Advances in modern medicine and better quality of life have resulted in higher life expectancies in the developed world. The population aged 65 years and over is projected to grow by around 50% over the next 23 years in the UK [1
]. The rural areas, in particular, have seen a larger increase in average age than urban areas [2
]. Additionally, internal migration statistics have shown an increase in population moving into rural locations [3
]. Such population migration trends present a number of challenges for the future, in particular regarding access to modern health services.
Abnormal gait and locomotion deformities can be a result of various causes. Ageing, disease, and accidental injuries are all common factors. The probability of a gait-related issue increases by 10% between ages 60–69 and by 60% over the age of 80 years [4
]. Observation of locomotion is a complex study involving internal and external factors, which can result in a variety of positional changes [5
]. Typically, the diagnostic triage of such injuries and pathologies requires the physical presence of the subject for orthopaedic and locomotor examination. Vital clues for diagnosis can be obtained from visual observations by trained professionals. Indications such as swellings and bruising can ultimately aid in determining the cause and result in a diagnosis. However, long-term patients with locomotion-affected pathologies, acute and chronic pathologies, such as osteoarthritis can be required to visit clinics routinely.
This could involve such patients having their gait assessed over a period of time. Such gait-related observation utilises the use of a motion-capture system to record the patient’s movement to be analysed in the specialised motion analysis software. Gait analysis facilities enable the acquisition of high-quality, quantifiable biomechanics data that are analysed objectively [7
]. However, such facilities are very limited, are mostly situated in major cities, and require substantial space and costs to operate. Furthermore, concurrent global financial issues, limited resources, and the constantly growing population have prompted interest from various governing bodies to reform and seek alternative methods for delivering high-quality healthcare [11
]. The advancements on sensors, 3D visualisation and the increase of communication innovations have made telemedicine a promising solution for many issues faced by the current healthcare systems [13
]. To this end, telemedicine-enabled VR gait analysis systems could offer a viable alternative, as presented in previous preliminary studies [15
Adhering to the above, the aim of this work is to identify a precise and financially viable rehabilitation alternative, which could be deployed in a small footprint room (i.e., an average house bedroom). As such, we propose a novel system that utilises advanced 3D/VR visualization, motion tracking sensors, and remote data analysis for the acquisition, presentation, and analysis of motion data derived by the patient. As two of the major restrictions are the cost and space of current rehabilitation facilities, our system employs cost-effective motion-capture, built from off-the-shelf equipment, to be utilised with our immersive 3D virtual gait analysis clinic. The latter enables the collection of data related to biomechanics independently. In turn, the data are mapped on to a unique 3D anatomical data set, which is then visualised in the 3D virtual gait analysis clinic application. This paper presents a pilot study designed to gauge subjective responses and user experience feedback by ten participants, which used the prototype system in this experimental form.
The following section presents the musculoskeletal issues that require gait analysis and rehabilitation systems. In turn, the paper discusses the contemporary state of the art facilities and the issues that require an alternative approach to the solution. Technological advances and equipment are also discussed in detail. Section 4
presents the proposed system solution, whilst Section 5
presents the evaluation process and the questionnaire used to acquire the subjective feedback form the participants. The paper concludes with the overview of the work undertaken and the results derived during the evaluation. Finally, the paper presents a tentative plan of work based on the received feedback and recorded users’ experience of the proposed VR gait analysis system.
4. Proposed System Solution
Our system was developed through an iterative and incremental development process, which enabled the developing team to build and test the software application in multiple stages, aiming to eliminate major usability and functional issues. During system development, a number of specialist medical doctors were consulted for the accuracy of the system, regarding both the data collection capacity as well as the correct presentation of human anatomical and pathological data. Throughout this process, the project’s aim was to enable health professionals and clinicians to interact efficiently and in a timely manner with the virtual dataset.
