Ergonomics: Introducing the Human Factor into the ...

PEER-REVIEWED PAPER

Ergonomics: Introducing the Human Factor into the Clinical Setting Robert S. Bridger, PhD and Mladen A. Poluta, BSc (Eng) Health Technology Management Program Department of Biomedical Engineering University of Cape Town/Groote Schuur Hospital UCT Medical School Observatory 7925, South Africa

Applications in the fields of user-interface design, medical device design and accidents and injuries in the hospital setting are described to illustrate the ergonomic approach. Data on the cost-effectiveness of ergonomics are cited. Finally, some international standards for ergonomics and some recent regulatory trends are described. Index under: Ergonomics, Human Factors, Human Error, Medical Equipment Work Environment.

INTRODUCTION

T

he name "ergonomics" comes from the Greek words "Ergon", which means work, and "Nomos" which means law. Ergonomics is the study of the interaction between people and "machines" (technology) and the factors that influence the interaction. Its aim is to make systems work better by improving the interaction between people and machines. This can be done by "designing-in" a better interface or by "designing-out" factors in the work environment, the task or the organization of work that degrade human-machine performance. For example, work systems can be improved by: • making the user-interface easier to use or making it resistant to errors that people are known to make; and • changing the work environment to make it safer and more appropriate for employees or changing the task to make it more compatible with the user's expectations and knowledge and/or changing the way work is organized to accommodate peoples' psychological and social needs. The implementation of ergonomics in system design should make the system work better by eliminating undesireable, uncontrolled or unaccounted-for aspects of system functioning, such as: • Inefficiency - when worker effort produces suboptimal output; 180

Journal of Clinical Engineering • May/June 1998

• Fatigue - when people tire unnecessarily in badly designed jobs; • Accidents, injuries and errors - due to badly designed interfaces and/or excess stress (either mental or physical); • User difficulties due to inappropriate combinations of sub tasks making the dialogue/interaction cumbersome and unnatural; and • Low morale and apathy. Absenteeism, injury, poor quality and unacceptably high levels of human error are seen as system problems, rather than "people'' problems, and their solutions are seen to lie in designing a better system of work, rather than in better "human management" or incentives, by "motivating" workers or by introducing safety slogans and other propaganda. Similarly, human requirements are seen as system requirements rather than secondary considerations, and can be stated in general terms as requirements for: • equipment that is usable and safe; • tasks which are compatible with people's expectations and limitations, • an environment which is comfortable and appropriate for the task, and Journal of Clinical Engineering • Vol. 23, No.3, May/June 1998 Copyright © 1998, Lippincott-Raven Publishers

Bridger & Poluta

.

• a system of work organization that recognizes peoples' social and economic needs.

Cost Implications

-

Failure to consider the human factor in equipment or facility design or in equipment procurement usually has t!nancial implications in either the short- or long-term. Increased costs get passed on to consumers either directly through higher healthcare costs and insurance premiums in consumer-funded systems or by higher taxation or poorer service in socialized systems. The cost of health care is much-debated nowadays, and its impact is thought to be exacerbated by demographic changes in the population of many industrially developed countries. According to Shelton and Mann-Janosi (1992), the US. has the highest per capita spending on health care of all industrialized nations, and yet the largest percentage of people not covered. Some examples serve to illustrate the ripple effect of high healthcare costs, and therefore, the economic justification for making improvements. Stamper (1987), then Vice President of Boeing, reported that the company spent 53,500 per employee per year on health care ($350 million annually) - more than it spent on aluminum to build its airplanes. Also in the 1980's, Chrysler

found that its healthcare obligations accounted for 10% of the cost of a basic car and that an insurance carrier, not a steelmaker, was its single busiest supplier (Stamper, 1987). Shelton and Mann-Janosi (1992) cited similar cost-estimates for the Ford motor company - $400 per vehicle in direct costs and $675 per vehicle if the health care costs of Ford's suppliers are included.According to these authors, the U.S. healthcare system competes on services, not price, which tends to drive prices upwards. Prices are generally higher in competitive, multi-hospital markets than in areas served by a single supplier. Some of these issues have been addressed by the move towards managed health care, but new concerns relating to freedom of choice and quality of care have been raised. The increasing sophistication of much medical technology is one factor that may increase healthcare costs. Another is the unsophisticated, stressful environment in which technology is used. Many occupational injuries and diseases are caused by work environment and work organization factors and, because these injuries are not always recognized as being occupationally related, the actual cost of work-related employee ill-health may be underestimated, and therefore covered by employees' health insurance rather than workers' compensation schemes (Rey and

