Secondary Logo

Journal Logo


Development of an Inexpensive Upper-Extremity Prosthesis for Use in Developing Countries

Sitek, Alison J. BS, MS; Yamaguchi, Gary T. BS MS PhD; Herring, Donald E. BA BS MS; Willems, Christopher J. BSIE, MBA; Boninger, David BA, MA, PhD; Boninger, Ronald M. BS, MBA

Author Information
JPO Journal of Prosthetics and Orthotics: July 2004 - Volume 16 - Issue 3 - p 94-102
  • Free

The upper-limb amputee population in developing countries is growing at an alarming rate. Estimates indicate the number of amputees needing prostheses in India alone increases by 17,000 annually. 1 Although there are many causes of amputation, including firearms, war, farm and auto accidents, and disease, farm machinery accidents and war are the largest contributors in many developing countries. 2 Countries undergoing internal turmoil tend to experience large increases in their amputee populations because of combat, land mines, and war atrocities. For instance, rebels in Sierra Leone inflicted hand amputations and mutilations on hundreds of civilians during that country’s civil war (J. Egan, unpublished data, 2001). This unrest has led to incidence rates of 1 amputee per 650 citizens in Afghanistan, 1 per 400 in Cambodia, and 1 per 350 in Angola, compared with 1 per 22,000 in the United States (J. Egan, unpublished data, 2001).

This is a problem at both the national and individual level. The majority of the population in most developing countries depends on manual labor for employment. When a significant percentage of the population lacks the capacity for physical labor, the nation experiences decreased productivity and slowed economic development. Personally, the amputee is at a disadvantage in the job market without a prosthesis. Families are strongly affected, especially if the amputee is the main wage earner. Some even turn to begging to help provide for themselves and their family. 3

Although necessary for income, prostheses are expensive, and Third World amputees cannot afford to purchase them, let alone maintain them. In the United States, a body-powered, above-elbow prosthesis can cost upward of US $15,000. Even a prosthesis manufactured by an aid organization such as the International Committee of the Red Cross (ICRC) can cost US $1,000 to fit and fabricate. 3 Cost aside, many prosthetic limbs manufactured for use in the United States are not suited for developing countries. The intense humidity of many Asian and equatorial countries, coupled with prolonged exposure to dirt and water, quickly wears down complex components, such as the standard locking elbow. 4 The terminal device control cable often fails during repetitive, heavy tasks (R. Kistenberg, personal communication, 2000). Finally, most prostheses currently in use in the United States do not facilitate activities typically seen in other cultures, such as the use of chopsticks.

In view of the large population and variety of complicating factors, there is a need for a prosthetic arm specifically designed to be inexpensive and adaptable for amputees in developing countries. The arm was designed with the following goals: functionality at a cost less than or equal to US $100, durability, simplicity, ease of repair, adaptability to local materials, and cultural acceptability. An inexpensively designed prosthesis could facilitate a Third World amputee’s return to gainful employment, aiding his family, and boosting his country’s economic growth.



In the fall of 2000, a team comprised of members from Arizona State University’s (ASU) School of Design and Department of Bioengineering tackled the challenge of building a prosthetic arm specifically for use in developing countries. A survey was sent to prosthetists working in the Third World to ascertain any special needs. One key response indicated the frustration with actuation cables, which frequently break during repetitive, heavy use and are difficult to repair (R. Kistenberg, personal communication, 2000). Drawing from this information, the team developed several concepts for prosthetic arms that do not rely upon cables, including terminal devices that are not cable actuated (Figure 1). Other key concepts, such as a polyethylene “living hinge,” eliminated a dependence upon specialized parts, which often wear quickly in gritty environments and are difficult to replace.

Figure 1.
Figure 1.:
Prototype upper extremity prosthesis design concepts produced by Human Factors Systems and Documentation students. Prototypes are described from top to bottom. (a) A neoprene elbow cuff, PVC hinge and components, and a “window-shade” cord lock. Cords to hold elbow position were found to be unacceptable when the elbow was elevated. (b) Bamboo upper and forearm components joined at the elbow with a section of rubber fuel line hose. The elbow was mounted to the waist via a belt, and the forearm was actuated by means of the cuff attached to the stump. (c) Starting from an extended position, the 45 degree rotating elbow could produce simultaneous elbow flexion and pronation. A pin was used to lock the elbow in any angular position. The socket was made from ABS piping, and the terminal device was made from the hood latch of a car. (d) Wooden prototype living hinge concept. One fixed piece of polyethylene sheet (outer radius of elbow), and one sliding piece (inner radius) allowed the elbow to flex. A pin was inserted through the forearm component and holes locked the elbow in position. (e) One-piece molded “clamshell” above-elbow and forearm components joined by a living hinge. This prototype was vacuum formed, and could be custom-fitted by cutting the molded components with shears. A terminal device actuated by pressing it against the table (not shown) was also developed with this prototype.

