Rheumatoid Arthritis of the hand is a disease which affects a significant part of the population. Due to population ageing, it can be expected that rheumatoid arthritis will become an even greater problem in the future. Currently, there are many possible treatments, one of them being physical exercises. Augmented reality, together with computer vision and machine learning open new possibilities for a successful physiotherapy. Turning repetitive tedious exercises into a game and motivating a patient to keep practising at home could lead to a significant improvement of their state. Telerehabilitation systems are already available for many illnesses such as cerebral palsy or back pain, and their results look promising. The aim of this article is to research available rehabilitation systems for rheumatoid arthritis, as well as to present a solution including low cost sensors, serious games and those designed to be used at home or on the way.
Rheumatoid arthritis (RA) is an autoimmune disease which affects about 0,5% of world population. It attacks various joints in the body, typically wrists and knuckles but also knees and feet, and leads to fatigue and ache in the muscles. Further symptoms include swelling, stiffness of the joints, loss of motion and strength and in serious cases joint deformities (fig. 11). The illness complicates daily tasks and can lead to difficulties with employment, psychical discomfort, as well as lower life expectancy on account of further complications such as an accelerated atherosclerosis [1–4]. A typical patient newly diagnosed with RA is a woman in her forties to fifties. Women have a two to four times higher risk of being affected compared to men due to the female sex hormones and their effect on the immune system .
In rheumatoid arthritis white blood cells attack the synovial membrane (fig. 10), a sac which protects a joint and produces a fluid which oils it. The joint becomes inflamed and swelled. Increased blood flow makes the joint warm. The inflammatory cells cause osteoclast generation which erodes the cartilage and bone of the joint leading to its weakening and destruction [1, 2, 4]. The illness affects both sides of the body symmetrically and becomes less aggressive with time . The RA diagnosis is carried out with a rheumatoid factor or sed rate blood test, CRP test or imaging using X-rays, ultrasound or MRI [2, 4].
The therapy focuses on early intervention with the aim of preventing disability and irreversible damage, thus retaining life as normal as possible [2, 6]. Besides anti-rheumatic drugs (DMARDs), medications reducing the cardiovascular risk (statins), injection of radioactive particles (Radiosynovectomy) and surgical treatment, the common therapy method is physical therapy [7, 8].
Physical therapy exercises
Rheumatoid arthritis physical therapy includes cold/heat applications, electrical stimulation and hydrotherapy. Rehabilitation can be carried out using joint protection strategies, massages and exercises accompanied by patient education .
The exercises focus on improving the mobility and strength of the hand and its joints. Whereas some aim to maintain or increase the range of motion (ROM), others strengthen the muscles to provide greater joint support, reducing load and stress on them. Physical exercises also help reduce bone loss caused by inflammation, inactivity and drugs [9, 10]. A crucial aspect of a successful therapy is that the patient develops habits and exercises on a regular daily basis [2, 12]. However, this can be a big challenge because the activities are often dull, repetitive and boring. The patient rapidly looses motivation and does not comply to the programme [11, 12]. The sudden difficulties to perform everyday tasks as well as the experienced pain are a big psychological challenge for RA patients as well. It’s not uncommon for them to become depressive and feel isolated [2, 13]. Therefore, a good therapy should not only include sessions with a therapist but also motivate to exercise at home in order to be effective.
However, already a brief literature research reveals that the effects of physical exercises of RA patients remain unclear. Whereas some works take its benefits for granted [10, 12], other authors point out that the topic is not researched well enough in order to make any conclusions . An extensive review by Williams et al.  states that:
„Exercise probably improves hand function with a minimal clinical benefit in the medium and long term. It is uncertain whether exercise improves pain in the short term, and it probably has little or no difference in the medium and long term. It is uncertain whether exercise improves power grip strength and pinch grip strength in the short term, but shows little or no difference on these outcomes in the medium and long term”.
State of the Art
Serious games for rehabilitation
Low compliance to the therapy programme is a big hurdle for any long-term treatment . Due to their addictive nature, games can be a solution to this problem, turning boring exercises into enjoyable, challenging activities and thus motivating the patients [14, 15]. The use of the so-called serious games in physiotherapy is nothing new and dates back to the 80 s . However, their use in practise has traditionally been limited by the high cost of the VR (virtual reality) systems or the need to carry uncomfortable sensors or attach special markers .
In recent years, the development in the video games industry has brought to the market a number of affordable input devices which has opened a way for research in the field of cheap home-based rehabilitation systems. However, for patients with reduced mobility and strength, commercially available video games are not suitable [12, 16]. Necessary fast interaction, as well as full range of motion represent unreasonably high requirements that may produce strain and anxiety. Besides, most video games are far too complex in order to be understood and enjoyed by elderly population, typical RA patients .
In order to develop an enjoyable yet effective rehabilitation system based on serious games, a set of rules appears to be necessary. As proposed in , rehabilitation game engines should follow two goals – provide all game functionalities, as well as real-time monitoring and advice to the patient. Monitoring of maladaptation and wrong postures are important not to „make rehabilitation more harmful than effective”. Another work states that the principles of a meaningful play and challenge should be followed .
Meaningful play means the player’s choices should have short- and long-term effects on the game. The player receives aural, visual or haptic feedback to signify correct or incorrect action. It is important that the game handles failures in a positive way so that „rehabilitation players are more likely to remain engaged and not feel that failure in the game stems from their impaired physical abilities” .
Challenge signifies that games typically start with a low level of difficulty when the player is unfamiliar with them and increase the level as the player makes progress. Setting the challenge too high or too low can result in either frustration or boredom. Thus, keeping the challenge at an appropriate level is crucial. This can be done by altering the game speed, position of its elements or their size.
