Consideration of requirements of different interest groups
Tense contrasts exist between promises, hopes or predicted failures on the one hand, and technical realities on the other hand. The fact that current myoelectric technology lost the Arm Prosthesis Race of the CYBATHLON 2016 against body-powered technology contrasts with high hopes and promises going with the new “bionic” hands [
7,
42]. The fact that body-powered prosthetic split hooks can be powerful prosthetic aids contrasts with the problem that they are vilified [
32,
118,
119]. These contrasts affect various interest groups that relate to prosthetic arms differently.
UBEA often find that for ADL or light work, the stump is the best prosthesis [
120]. In the age group 2–20 years, UBEA without prosthetic arm outperformed both wearers of prosthetic arms as well as people without disability for ADL across freely distributed bimanual tasks [
121,
122]. Prosthetic arms are not of proven value to help psychosocial adjustments [
123]. Moreover, arm amputees may regard not wearing a prosthetic arm as part of affirming a public image of different ability [
124] particularly in the light of social pressure. The prevalent non-usage of prosthetic arms may be the best functional, economical, proudest and thus rational choice for ADL and light to moderate work [
10].
Users that expose themselves to their devices may end up as the ultimate experts [
4,
118]. They try to get their consumers’ complaints to be taken seriously, but there may be powerful social and neurological mechanisms that prevent this [
125]. The bare arm amputee risks to upset others visually so much [
119,
126], that expensive gadgets have now shifted towards the center of a sociological demarcation process [
127]. Thereby, societal mechanisms exert a strong push towards amputees to stereotypically cover their stigma [
126,
128]. Conversely, the few amputees that do feel personally concerned by that push may offer to comply with that request by exclusively accepting expensive or futuristic-looking rather than functional technology [
129‐
131]. Within that discourse, raw mechanical functionality risks to deteriorate from being a core property to being, at best, a superficial label, while affinity-driven product ratings [
132] may risk to distort public perception of their advertised (but not actual) technical performance. In a further twist of society attributing stereotypes, amputees wearing “bionic” hands risk to be perceived as “cold” and as “high-tech”, and thus as a social threat [
133].
Families of amputees or prosthetic technicians have assumptions regarding the role of prosthetic arms that differ from those of amputees [
123], as do engineers [
134]. Current prosthetic arm research and development mainly focuses on myoelectric [
118,
135,
136] technology and, more recently, 3D-printing [
137]. If nothing else, these devices are marketed to conform to the requirement of a social standard of costly modern technology [
32]. Myoelectric and 3D-printed arms are thus assumed to support at least light work or ADL. But only 23% of the users rated the weight of a myoelectric hand as acceptable [
138]. Only 12% of the male users found the noise of their myoelectric hand to be not disturbing [
138]. Usage of myoelectric arm was indicated most often for using cutlery (76% of men), handicrafts and even opening/closing doors (71%) [
138].
A more definite role for myoelectric arms to play particularly in UBEAs’ lives may thus depend on what real needs this new technology manages to cover [
128,
136,
139]. However, the list of known issues relating to current myoelectric arms, remains long. It contains electrode related skin rashes [
98,
99], sweat interference with electrode functioning [
84], postural interference [
140], high weight and distal center of gravity, insufficient durability [
47], noisy distraction [
141], absent proprioceptive feedback [
142], uncoordinated grips [
93], fragile prosthetic gloves [
143], extreme costs [
144] and unattractive appearance [
45,
145].
Arm amputees with PDW to deliver are far more constrained regarding the choice of their prosthetic arm build, controls or components: they will more likely have a vital need for prosthetic arms that function, also under harsh conditions. Body-powered arms also dominate the market of prosthetic arms that are indispensable for PDW as well as sports [
3,
44,
146]. For work with occupational heat exposure, biological or chemical hazards, large weights or widely ranging ambient temperatures, there is no other technology. Development of body-powered technology currently is only conducted by a small number of individuals and groups (e.g., Randall Alley [
147,
148], Bob Radocy [
40,
149], Bradley Veatch [
74,
150,
151], Dick Plettenburg’s group [
152‐
156], Aaron Dollar’s group [
157] and John Sensinger’s group [
158]).