Additionally, the system was designed to offer a plethora of tools to the health professionals, to confidently analyse and monitor patients’ gait idiosyncrasies and their rehabilitation progress. As such, the system could present in real-time the patient’s motion in a VR environment and present simultaneously the related motion graphs. The latter could highlight instantly the “peak to peak” variations in the patient’s motion pattern and enable the medical practitioner to adjust future rehabilitation exercises. The VR /3D model could be manipulated by both the patient and the medical practitioner in a 3D space, and it could be rotated or panned in order to present the best view of human body motion. The complete system is presented in action, capturing and analyzing a user’s walking cycle, in Figure 4
For the particular prototype system, it was deemed useful to employ a set of first-generation Microsoft Kinect sensors system, purely in order to prove the point that the hardware should not be a limitation or the major cost of such systems. A combination of Microsoft Kinect sensors was positioned diagonally in a room (3 × 4 m). The rationale behind constraining the experimentation area was in order to resemble the typical size of a bedroom or a small office in a house environment.
5. System Requirements
The demand for motion capturing systems is greater than ever, with a range of industries understanding the potential of such systems and its capabilities. Motion capture systems have surfaced into many cross-platform disciplines.
As previously mentioned, motion capture setups have been expensive, required specialised costly equipment, required professional setup, trained professionals, and a dedicated large space.
Rapid technological evolvement, specifically in the last decade, has rendered current large-scale production motion capture systems inexpensive (approx. £800), and it has enabled the use of marker-less motion capture, which negates the need for reflective markers on the subject and is easily accessible (utilising off-the-shelf sensors).
Whether an off-the-shelf motion-sensor device (i.e., Kinect (Microsoft ©), PS eye (Sony ©), Xtion (Asus ©) could be utilised to obtain data that can be of diagnostic value, to the clinician or the accuracy of their results, in comparison to the traditional marker-based Mo-cap system for gait analysis had not been tested prior to our work.
Game engines are of particular interest, in contrast, to fully custom developed engines. During the exploration of existing systems and game engines regarding cost efficiency, community support, existing libraries, and ease of development and use, this work deemed that the Unity 3D engine complied with most the aforementioned characteristics. As such, the particular game-engine was employed to build upon the prototype software application. Furthermore, the particular engine allowed the development of customizable tools, and it was compatible with various types of motion capture data whilst being able to visualise high-quality 3D data in real-time. Additionally, this is a cross-platform engine, which enables the proposed software application to be easily modified to operate on a host of different platforms (Windows, Apple, and Android) and devices (computers, phones, and tablets).
5.3. 3D/VR High Definition, Accurate Anatomical Models
Another major consideration for the development of a VR gait analysis application is not only the acquisition of data but also their presentation to both patients and medical practitioners. The photorealistic and accurate presentation of human anatomy forms a common ground for communication for both groups, as this was revealed by previous studies [42
]. As such, the accuracy of the provided data both in motion and visual representation could enhance further the patient/user experience and convey in a more understandable manner the complex medical issues that might hinder the patient’s mobility and/or progress of rehabilitation.
For this study, we opted for an anatomical data set provided by Core Lab + Ltd. [46
]. The highly detailed and accurate models offer real-time photorealistic medical 3D visualisation. Core Lab + 3D models are constructed and verified by specialised medical practitioners from relevant fields such as orthopaedic surgeons, radiologists, and general practitioners. The models’ 3D meshes were further optimized for smooth real-time interaction and for use in different operating systems for VR applications, as presented in Figure 5
The particular models consisted of more than 300 different anatomical elements including muscles, ligaments, fasciae, blood vessels, nerves, lymphatic drainage, and bones, which were developed to move and follow the leg and body motion. As such, these models enabled a photorealistic representation of a patient’s gait analysis movements, which could highlight clearly the section and part of the leg that operated sub-optimally and affected overall motion.
The different layers of the human anatomy models could be removed or applied in real-time so as to inform the patient further regarding the different anatomical issues that he/she was encountering. This process was particularly useful as the patient was educated to understand their personal medical condition and their own circumstances. By understanding the rehabilitation rationale, the patient could further improve the exercise regime. At the same time, the motion data and 3D visualization offered the health professionals an instant overview of the patient’s progress.
The aforementioned three elements were utilised in the development of the proposed VR gait analysis system, as presented in the following section.