Recycling Waste products_ } ReJects Accidents

BYPRODUCTS

Matter

\----~ Products

}~

....,__ ___,. Energy(transformation

Energy

1-----~

Knowledge

~

::>

0

Information

}

BYPRODUCT$

Heat Recycling

FIGURE 1

Generic model of a worksystem. In a hospital, the inputs include patients, energy, medical technology and knowledge. Within the system, personnel (H) interact with technology/machines (M) in a work environment (E). The intended outputs are patients who have received treatment and, in the case of academic hospitals, medical knowledge. Byproducts, or unintended outputs, include waste, pollution, injured patients and/or employees, etc.

Bridger & Poluta


181

Busquet. 1995). Nurses, in particular, are a high-risk group for injury - particularly musculoskeletal injury - due to the heayy manual work they are required to carry out as part of their jobs (e.g. Garg and Owen, 1992). When it is remembered that one day·s sick pay can cost the employer up to 3.5 times the daily wage (Oxenburgh, 1991) the costs of bad ergonomics can be large.

Ergonomics and latrogenics latrogenics refers to medical problems caused by the system of medical care, including clinicians, managers and technology. Iatrogenic complications can have ergonomic causes (Gardner-Bonneau, 1993).A great deal of the equipment in hospitals has no" owner". It is often on wheels and is handled by many different people, sometimes in emergencies. and is mechanically "traumatized'' (e.g., out of calibration or dysfunctional). Disposable equipment (e.g .. syringes) is sometimes reused, but doesn't survive autoclaving. Hospitals around the world often lack formal systems of quality control and formal maintenance standards (in the industrial sense of the term) and there is either no feedback or delayed feedback about equipment performance and reliability. Shortcomings or bad practices may persist due to lack of negative feedback. Professor L. Levin of Yale University School of Medicine (1993) reports that about 10,000 people die per year in the U.S. as a result of anesthesia error. much of which is related to human error. A study of 1,800 diagnoses revealed an error rate of 20%; half of these errors may have contributed to patient death, despite that fact that clinicians order large numbers of tests on patients before reaching a diagnosis (almost two thirds of the tests are inappropriate or unnecessary).According to Professor Levin. blood pressure cuffs and x-ray machines are frequently out of calibration. damaged or leaking (in the case of x-ray machines), reinforcing the view that many of the iatrogenic problems in medicine stem from the interaction between machines and their sometimes diverse body of users.

Applying Ergonomics to Clinical Settings Two main areas of focus, both very relevant to clinical settings. are the user-interface and its design to ensure usability in the performance of work tasks, and the general work environment and its effects on hospital workers.

HUMAN-MACHINE INTERACTION ISSUES Some questions equipment designers should consider at the early stages are: Will users understand the product? How easily and quickly will they learn to use it? Is it easy to use when learned? Does the product build on existing user knowledge or does it require new knowledge? Are similarities between the product and its predecessors realistic? Are comparisons between the product and other

182

journal of Clinical Engineering • May/june 1998

products valid (does the same "mental model" apply)? • Are the operations and procedures compatible with people's general knowledge of operations and procedures and with similar or related products?

Medical Devices Mosenkis (1994) has reviewed some of the human factors aspects of medical devices emphasizing (i) the need to reduce errors and the consequences of errors and (ii) correct procedures for use. As discussed elsewhere (e.g. Reason, 1990) the way a system is designed determines the errors that are possible. Some design faults can be summarized as follows: • Unusual or inconsistent operation. Users may assume that if they know how to use one type of infusion pump or controller, they know how to use all types. Problems can occur if different makes of a product have different control logic. On the other band, Dixon (1991) has shown that positive transfer of learning can take place between different devices if there is a high degree of operational similarity, even if there are large conceptual differences (e.g. landing an airplane and closing down a nuclear power station). • Lack of protective incompatibility. Some medical devices can be misconnected due to the design of their leads and sockets resulting in equipment damage or patient death. Mosenkis (1994) reports how a design of apnea monitor lead had a chest electrode at one end and a plug at the other. The plugs could be incorrectly connected to AC power connectors. Several infants were electrocuted because people saw unconnected leads close to power connectors and assumed that the two should be connected. • Unclear or incompatible control/display relationships. Traditionally, ergonomics has argued for compatibility between display and control configurations. In analog mode, this usually means that there should be a good spatial mapping between the two. Movements of a control should produce an effect on the display according to user expectations (e.g., turn the knob to the right to increase the voltage on the electrothermal unit or to the left to increase the flow of a liquid or gas. Well-learned relationships such as these can be exploited to standardize design decisions and make the operation of inter-faces "intuitive". However, the shift from analogto symbolic modes of representation can introduce complexity in unanticipated ways. Mosenkis reports the case of a defibrillator/monitor that displayed alphanumeric codes whenever a problem arose.The codes were explained in a table in the manual. Staff solved the problem by cutting out the table and taping it to the monitor. Arguably the provision of a