According to our results, even some of the current commercially available body-powered components are nowhere near sufficient for PDW as outlined here. When facing such a situation as a consumer, discarding the faulty product is a far more likely reaction than trying to fix it, which can be very difficult [
159]. We worked on two fronts for that: we tried to optimize both body-powered and myoelectric technology, both within the available options.
Approaching solutions for a PDW workplace from a general development and research position, one will consider that most of the hazardous, intense, sweaty or manually challenging work aspects cannot be changed [
160,
161]. Also in the future, decomposing, heavy and slippery bodies will be found, also in narrow confined spaces, also of messy premises. Also for years to come, lifting, retrieving, turning, undressing and examining bodies in such situations will remain strenuous and require tough, light weight, durable prosthetic arm components with high tensile and compressive strengths for the experts that perform these duties. Occupational tasks of this specialized and individual nature will require concessions and compromises also concerning posture [
160]. One may have to work out regularly to achieve and maintain fitness for such work [
162]. Long and drawn out death scene examinations when wearing protective gear have aspects of “mini-expedition” style missions: one goes in, then one is in there under full strain, with executive and manual challenges and responsibilities, without any easy option to exit or troubleshoot, until only hours later, when that mission is over. And so there are other instances where equipment has to conform to harsh occupational requirements, and equipment specifications seem not too different: for large expeditions, reducing weight, improving performance and extending longevity of equipment can attain game changing significance [
163]. So, research and development has proven, elsewhere, that it can understand and integrate such concepts outside the circle of amputee problems.
Narrowing technical options
An ideal mission-critical design [
164] – as a necessary property for a prosthetic arm – will deliver reliable and largely error-free performance that at least approximates industrial quality standards as well as delivering performance across the specified exposure. A conformant prosthetic arm is built to minimize ill side effects, bodily injury or damage. It is built with a modular design that allows fast user repairs with widely available and affordable materials. It offers protection from overuse in the light of bodily asymmetry and heavy bi-manual work [
17,
19‐
22].
Studies that discuss prosthetic use and overuse never normalize or stratify for actual work exposure, prosthetic arm proficiency for intense work, and actually delivered manual work. In our case, a supportive prosthetic arm allowed to perform hard work at the same functional level as peers, whereas a wrong design would cause severe shoulder pains after 1 day of regular typing work.
Mission-critical design requirements are not met by some of the current prosthetic parts that we encountered. Clinically relevant side-effects are a reason to reconsider design aspects of a prosthetic arm once lesions take too long to heal or when they risk causing permanent damage. Sudden or erratic failure while wearing a prosthetic arm can be a dramatic and stressful event; this is remedied by pushing a system to exhibit graceful degradation, which gives the user time to intervene.
Body-powered prosthetic arms are very intuitive to use. But actual motor skills including fine motor skills are acquired only by sufficiently specific and sufficiently extensive training [
165,
166]. To no surprise, absent proficiency of large shoulder and trunk muscles to perform fine grasps with a body-powered control in untrained non-amputees causes their control attempts to deteriorate at higher pinch forces in a study that makes a great case for training [
156]. Also, absent sufficient specific training appeared to be the reason of fatigue in most non-using amputees when trying out body-powered arms, whereas the only actual daily user of a body-powered arm in that case series did not exhibit any significant restriction (study subject number seven [
167]). The first user of this study had therefore been advised by his physiotherapists early on, to not just try out body-powered technology, but to really wear it for a few years. Ultimately, large arm, shoulder and trunk musculature may be trained for heavy lifting and subsequent fine control even more efficiently than hand muscles [
168]. Conversely, electric motors or batteries may simply be dead weight for a UBEA that delivers PDW over years and that has sufficiently extensive and sufficiently specific strength to provide forceful body-powered grips.