Much of the cost in body-powered prostheses comes from the need for expensive fabrication equipment and skilled labor. To eliminate much of this, the team developed a method for fabricating a socket directly on the amputee’s stump similar to the below-knee interim prosthesis developed by Wu et al. 5 The amputee’s stump is covered with a plastic bag and then wrapped with fiberglass casting tape to create a negative impression. The rough fiberglass socket is impregnated with epoxy to provide strength and to smooth out the porous material. The materials are inexpensive and widely available, and the process does not require a high level of training. Both cost and time to completion are greatly reduced by eliminating the need for specialized equipment and the standard negative-positive-negative process.

Because most amputees prefer it, the Hosmer Dorrance (Campbell, CA) aluminum split hook (Figure 2A) was used as a model for the terminal device and modified according to the survey feedback and goals of the project. The moving hook was mirrored across the rigid hook so it opened laterally relative to the rigid hook, as opposed to medially (Figure 2B). The lever arm was then enlarged and angled downward to allow split hook actuation by grasping with the opposite hand or by exerting pressure against a surface, such as a table. The cross-sectional dimensions of the hook were increased to make it stronger and more rigid when molded from plastic.

Figure 2.
Figure 2.:
A: Hosmer Dorrance 5× split hook (right version). B: First split hook prototype (right version).

In a standard split hook, the bearing is a specialized component of specific size. Typically unsealed, it is vulnerable to sand and grit and difficult to access with ordinary tools, making repair difficult. Although this may not be a problem for amputees in North America or Europe, replacement can be a long and expensive process for amputees in developing countries. To avoid this inconvenience, the bearing housing was redimensioned to hold a skate/roller blade bearing (Figure 2B). This bearing is high quality, low cost, sealed on one side, and commonly available in a standard size around the world. Press fitting the bearing allows for easy replacement.

Prototype 1 was examined by local prosthetist, S. Nassan, who recommended filling in the diamond-shape opening between the rigid and moving hooks and cigarette notch, because individual users could easily add these to their own hooks, if desired. He also recommended additional strengthening, esthetic changes, and a better hook angle. These suggestions were followed, and the new design was sent to Solid Concepts, a California-based rapid prototyping company. Twelve samples of Prototype 2 were cast from AFP3000-G25, a 25% glass-filled advanced formula polymer that simulates filled polycarbonate (Figure 3). The articulating surface of both the rigid and moving hooks was coated with a nonslip silicone for a better gripping surface.

Figure 3.
Figure 3.:
Second split hook prototype (right version).

The polyethylene living hinge was also revisited. Light testing indicated the locking mechanism needed additional development; plastic components were being worn by the metal contact. A new mechanism was constructed from a releasable nylon cable tie. The cable tie was modified and situated in the elbow, as shown in Figure 4A, allowing the arm to flex in approximately 7° increments and to extend when released.

Figure 4.
Figure 4.:
A: Final elbow component design. B: Exploded view of the connection between the split hook and the forearm.

Other components were also constructed and optimized for strength, weight, and manufacturing ease. Schedule 40 PVC pipe ¾″ (1.905 cm) was chosen to provide the body of the prosthesis. A threaded bushing was designed to fit inside the pipe and join the hook to the forearm component (Figure 4B). Proximally, the socket connection was redesigned to distribute stress. The completed below-elbow and above-elbow prostheses can be seen in Figure 5. Both these prototypes were manufactured and fit for subjects. The cost of materials for each was approximately US $34.

Figure 5.
Figure 5.:
A: Completed below-elbow prosthesis. B: Completed above-elbow prosthesis.


Six subjects were recruited from local prosthetic clinics, five males and one female, ranging in age from 22 years to 63 years. One subject was unable to complete the study. Of the five remaining, four were below-elbow amputees, and one was above-elbow. Four were experienced prosthetic users, whereas one was undergoing fitting for his first prosthesis.

Testing was divided into three main stages, similar to the study conducted by Procter and LeBlanc. 6 All procedures were approved by the ASU Institutional Review Board. Accordingly, consent was obtained from each subject. In stage 1, subjects were videotaped at ASU completing a set of activities of daily living (ADL) with a prototype customized to fit. The ADL chosen were a combination of household tasks and medium-duty work chores (Table 1).