6. System Evaluation and User Feedback
The proposed system was designed and implemented following the strict requirements of the contemporary, high-fidelity, rehabilitation systems currently used in the clinical environment. In order to ensure the proposed VR systems’ compatibility with the National Health Service (NHS) requirements, the system was also tested against the high-end motion capture system available at the School of Health and Life Sciences (SHLS) at Glasgow Caledonian University (GCU). The data were retrieved simultaneously, with both systems capturing the same locomotion data, and were almost identical and within the required limits.
The study acquired the relevant ethical approval from the University’s Ethics Committee. In order to achieve this, a number of experiment areas were prepared meticulously, and they followed all the current NHS guidelines for musculoskeletal rehabilitation. The process was also explained step by step to the participants, and their written and signed consent was acquired prior to the experiment.
This paper presents results of a pilot study for the user evaluation trials, designed to identify the users’ technology acceptance levels. Notably, the main points of interest were user views regarding the utilization of VR for rehabilitation and the perceived ability to deploy such system in small spaces such as typical rooms in a home, hospital, or general practitioners’ surgery. In addition, this work was interested in revealing any additional benefits or drawbacks that could potentially surface through the evaluation and the users’ subjective feedback. The user experience and the level of technology acceptance received from this experiment would further inform this project and other future endeavours towards developing more user-friendly and cost-efficient rehabilitation systems.
For this evaluation, ten participants took part in this pilot study, 80% male, 20% female. A range of age groups was included in an attempt to minimise age bias. Fifty per cent of the participants were 25 to 29 years old. Thirty per cent of the participants were 45 years of age and over. Weight and height were also measured for each participant. The heaviest participant was 133 kg and the lightest was 57 kg. The tallest person was 1.95 m and the shortest participant in the study was 1.64 m.
Regardless of the height, weight, and different body types, the Mo-cap data of all the participants were collected and successfully processed into the prototype system.
It has to be noted that this paper presents a pilot study designed to identify the users’ intentions towards this type of technology and rehabilitation method. Consequently, the ten participants’ evaluations do not yield any statistical information, however, it provides an informative appraisal of the users’ expectations of the proposed system and the system’s performance.
The subjective feedback provided by the patients was recorded in two phases, namely, in the pre-test and post-test questionnaires. The following sections present the results indicative of proposed and future system requirements.
6.1. Pre-Test Results of Users’ Subjective Feedback
This questionnaire was carried out prior to the patients being exposed to the proposed telemedicine system, before the data acquisition process (Mo-cap), presentation, and 3D demo. The objective was to try and get information from the patients without influencing their answers with the proposed research.
The first set of questions targeted the concept of telemedicine. The objective was to determine how familiar the participants were to the practice and if they had utilised it in the past. Findings indicated that even though 50% of the participants were not familiar with the term telemedicine, approximately 90% had used some form of telemedicine service previously. This shows a lack of familiarity with the term “telemedicine”. Data shows that the most common form of telemedicine service used previously by the participants was NHS 24/Direct, which is an out-of-hours service; ninety per cent of the participants had used it previously. The NHS’s symptom checker website was reported to have been used by 70% of the participants; again, this service is also considered to be an out-of-hours service for primary care. Only 20% of the participants had used the telecare monitoring services, which were related to other long-term pathologies. None of the participants had used any telemedicine service related to their gait pathology or injury. Although the patients had utilised telemedicine services, they did not know they were utilising a telemedicine system. The objective of this question was to help the patients make a connection to what they had used in the past and what the undertaken research was proposing. All the patients had used some sort of telemedicine-based service in the past; none of the services was used for their pathology related to gait-related injury or locomotion. By asking the participants about their previous experience of using telemedicine, the concept of practical telemedicine was explained, making it simpler for the participants to understand what this research was proposing. The next set of questions was designed to acquire information about the patient’s history related to their pathology, appointment waiting time, and how frequently the patient visited health professionals regarding their pathology.
