limited clinician knowledge (156; 78%), lack of institutional support (130; 65%), data privacy concerns (120; 60%), high implementation costs (110; 55%), and insufficient training programs (85; 42.5%). Experience level was significantly associated with educational and organizational barriers (p=0.03), and private sector clinicians reported higher technical barriers compared with public sector counterparts (p=0.04). Conclusion: The findings demonstrate that while awareness of AI exists among physiotherapists in Peshawar, practical application is severely constrained by multidimensional barriers. These include inadequate training, insufficient institutional support, technical limitations, ethical concerns, and high costs. Targeted strategies such as integrating AI literacy into curricula, strengthening institutional frameworks, investing in infrastructure, and developing cost-effective, context-appropriate solutions are vital to promote adoption. Enhancing clinician readiness through structured training and organizational policies could bridge the gap between awareness and effective utilization in rehabilitation practice.