Should vet rehab professionals use AI to objectively measure gait: International Association of Veterinary Rehabilitation and Physical Therapy

Should Vet Rehab Professionals use AI to objectively measure gait?
Objective measurement of gait parameters is becoming more commonplace in the veterinary industry due to the known levels of reliability both within and between practitioners observing gait [1]. The set-up costs and ease of use of technologies has allowed application of devices in clinical practice and not just in research laboratories or veterinary hospitals. Formerly technologies to measure metrics such as head and pelvic asymmetries, stride length or limb joint range of motion were restricted to lab use because of the equipment required, how it needed to be applied and the complexity of the analysis. However, the once inaccessible has now become available to those working ‘in the field’.
The advance in modern technology means that high-frame rate video cameras are in our pockets, inertial motion sensors are the size of coins and marker-less video analysis by AI can be used to quantify motion asymmetries. The costs and time required to set up and use these devices have reduced considerably and so they now have the potential to be used in therapy and rehab practice, in addition to their use in veterinary examinations.
Use of objective measures, such as technologies to measure gait symmetry, are recommended to improve reliability by removing the human, subjective element to evaluation. Observer bias is inherent and a risk to the reliability in recording measurements and thus the truth of the effects of treatment [2]. Irrespective of the profession or pathology being assessed, measurement of factors that are influenced by the condition and subsequent intervention should be used [3]. If measurement technology such as inertial motion sensors or marker-less artificial intelligence systems are available, these can be used as they can classify asymmetry, limit observer bias and remove the limitations of the human eye to detect mild lameness [4].
Whilst innovation and technology are exciting and suggest evolution in the ability to be objective, the use by non-veterinary professionals could be considered controversial and beyond scope of practice. Training to apply or use the devices is not a challenge but the interpretation of the data collected is more so. If this interpretation is believed to identify the nature of the problem, then it could be defined as a diagnosis, however if the output is a single measure and not specifically identifying an underlying pathology this arguably is not a diagnosis. Musculoskeletal practitioners working with animals are limited in their scope of practice by their knowledge, skills and experience as well as by law in some countries [5]. Whilst not permitted to diagnosed, the measurement of movement and function a core competency within the profession.
How data driven discussions can be used within a multidisciplinary team, focused on the best interests of the animal patient is a key consideration. The use of objective assessment is critical to being able to evaluate practice and evidence effect. Understanding where the ethical and professional ‘line’ is, what the data can and can’t tell you, the relevance of AI for Vet Rehab Professionals and how to use the information gained, will be the focus of this talk.