The most common waiting time for an initial consultation amongst the participants was 12 weeks; this was mostly to see a specialist health professional regarding their pathology. After their initial appointment, 70% of the patients were visiting the health professionals regularly every 1 to 4 weeks for the course of their treatment (Figure 6
The next set of questions examined the length of time patients had experienced gait symptoms and explored the level of patients’ familiarity with their pathology and pathology-related terminology.
According to the results, 70% of the participants experienced gait-related problems for over six years, whereas for the remainder 30% the participant’s gait abnormalities had persisted for two to five years (Figure 6
On questioning patients on their understanding of their pathology, 40% graded their understanding as moderate, 30% as light, and only 20% of the patients felt they understood their pathology extremely well (Figure 7
(left). Participants were further asked on their understanding of the associated pathology-related terminology. Seventy per cent of the participants felt they were not at all, slightly, or only moderately familiar (Figure 7
(right)) with such terminology. On further analysis, it became apparent that the only patients that had subjectively graded themselves as extremely familiar with their pathology or terminology were from medical and health-related educational backgrounds. Overall results highlighted that although the majority of the participants had gait-related problems for a long period of time, their understanding of the underlying pathological process or relevant terminology was limited. It was also evident that current interactions with the health professionals were limited in their educational value regarding the conditions under treatment
To this end, the next set of questions aimed to establish which techniques had been previously utilised to aid the patient’s understanding of their pathology or injury, and which aspects of the information provided by the health professionals did the patients find difficult to comprehend.
Health professionals utilised verbal consultations as the main means of information sharing, and it was the main means experienced by 80% of the participants. The second most common means of conveying information was “images”, and 40% of patients had some experience of this in clinical practice. Few of the participants had also experienced the use of “props” such as a skeleton model in order to visualise and understand their pathology.
3D animation or 3D visualisation techniques had not been experienced by any of the study participants in the past as aids in their understanding the pathological changes of their condition or injury (Figure 8
Participants were further questioned whether they had difficulties during their consultations with understanding medical information. Findings revealed that 80% of the participants found understanding complicated terminology problematic, with difficulties encountered in visualising their pathology and/or injury, and a further 50% were challenged with the visualization of the underlying anatomy (Figure 8
It is of interest that despite 40% of the participants subjectively claiming that they understood their pathology moderately well, 80% admitted to having difficulties visualising their pathology.
The next section presents post-test user feedback, which further examines the users’ intentions towards the use of emerging technologies.
6.2. Post-Test Results of Users’ Subjective Feedback
After the data acquisition process (Mo-cap), presentation, and 3D demonstration, the users completed a post-test questionnaire. The objective was to obtain their views in regards to a telemedicine-based musculoskeletal diagnosis service, the prototype system, the data acquisition process (Mo-cap), and the method of 3D visualisation as a means of communicating patient information regarding underlying pathology.
The first set of questions explored patient views regarding the telemedicine-based musculoskeletal diagnosis service, including the prototype system that was demonstrated and presented to the participants. Participants were questioned on their views regarding the proposed system and whether they would prefer to use it (Figure 9
Regardless of factors such as gender, age, familiarity with telemedicine, and duration of gait-related problems, findings revealed that 100% of the participants either “agreed” or “strongly agreed” that they would utilise the 3D virtual reality telemedicine service if it was implemented at their local general practitioner clinic or health center, whilst 80% either “disagreed” or “strongly disagreed” that they would prefer to utilise their current method of receiving health care rather than using this telemedicine service. These results indicated that participants had a very positive attitude towards the proposed telemedicine service. Again, 80% of the participants indicated a preference for utilizing the system independently rather than attending at a facility (Figure 9
The following questions aimed to identify which aspects of the system were considered most and least appealing. Findings revealed that with an average rating of 2.2, the most appealing factor in system implementation was the potential for regular gait assessment, followed with an average rating of 2.3 for the prospect of regular feedback regarding their progression. Results interestingly also indicated that the least appealing factor was time and money, with an average rating of 3.9 (Figure 10
Participants were further questioned on who in their view such a system would benefit the most. Eighty per cent of the participants thought that such a system’s implementation would benefit patients living in remote areas the most. Participants thought it would be least beneficial for patients who routinely did not attend appointments ((Figure 10
The utilisation of such a system directly from home was explored, with participants asked to indicate the most and least concerning aspect of such an endeavour (Figure 10
(right)). Participants were mostly concerned with the learning curve associated with being able to utilise the system properly (60%), whilst the majority of participants indicated that the least concerning statement was “lack of trust in new technology and its accuracy” (80%). Overall findings suggest that participants were willing to trust new technologies that would help them assess their gait more regularly.