Dermatological side-effects of prosthetic arms
Friction rashes are a frequent side-effect of wearing a prosthesis [
72]. Conventionally, polyurethane or silicone liners are worn directly on the skin. When sweat disrupts close liner adherence to the skin, the sweat soaked outer layers of the skin will easily abrade and develop a rash or blisters, as early as after a few hours. It may take days for a rash or blisters to heal, during which the prosthesis should not be worn. Tight cotton is known to effectively treat ’acne mechanica’ in soccer players [
169]. We employed tight tubular gauze to be worn under a gel liner. It interfaces with the skin through micro-compression by way of many tiny fabric strands. These swell up to a degree as sweat fills up the cotton, while the outer skin layers remain relatively dry [
170]. With a body-powered arm, the socket does not contain electrodes that sit on the skin and provide ridges where soaked soft skin layers risk to get abraded. So protection from friction rashes can allow for far greater exposure under sweating with a body-powered arm.
Skin burns are not uncommon to develop in the vicinity of myoelectrodes [
98,
99]. Here and under our observation, these lesions came about under moderate amounts of sweat that had not acutely disrupted myoelectric control and took about four to six weeks to heal. As described elsewhere, we also observed blister configurations as part of these burns. The underlying technical aspects of these burns appear to also affect implanted electrodes [
171]. Furthermore, heavy sweating would disrupt myoelectric control as early as 10 min into PDW [
84]. Research into non-electric modes of control of devices as so far yielded both subcutaneous [
78] as well as surface shape [
172,
173] derived control signals as viable alternatives, at least from an academic research angle. From a PDW aspect, too much equipment is not a practical option [
174]. With regard to skin preservation under PDW conditions, we found that body-powered suspensions could be coerced to conform best.
Typing contains its own perils. A long duration of repetitive small stroke actions can be hazardous, so even small differences in weight amount to large effects at the end of a day. Myoelectrodes’ ridge structures pressing into the skin caused a significant friction rash and large blisters, just after one day in the office with typing work. The socket will experience larger repetitive motions also due to a higher myoelectric terminal device weight. A tightly fitted body-powered configuration with a light aluminum split hook performs with less amplitude and less momentum. This is the case particularly with deadline work and long hours of writing [
175].
Sudden failure rather than graceful degradation
Graceful degradation of performance even under adverse conditions is essential for mission-critical reliability [
164]. Research and development will have to address this aspect consciously.
A predictable grip geometry is required for efficient forward-planning of dynamic push-release or reach-grasp trajectories. A multi-articulated hand that lacks finger tip coordination cannot guarantee a reliably repeatable grip configuration [
176]. Lack of geometry control invariably will cause grip failure that may surprise the user, causing “sudden” or at least unexpected problems on a functional level, as seen at the CYBATHLON 2016, where a rigid gripper with just two claws outperformed some of the demonstrated multi-articulated hands due to this problem [
7,
177]. Plannable grips so far benefit from rigid or constrained grip geometries. This to a degree may explain the various split hooks’ models success within amputees [
41‐
44,
146]. The design of multi-articulated hands could possibly be improved, as researchers have identified and understood this problem [
93].
Posture or stump position may negatively interfere with myoelectric control. Even professional training levels and trained controlled circumstances cannot prevent sudden occurrences of this phenomenon [
7]. Typical myoelectric control uses two electrodes to control a single degree of freedom. They are placed on the flexor and extensor locations with best signal-to-noise ratio. Incidentally, these locations typically contain muscles that are also activated during elbow flexion, extension, or during stump pronation or supination, regardless whether the user intends to open or close the myoelectric device. Flexing the elbow, standing up or changing the position of the torso while keeping the hand in a constant position (which will entail elbow extension or flexion) or other changes in the limb position risk to trigger unintended signals [
91]. Different stump positions are also known to interfere with multi electrode control [
140]. This problem results from employing intrinsically polyvalent muscle groups for single function controls [
90]. Especially when the user is distracted, and during dynamic work, this can drive up myoelectric performance error rates fast. While body-powered arms exploit posture of elbow, shoulders and back to directly transmit their shape change to achieve an analog cable tension actuation, myoelectric arms exploit polyvalent forearm muscles for digital single function control in UBEA.