Table 1
Table 1:
Activities of daily living (ADLs) tested

Each subject rated the ease of completing each activity using a five-point scale. During stage 2, each subject used the prototype in freely chosen activities for approximately 3 weeks and recorded usage on a minicassette recorder. Subjects then returned to ASU for stage 3, in which the ADLs were completed for the second time and followed by an interview about the entire procedure. A local prosthetic professional (referred to as the rater) viewed the videotaped ADL sessions and, using a five-point scale, rated how well the subject performed each task.


In addition to human subject testing, the material properties of the split hook were characterized. The split hook was secured in a vise below an axial hydraulic table top test frame (Model: 303.1, Shore Western Material Systems, Monrovia, CA). It was compressed until failure at a rate of 25 mm/minute. Three split hooks were tested.


While recognizing the limitations of the prototype compared to their current prostheses, all subjects considered the prototype design suitable for its intended purpose. Activities were rated as easy or very easy to complete using the design more than 70% of the time. The low weight of the prototype was considered one of the best qualities; the elimination of the split hook cable did not rate as well. The most common suggestions included increasing the split hook grip force, allowing for rotation of the split hook, and improving the comfort of the socket.

Results of the mechanical testing placed the average stiffness of the split hook material at 177.85 N/mm. The average yield and maximum values were 964.33 N and 1179.39 N, respectively. Table 2 lists the scores given by the subjects and the rater for both ADL sessions. Subject 1 scored the second session of ADL as being more difficult overall than the first. This decrease in scoring is not consistent with the rater’s perception. The rater scored both sessions more equally and higher than the subject did. This decrease is also not consistent with subjects 3, 5, and 6, who scored all the second session activities, except for one or two, at either the same level or higher.

Table 2
Table 2:
Subject and rater scores for ADL sessions

In general, the rater scored more conservatively than did the subjects, handing out few ones and fives, except to subjects 2 and 6. Subject 2 received several poor scores from the rater. Because subject 2 was the only above-elbow amputee, this concentration of lower scores may indicate a weakness with the elbow design or some other aspect of the above-elbow prototype. However, without another above-elbow subject for comparison, it is difficult to tell whether this lower scoring is solely design related or possibly subject related. Conversely, subject 6 received practically all fives. This supports the rater’s assertion that subject 6 was the best problem solver.

Figure 6A depicts the number of times each subject received the five scoring levels during the first ADL Session. For each subject, the rater’s scores were combined with the subject’s, for a total of 22 scores. However, subject 1 did not feel comfortable completing task 8b, so he has only 21 scores. When a subject could not decide between two scores for a task, each was included.

Figure 6.
Figure 6.:
A: Frequency of times subjects received each scoring level during first ADL session. B: Frequency of times subjects received each scoring level during second ADL session.

Figure 6B depicts the number of times each subject received the five scoring levels during the second ADL session. For each subject, the rater’s scores were combined with the subject’s for a total of 22 scores. Again, when a subject could not decide between two scores for a task, each was included. Comparing Figures 6A and B reveals an overall shift to higher scores from the first to the second ADL session. More importantly, both figures depict the high percentage of easy and very easy scores for each session, 73% and 75%, respectively. (Subject 2 was not included in either Figure 6A or B because no comparison over time could be made.)

Table 3 summarizes the interview responses. Examining the interview responses revealed some similarities among subjects 1, 2, 3, and 6. All four indicated the prototype weighing less than their regular prosthesis was a positive trait. They also considered the elimination of the cable to be a poor modification. All four stated they were unconcerned about the appearance of the prototype. In fact, three subjects stated they would rather not have a natural-looking prosthesis. However, subject 5 did not mention the weight of the prototype. He also preferred manual operation of the split hook to cable actuation and was more concerned about the appearance of the prototype.

Table 3
Table 3:
Selected Interview Responses

As discussed, subject 2 was the only above-elbow amputee. Although this prevented any intersubject comparisons for the elbow, subject 2’s feedback was positive nonetheless. Subject 2 was very satisfied with the flexing capabilities of the elbow along with the load it was capable of supporting. Although noting the convenience of having one free hand when using a cable-operated elbow, subject 2 was very satisfied with the prototype elbow overall.

Flat and cylindrical objects were considered the easiest to grasp by subjects 1, 3, 5, and 6. Subject 2 found it easiest to hold thin, rectangular boxes, such as those containing cooking mixes. All subjects found it easier to open the split hook using their hand than by pressing against another surface. In fact, subject 6 never used the surface method. Finally, although subjects 1, 2, 3, and 6 rated the split hook lever as less functional than a cable, only subject 6 attributed any difficulty to the shape or location of the lever.