Participants were asked to respond to a number of statements regarding their views on their current method of receiving healthcare and their views on the new proposed telemedicine system. Findings from these results were very positive towards the proposed telemedicine system (Figure 11
(left)). This was based on average ratings, with a value of 1 being “strongly disagree” and a value of 5 being “strongly agree”. With the highest average rating of 4.7, 70% of the participants strongly agreed that there were more advantages than disadvantages to the new proposed system. Most of the participants “strongly disagreed” that they would carry on using their current method of receiving healthcare as they are used to the routine (70%).
Furthermore, the results showed that 100% of the participants either agreed or strongly agreed with the following statements:
“I would prefer to use the 3D virtual reality system as I wouldn’t have to travel as much if it was implemented at my Local GP’s or Health Centre.”
“I would prefer to use the 3D virtual reality system because I feel there are more advantages to the proposed system than disadvantages.”
“I would prefer to use the 3D virtual reality system because I feel the proposed system would benefit me more in comparison to the current method.”
Additional interesting findings from this data included that between 80–100% of the participants either “disagreed” or “strongly disagreed” with the following statements:
“I would rather carry on with my current method of receiving health care because I am used to the routine.”
“I would rather carry on with my current method of receiving health care because I don’t feel my GP or Health Centre has the specialist training required to record my gait.”
“I would rather carry on with my current method of receiving health care as I feel the system would benefit the health professionals more than me.”
These findings illustrated that all participants indicated a preference for the use of the proposed system in comparison to their current method. The next set of questions focused more on the aspects of gait assessment and the methodology (Mo-cap) employed by the proposed system.
The participants were asked about their previous experience regarding gait analysis methods. Some (60%) of the participants previously had their gait recorded. Furthermore, participants were asked about the gait assessment methods previously used to assess their gait. By far the most common (90%) was a visual observational assessment method by the health professionals. Just under 40% used markerless and marker-based methods in the past (Figure 11
(right)). Thus, it was of interest to note that, although most of the participants had gait-related problems for over six years, nearly half of them had never had their gait recorded. Additionally, for the vast majority of participants, only observational assessments were utilised for gait analysis. As per our results, 90% of the participants considered our proposed system more useful in their gait analysis than any other system they had utilised (Figure 12
(left)). Participants were also presented with statements regarding the motion capturing process. The result showed that 100% of the participants either “agreed” or “strongly agreed” with the following statements:
None of the participants found it difficult taking part in the motion capturing process. The majority of the participants liked the fact that they did not have to walk any significant distance while having their gait assessed (see Figure 12
(right)). After the data acquisition process (Mo-cap), presentation, and 3D demonstration, the users completed a post-test questionnaire. The objective was to obtain their views in regards to a telemedicine-based musculoskeletal diagnosis service, the prototype system, the data acquisition process (Mo-cap), and the method of 3D visualisation as a means of communicating patient information regarding underlying pathology.
The following sets of questions regarded the 3D VR tool (proposed system and osteoarthritis tool). Participants were asked questions regarding their thoughts on how helpful they thought the 3D VR tool was to them. The findings were extremely positive towards the 3D visualisation tools (Figure 13
The result showed that 100% of the participants thought the VR tool (proposed system) would let them visualise their gait on a model, and help them understand what was happening with their locomotion. Additionally, 100% of the participants also thought that with VR tools (visualization regarding stages for osteoarthritis), they could visualise different stages of the pathology being simulated, and felt that a similar simulation for their own pathology could benefit them to prepare for the future. Almost 100% of the participants either “agreed” or “strongly agreed” that the VR tool (proposed system) would help enable them to name and visualise all anatomical structures related to their lower limb, and that it would increase their understanding of their pathology, which would be hugely beneficial for them.