It is thus fair to say that myoelectric arms are or can be also, to a degree, body-powered [
178]. The art consists in making that a wilful and consciously controlled act. With that, there are two distinct differences to proper body-powered control. In body-powered arms, cable tension is built up gradually, and there is considerable proprioception of the analog control state, to a degree where body-powered VC devices can be used to precisely vary grip power from very subtle [
27] all the way to over 200N. Myoelectric arms lack an analog proprioception across any control range. Secondly, the muscles used for body-powered control allow for a relatively intuitive separation of gripper actuation versus limb position change. As a key property of the control system, it results that body-powered control degrades far more gracefully when changing limb or body position. The user always feels the cable tension. While it is a training paradigm that myoelectric arms allow for precise and fluid motions [
179], we found that controlled stop-and-go procedures can be more effective to prevent the limb position effect.
With both myoelectric and body-powered systems following bodily motions, both can be thus used in a freestyle way, or ’tricked’. One useful posture trick, given conventional myoelectric systems, is for the user to not at all move the stump, elbow or shoulder while performing critical grip maneuvers. An elevated shoulder and stiff elbow in an attempt to avoid posture effects will eventually cause overuse symptoms on the shoulder and neck of the amputated side, but may be relatively efficient when carrying valuable items [
178]. Another useful trick, for both body-powered and myoelectric controls, is to switch off or let go of the prosthetic actuation entirely, to avoid any postural interference with the gripper.
This has been the solution for the winner during the hot wire loop test at the CYBATHLON 2016 [
7]: the pilot locked down his body-powered VC system’s control cable [
180] before he started with the hot wire test. He was then free to focus on the loop position fully. He only unlocked the cable afterward. The other competitors did not appear to have visibly incorporated that body-powered aspect into their myoelectric race strategy [
8].
Immediacy and option to manually intervene in real time, at every step of a manipulation, is far easier with body-powered arms. Being in full control over one’s own work pace is a key factor in successfully delivering PDW [
181]. Manual overrides or visual signals could be added to myoelectric devices with little extra weight. Overall, due to a very intimate link between cable tension, proprioception and terminal device actuation, we found that a body-powered control was always far more reliable than a myoelectric system.
Grip quality and grip strength
Soft covers of grip devices are a relevant issue [
111]: in the presence of friction, form closure of any object places less emphasis on the grip geometry (gripper shape, number of fingers or claws). There exists a negative relationship between softness and longevity of a gripper surface [
182]. The softer the surface, the firmer an object may be held even at low grip forces, but the more frequently it decays and needs to be replaced. Then, user accessibility and very affordable materials become a critical issue.
For prosthetic hands, soft covers are typically gloves. The durability of gloves is important; it was mentioned as a relevant factor already in 1980 [
143]. The constraints that exist are manifold: Firstly, manufacturers of prosthetic hands make narrow specifications for allowed gloves. Secondly, gloves mechanically impede actuation [
183], so weak prosthetic hands are equipped with thin and fragile gloves. Thirdly, perforating damage usually calls for an immediate stop to usage as gloves protect the hand from dirt or fluid. With myoelectric hands being rather weak and heavy already, hand geometries deviate from a normal human hand in efforts to maximize efficient grip geometry. That again makes it hard or impossible to fit these hands with normal gloves that fit normal human anatomical hands. The softer the glove, the better the grip but the faster it is damaged [
182] and needs replacement. Humanly proportioned gloves are mass produced at a wide range of makes and qualities for relatively low prices. Any terminal device that works without these constraints is at a clear advantage.
There is one adaptively gripping very precise and robust prosthetic hand that excels there. The body-powered Becker hand [
109,
110,
184] is a very affordable, robust body-powered hand with a reliable precision grip and an adaptive grip, that fits standard gloves including regular work gloves, including those sold at convenience or hardware stores. Its mechanical design is technically very evolved. It is not clear why the 3d-printing community, that claims to search for affordable durable solutions with respect to prosthetic hands, has not identified that hand as an answer to their quest.
Equipping a standard metal split hook with silicone tubing or cutting sheet rubber to fit a V2P or TRS prehensor device is fast, supported by warranty regulations, and easy to perform for the user.