A rehabilitation system for RA patients can obtain necessary data based on various principles.
Datagloves are a common input device due to their precision similar to expensive vision systems and a lower cost compared to them . The cited work by Connolly et al. identifies a good fit of the glove and the need for calibration as the main disadvantage limiting datagloves’ precision. Standard calibration procedures require a user to achieve the maximal bent and stretched positions which may be unreachable for RA patients, thus leading to lower precision. The finger position can be detected by IMUs (inertial measuring unit) using a combination of accelerometers, gyroscopes and magnetometers. A good alternative are fiber optics sensors. These make use of varying light intensity in plastic optical fibers which decreases with bending .
Commercial video consoles are also widely used for physiotherapy. Borghese et al.  developed a complex rehabilitation system for stroke patients including therapists, virtual therapists, as well as parameter adaptation and knowledge base. The set of minigames works with various input devices such as Nintendo Wii Balance Board, Microsoft Kinect, 3D touch device Falcon by Novint technologies or specialised balance board Timo plate by Tyromotion. Another work which focuses directly on RA patients was done by Metsis et al. . Here, Microsoft Kinect was used to track full body motion. Commercial video consoles are affordable, however, it would have to be determined whether they provide enough sensitivity to recognise finger joints’ positions correctly. In the case of Kinect, the system is capable of tracking head, legs and arms but hand remains represented as a single point.
Magnetic sensors provide a good alternative. As in the work of Lockery et al. , small sensors can be incorporated into objects of daily use to create a gaming system for local and remote therapy sessions. The patient moves the objects such as a ball, cup or a clothes peg in order to control the mouse movement in the game. This technology seems to be suitable for individualised and task-oriented therapy but not for standardised rehabilitation using serious games.
Accelerometers can be used not only in data gloves, but also in devices called air mouse, such as in the work of Srikesavan et al. . The authors propose a trial focusing on a task-oriented therapy, where the air mouse is attached to objects of daily routine to control standard computer games. As in the case of magnetic sensors, the installation is a problem. It is not a plug-and-play system, so that a home visit of a therapist would probably be needed in order to install it.
Smartphones can also be used since they typically incorporate a variety of sensors, such as accelerometer, magnetometer, gyroscope, GPS and many others . Shin et al.  developed an Android Smartphone App which uses accelerometer and GPS data to play virtual golf. The patient has to swing the arm to hit the golf ball and is then encouraged to walk to make the next stroke and keep fit. The use of smartphones in serious games looks promising as they incorporate a lot of sensors and because they could eliminate the need for special hardware and additional costs. The question remains whether such a system would be easy-to-use for RA patients, often elderly people without previous experience with this technology. Another open question is the precision of such a system.
A force sensor together with a draw wire position sensor was used by Pani et al.  in a custom-built device to monitor home-based therapies of RA patients, specifically having problems with hand joints. Such sensors would be a good option for a special stand-alone device, whose affordability is questionable.
EMG (electromyography) can provide a useful insight into the function of body muscles surrounding the arthritic joints. A paper by Nair et al.  suggests that machine learning can be used to analyse EMG data and to better understand how arthritis affects body muscles, whose reduced function constrains body motion. However, due to its invasive nature, this technique is more relevant for medical research than for a home-based therapy.
Perhaps the most promising is the work of Corona et al.  which developed a set of games for hand therapy of young patients suffering from juvenile idiopatic arthritis. Two near-infrared cameras inside the Leap Motion Controller provide an exact tracking of finger positions. The device is compact, easy-to-use and with a price of about one hundred dollars affordable.
A proposed RA rehabilitation system should be affordable, mobile, easy to use and be precise enough. According to the literature research, an optimal input of such a system could be the Leap Motion Controller, which uses two near-infrared cameras for precise tracking. Its installation is very simple and could be done by patients themselves since it only requires plugging the sensor into a computer.
The manufacturer offers SDK to access the tracking data directly in a C program, but also offers support for Unity and Unreal Game Engines. The SDK includes hand gestures recognition, interaction module for object manipulation and a support for VR headsets such as Oculus Rift or HTC Vive .
Choice of exercises
Virtual rehabilitation system for arthritis patients should be built up upon an established set of exercises. The movements of the hand and fingers should be recognised by the software as gestures and enable controlling the rehabilitation game easily. Literature, such as  and , recommends the RA patients doing following set of exercises.
In the Thumb walk exercise, the patient makes an „O” with their fingers, as seen in the figure 3. After touching the thumb with the finger, one stretches the fingers and continues with the next one.
Another exercise, found in  consists of three movements. First, the patient opens the palm widely. Then, they bend the finger joints to form a „hook”, as seen in figure the 4. Finally, the patient makes a fist and repeats the movements.
An exercises also mentioned in  focuses on stretching the fingers by moving them sideways, as seen in the figure 5.
The last standard exercise to be introduced appears in  and . Its goal is to lift the fingers one by one and leave them lifted for a moment. Then the finger is lowered back and one proceeds with the next one, as seen in the figure 6.
These exercises can be used to control simple serious games tailored for the RA rehabilitation purpose. The following chapter suggests such concepts.
This proposed game incorporates the „palm-hook-fist” exercise shown in the figure 4 and described above. The game uses classic tower defence principle. Troops of soldiers are attacking a castle and a player/patient has to react in order not to let them des...
Pozostałe 70% treści dostępne jest tylko dla Prenumeratorów
- 10 wydań czasopisma "Praktyczna Fizjoterapia i Rehabilitacja"
- Nielimitowany dostęp do całego archiwum czasopisma
- ...i wiele więcej!