Table 4 lists some of the modifications suggested by subjects, along with how many subjects noted them. The split hook was commented on most frequently. Specifically, increasing the grip force of the split hook and increasing the durability of the hook coating were the two most suggested improvements. Reversing the position of the moving and rigid hooks was mentioned for two reasons. One subject thought it would be easier to grasp the split hook lever in an upward facing position. The second indicated ADL task 8a (carrying a bucket by grasping the handle) would be easier to accomplish if the bucket were resting against the rigid hook, instead of forcing open the moving half.

Table 4
Table 4:
Suggested modifications grouped by affected component

Two subjects suggested modifying the connection between the hook and forearm components to provide both rotation and stabilization. Finally, two subjects were displeased with the comfort of the socket. One suggested adding more padding, and one suggested manually compressing the tape during molding to create a tighter fit at the distal end.

The subjects chose to test the prosthesis in a large range of activities. Most activities in which the prototype worked well were household tasks or light outside work. Two of the most notable were unloading luggage and driving, even a manual transmission car. The prosthesis did not perform well in grasping heavier items, tying shoes, or riding a bike. No occupational activities were reported because subjects were instructed to use the prototypes in and around the house.


The prototype design met or exceeded the requirements for cost, simplicity, ease of repair, and adaptability. The prototype was successfully used in a variety of activities by the subjects and was considered suitable overall for light use. Although the design met the initial criteria, several of the subjects’ suggestions will be implemented to additionally strengthen the design. Durability, cultural acceptability, and the feasibility of remote fabrication will be the focus of future research.

This prosthesis was designed for amputees in economically disadvantaged countries who cannot afford the standard technology. They will most likely be first-time prosthesis users. Because of this, the scores should be interpreted carefully. The experienced amputees tended to focus upon any lack of function compared with their existing prostheses and this influenced their suggestions for improvement. Of the five subjects who tested this prosthesis, subject 5 may be the most representative of the target population. As a first-time user, subject 5 did not have to overcome a bias for cable actuation. His increased concern with appearance indicates he may not have fully come to terms with his amputation at the time of the study. This may also be true of the intended users because of societal and cultural values. Because of subject 5’s similarities to the target population, his ADL scores were examined to ascertain how easily other users would adapt to the prototype. Subject 5 scored 64% of the activities as very easy in the first session and 82% as easy or very easy in the second session. His scores in the second session were the same or higher for all but one activity. The high percentage of easy and very easy scores awarded by subject 5 indicates the ease of learning to operate the prototype. This was also reflected in the combined percentages of easy and very easy scores for all subjects (73% for the first session and 75% for the second session).

Increasing the grip force of the split hook was identified by three subjects as a needed improvement. The number of O-rings securing the split hook determines the amount of grip force. More may be easily added. In addition, the diameter of the O-ring posts may be increased to preload the rings. In response to increasing the durability of the split hook coating, the current coating was intended only for the alpha prototype. A more effective nonslip surface will be provided by incorporating rubber inserts into the cast or injection molded plastic. The feedback from the subjects confirmed the necessity of this addition. The overall design of the split hook, especially around the tips of the hook “fingers,” will be examined and strengthened. Thickening the hook fingers should decrease their flexibility.

Although reversing the moving and rigid hooks was mentioned by two amputees, this suggestion will probably not be followed. Placing the split hook lever on the medial side of the arm would eliminate one of the two operating mechanisms: pressing against a surface. Without this capability, the split hook would have to be activated using the other arm, decreasing its usefulness.

Two other suggestions to improve the functionality of the split hook will be followed. The end of the lever will be squared to provide more stable contact against flat surfaces. In addition, a simple mechanism for easily varying the grip force will be investigated. Although this has been accomplished in other designs, prehensors with variable grip force have never been commercially successful. 7 Finally, the distal connection will be modified to provide both rotation and stabilization of the split hook, depending on the user’s needs. One subject implemented this on his prototype by installing two opposing set screws into the distal connection. These were tightened when desired to hold the split hook at a fixed angle.

A good socket fit is crucial to an amputee’s success with a prosthesis. Not only will a poor fit discourage use, it may also cause pressure sores to develop on the user’s residual limb. The comments regarding the socket fit of the prototype were carefully studied. Many amputees have little soft tissue to act as a cushion between their bone and their skin. This was the case with the subject who suggested thicker padding. A solution to this problem is to build up bony prominences when molding the socket, thereby creating reliefs in the socket when the extra material is removed. The subject who mentioned molding for a tighter fit at the distal end claimed this was how his regular prosthesis was molded. According to him, pinching the casting tape against the residual limb before it has fully cured will help to position the radius over the ulna (neutral pronation/supination) and create a more solid fit.