All the participants answered yes for the following questions:
“Do you think such 3D visualisation tools, would be helpful for you to understand your pathology?” (Figure 14
“Do you think such 3D visualisation tools could help you understand your underlying anatomy?” (Figure 14
“In comparison to other methods used to help you understand about your pathology, do you feel the 3D medical visualisation method shown today was better?” (Figure 15
The participants were additionally asked which specific aspect of the VR tools they would utilise the most. It was evident that participants wanted to use the system mostly for mapping their own gait on the 3D anatomical dataset for visualisation and to identify terminology and parts of anatomy connected to their pathology so they can understand health professionals and increase their overall disease understanding. (Figure 14
Through the evaluation process, it was highlighted by the participants that the technological advances used by the proposed system provided an efficient, informative, and enjoyable process. One of the major aspects was the ability to scan their gait motion in close proximity with discreet equipment. This was in direct contrast to the intimidating plethora of equipment typically hanging from the ceiling gantry of large-scale facilities and the multiple cameras/sensors staged across a long-distance corridor.
Another major aspect that was tackled with the use of the proposed system and the technologies employed was the markerless MO-CAP capability of the sensors. This was again in divergence to the white/reflective small spheres attached to patient joints before gait analysis and rehabilitation sessions. These markers tend to require precise positioning and fasting of any loose clothing, whilst the human factor of positioning the spheres could result in distorted results. This was emphasized as a major annoyance issue by the patients, who in some cases were unwilling to go through the process. The markerless technology embedded in the proposed system does not require any spheres or physical object or interaction between the staff and the patient. Furthermore, because the whole body is scanned in every step by the sensors, the points of interest for each joint are calculated by human motion and not by any additional artefact attachment. As such, the results are not affected by human errors in positioning the tracking spheres.
A third technological advancement that proved equally useful and was appreciated by the users was the VR visualization and presentation of the patient’s gait motion and analysis. This was deemed by the users as an essential part of any future gait analysis and rehabilitation system, as it helps in educating and informing the patients about their anatomical underlying issues and the progress of their rehabilitation.
7. Conclusions and Future Work
This paper presented current issues regarding gait analysis in the clinical setting and the significance of its use in further understanding and rehabilitating various pathologies. Additionally, it explored limitations of the current systems with a view to map the design requirements for a prototype telemedicine-based asynchronous virtual reality gait analysis clinic. The prototype system was tested for result accuracy and fidelity in comparison to current clinically used methods with promising outcomes. Further development concentrated on the system’s evaluation with regards to user experience, as the latter can present significant challenges in the incorporation of new, emerging technologies on the existing health provisions/services. To this end, pre- and post-test questionnaires were developed in order to acquire user/patient views towards technology acceptance of the proposed system. In turn, data analysis was performed with an indicative group of 10 users who were all experiencing or had experienced typical rehabilitation methods and gait analyses offered by public hospitals and/or private clinical facilities. As this was a pilot study, the acquired data were not statistically significant, yet provided an informative appraisal of the system’s performance towards user experience and technology acceptance.
Overall, the users’ experience presented positive results and evident preference for the adoption of the proposed prototype system within the National Health Service provision. The technological elements and advances employed, in conjunction with the bespoke developed software, could enable a cost-efficient and patient-accessible method of gait analysis and rehabilitation. The participants’ feedback highlights the simplicity and efficiency of the system in contrast to the users’ experience of the current state of the art facilities that are time-consuming and difficult to use. These positive user responses further informed future research plans for the development of a bespoke hardware and software VR system that could be deployed and operated directly from the patient’s home whilst monitored by the healthcare professionals.
Concluding, the overall experiment offered an intriguing insight into the requirements of future rehabilitation applications. To this end, it was evident that patients were welcoming the acquisition of medical information regarding their pathology. Yet, this should be presented in a simplified yet visually enticing manner. This type of information could further support the patients’ consent process and empower the patients to understand the methods and the required outcomes of their rehabilitation.