AI Breakthrough App
Merlin Fit has just completed a validation study for its AI-assisted physiotherapy pain management app, Merlin Physio, with Queen Mary University of London. The validation study demonstrated a number of very promising results. Merlin Physio achieved full equivalency with a human physiotherapist in guiding accurate frontal plane hip and knee range of motion across all exercises. This is remarkable, given the bio-mechanical complexities involved.
Introduction
The study looked at the potential use of Merlin’s real-time feedback physiotherapy app using artificial intelligence (AI) in providing accurate and effective feedback to patients doing rehabilitative exercises.
The observational study tackled the following question: Does artificial intelligence feedback result in equivalent kinematics during lower-limb exercises compared to physiotherapist-led feedback in asymptomatic individuals?
In other words, can Merlin Fit’s AI-powered physiotherapy system, Merlin Physio, match the quality of feedback provided by human physiotherapists when monitoring patients’ movements and body positioning during commonly prescribed rehabilitation exercises.
As highlighted in the 2007 Australian Journal of Physiotherapy systematic review analysis by Taylor et al., therapeutic exercise plays a vital role in effective rehabilitation protocols and recovery trajectories following musculoskeletal (MSK) injuries. However, a common challenge (raised in the 2009 Sports Health article by Kraemer et al.) lies in the inability for physiotherapists to monitor and correct patient form, technique and adherence when prescribing at-home exercises between supervised clinical sessions.
This is where the Merlin Physio AI mobile application by Merlin Fit, has the capability to bridge this gap. The Merlin Physio App provides patients with real-time audio and visual feedback in their preferred language on a mobile device as they perform prescribed exercises. Merlin Physio enables the real-time correction of subtle flaws in range of motion, posture and movement pace that may undermine goals of therapeutic exercise if left unaddressed over weeks of at-home self-rehabilitation. The App also encourages patients to adhere to a rehabilitation program more readily as it motivates and rewards patients for regular use, although this aspect was not covered by the study.
The Study
The validation study for the Merlin Physio mobile application employed a straightforward methodology to assess the AI-powered feedback system’s equivalency to expert physiotherapist guidance during patient rehab exercises. Researchers recruited a convenience sample of participants from the university population who attended two separate lab sessions. In each session, subjects performed identical sets of four lower limb exercises while receiving visual and audio cues in real-time from either the Merlin Physio app or a physiotherapist with 7 years of professional experience.
Kinematic data was collected across both conditions using a high-precision 3D motion capture system configured based on validated parameters from Lack et al.’s 2014 study. This system tracked key biomechanical metrics including total joint range, the maximum and minimum joint angles of the hip and knee. Analytics compared these key measurable outcomes between the AI-guided and physiotherapist-guided exercise sessions. Reliability was further assessed by calculating intraclass correlation coefficients between the datasets. Equivalency testing determined whether discrepancies fell within a margin that was no greater than measurement error previously reported to be evident between two physiotherapists. These measures validate if Merlin Physio’s autonomous form correction help patients to properly perform rehab routines matching the quality achieved under in-person physiotherapist supervision.
In order to make this determination, the study compared the kinematics – referring to precise motion paths and body alignments – achieved by asymptomatic test subjects when receiving corrective feedback either from an AI-assisted device employing motion sensors, video feeds, and analytical software or directly from a licensed physiotherapist observing their movements.
The Findings
By evaluating the equivalence of the resulting kinematics under therapist-led versus AI-led conditions, the following observations were made regarding the readiness of Merlin Physio’s AI technologies to automate and augment elements of physiotherapy treatment, and in future increase patient access and enhance outcomes, while upholding the standards of care delivered under expert human oversight.
The validation study demonstrated a number of very promising results. Merlin Physio achieved full equivalency with a human physiotherapist in guiding accurate frontal plane hip and knee range of motion across all exercises. This is remarkable, given the bio-mechanical complexities involved.
The 50% equivalence scores for range across the other two planes at the hip and knee are also impressive first benchmarks that showcase Merlin Physio’s ability to offer personalized, precision guidance when compared with expert practitioners.
Considering natural variability in the ways people move, the consistency and reliability reflected in the intraclass correlation coefficients further underscores the system’s technical capabilities to adapt to individual needs.
In presenting the findings, the research team also observed that patients inherently perform exercises differently. So to have high degrees of measurement equivalency and repeatability between Merlin Physio and experienced physiotherapists, despite such real-world conditions, is a resounding validation of this AI-assisted application’s readiness to augment and standardize tele-rehabilitation.
Finally, during and since the study, the Merlin Physio App has been updated and tweaked to improve start and end data points for patients helping to further improve feedback performance.
Further research is now underway with Pure Sports Medicine clinics in London, to assess the generalizability of such findings to symptomatic patients undergoing rehabilitation in real-world clinical settings. Nonetheless, this comparative study offers valuable insights on the capability of Merlin Physio’s AI-assisted physiotherapy app to support physiotherapy and pain management in this vital healthcare arena.
Merlin Physio
Merlin Physio allows physiotherapists to assign personalized exercise regimens through the app and remotely track patient engagement levels while allowing AI capabilities to handle the continuous, objective form assessments between clinic visits. Merlin Physio thereby aims to amplify clinical capacity and foster patient accountability.
Intelligent rehabilitation solutions like Merlin Physio can improve outcomes and access for MSK patients undergoing essential exercise therapies, and broader adoption appears imminent. This has the potential to shift the supervision and success of at-home physiotherapy science grounded in data-driven, individualized feedback accessible to all.