One of the goals of this project was to eliminate the labor-intensive socket manufacturing process. Although this means the prototype will not be as functional or durable as a professionally made prosthesis, a very good fit can still be made at a much lower cost. Additional sockets can be made quickly, optimizing fit via trial and error.

Examining the activities in which the subjects cited poor performance of the prosthesis was also beneficial, by highlighting much-needed alterations. Two of the activities in which the prototype did not perform well are associated with the level of grip force provided. Increasing grip force should help the prototype in grasping heavier items and tying shoes. It may also help in riding a bike, although altering the curve of the split hook may be more effective. Increasing the radius of the bottom hook’s curve and displacing it laterally would also aid in grasping other cylindrical objects, such as a broom handle.

Mechanical testing indicated the split hook material should be more than sufficient because loads experienced during regular use are not expected to approach, let alone exceed, the yield values. However, the split hook overall is not yet robust enough to be used in manual labor tasks as intended. According to the subjects, it is currently suitable for light duty tasks or a less active person. Applying the suggestions already cited should make it suitable for heavier use.

Although distribution costs are not currently known, it was encouraging that the materials cost for both the below-elbow and above-elbow prosthesis designs was approximately one-third the $100 goal listed in the design criteria. As determined by the 1-3-9 Rule, wholesale costs should then be in the neighborhood of $100. 8 Granted, the initial cost of the split hook injection molds was not included, but this cost will be reduced significantly by economics of scale.

Other design criteria included the development of a simple design, use of easily repairable components, and a design that is readily adaptable to individual needs. These were all satisfied by the prototype. After applying the design changes to the split hook and refining the cold molding process, testing of beta prototypes in several international locations is planned to better determine on-site durability, cultural acceptance, and remote fabrication. Local craftsmen, including amputees, will be trained to work with the casting tape and epoxy. Although the process is easy to learn, a few will likely excel and specialize in the procedure, ensuring future sockets are uniform in quality and well-suited for each individual amputee.


An inexpensive prosthetic arm has been built that was considered sufficiently functional for light duty tasks by trial users. The most common suggestions included strengthening the grip force of the split hook and increasing the comfort of the socket. These and other suggestions will be incorporated into a beta prototype that will be tested at several international locations. After determining the feasibility of manufacturing the device on location, the device will be packaged as a kit and sent overseas. Once refined and tested, this prosthesis has the potential to mobilize large numbers of amputees who are currently out of work. Having regained the capacity for employment, amputees and their families will enjoy an increased quality of life. A stronger work force should have national implications as well. Perhaps more importantly, examples of successful amputees will help decrease the fear and uncertainty associated with amputation in so many countries.


The authors acknowledge the support of National Institutes of Health Small Business Innovation Research grant R43-HD43512–01 for this research and the graduate assistantship provided by the Flinn Foundation. Thanks are given to S. Nassan, CPO, FAAOP, and R. West, CP, for valuable design feedback and testing assistance; M. Pack, CP, and J. Pongratz, CPO, FAAOP, for recruiting subjects; and J. Egan for the survey information. T. Campagna and the 3M Corporation are also acknowledged for their generous donation of Scotchcast® Plus casting tape.


1. Meanley S. Different approaches and cultural considerations in Third World prosthetics. Prosthet Orthot Int 1995;19:176–180.
2. Loro A, Franceschi F. Prevalence and causal conditions for amputation surgery in the Third World: ten years experience at Dodoma Regional Hospital, Tanzania. Prosthet Orthot Int 1999;23:217–224.
3. Coupland RM. Assistance for victims of anti-personnel mines: needs, constraints, and strategy [brochure]. International Committee of the Red Cross; 1997.
4. Kurzman S. Cultural attitudes toward prostheses: an anthropological approach. Capabilities: communicating the science of prosthetics and orthotics. October 1997. Available at: Accessed August 29, 2001.
5. Wu Y, Brncick MD, Krick HJ, et al. Scotchcast® PVC interim prosthesis for below-knee amputees. Bull Prosthet Res 1981;18:40–45.
6. Procter S, LeBlanc M. Clinical evaluation of a new design prosthetic prehensor. J Prosthet Orthot 1991;3:79–83.
7. Frey DD, Carlson LE, Ramaswamy V. Voluntary-opening prehensors with adjustable grip force. J Prosthet Orthot 1995;7:124–131.
8. Rondeau HF. The 1 (materials)-3 (cost of goods)-9 (selling price) rule for product cost estimation. Machine Design 1975;47:50–53.

amputee; developing countries; inexpensive; Third World; upper extremity prosthesis

© 2004 American Academy of Orthotists & Prosthetists