Bridger & Poluta

-

table on the device was the designer's job, given the questionable decision to use alphanumeric codes in the first place. • Defeatable or ignorable safety features. Auditory warnings attract more attention than visual ones (Posner, 1980). LCDs and other low-power visual displays are less effective than flickering lights, which can readily be detected by peripheral vision. Vibration and smell can be very effective non-directional warnings in some cases: vibration induced by ripples in road surfaces can warn of a hazard ahead and unpleasant odors can be added to otherwise undetectable gasses to warn of leaks. Miller and Beaton (1994) describe the problem of detecting emergency vehicle sirens by car drivers. The intensity of these sirens in the U.S. is 118 dB(A) at 10 meters. The car (assuming the windows are closed) attenuates this by at least 20 dB(A) and the siren needs to be at least 8-12 dB(A) louder than the noise inside the car (about 70 dB(A) with the radio off). If the emergency vehicle is travelling 30 km/h faster than the vehicle in front, the car driver will not hear the siren until the emergency vehicle is about 32 meters behind. This will leave the car driver about four seconds to get out of the way. Not surprisingly, emergency vehicles frequently pass us almost before we hear their sirens. This rather dramatic example illustrates an important point about auditory alarms the intensity of the alarm must be optimized in relation to the environment, and the two key considerations are the nature and intensity of the background noise and the distance of users from the alarm (signals are more easily masked by background noise of a similar frequency, a signal-tonoise ratio of about 10 dB(A) is needed, and sound attenuates rapidly with distance, particularly in open spaces). • Lack of cues to aid discrimination. Mosenkis gives the example of the contact lens package that contained two almost identical bottles, one containing a lens cleaning fluid and another containing saline to wash off the fluid. Small labels distinguished the bottles but were difficult to read. Eye damage was caused by users cleaning lenses with saline and rinsing with cleaner. These could be avoided by, tor example, using a blue, rounded bottle for saline and a red square or hexagonal bottle for cleaner (augmenting written warnings with color and shape cues to distinguish safe from unsafe products).

Improvement of Interface Design Given the abundance of modern design tools, there should be no technical barrier against developing and implementing user interfaces that meet established timetiona! and usability requirements. Unfortunately, many userBridger & Poluta

interfaces implemented in commercial, industrial and healthcare settings never realize their full potential on both functionality and usability scales (Maddox andAllen, 1993). Also, technologies have traditionally forced people to conform to the needs of machines. In this era of adnnced information processing, the power of the machine can be readily tailored to the needs of the user (Norman, 1993). Most texts in ergonomics emphasize knowing the users and the constraints they operate under and using a highly iterative approach with user trials undertaken from the very earliest stages (e.g. Bridger, 1995, Cox and Walker 1992,Preece.1992).

Benefits of Improved Designs There are many convincing case studies from areas outside of health care. Chapanis (1991) cites several studies ,perhaps the most convincing of which is the implementation of centered, high-mounted brake lights on cars (Malone, 1986). Four groups of 525 cabs each were monitored over a one-year period in Washington, DC. Data on rear-end collisions, other collisions and mileage were gathered. Non-rear-end accidents were the same in all groups, whereas a 54% reduction in rear-end accidents was found in the experimental group with the centered. high mounted brake light. Similar reductions have been found in other studies of these brake lights .The cost of installing a light in 1977 was S4.00 per vehicle, while the average vehicle repair cost for rear-end collisions was $317.00 for vehicles with conventional brake lights and S 194 for those with high centered brake lights. The cost saving for the 3.2 million accidents in 1977 would have been $1.465 billion, not including medical, legal and lost time costs. Clearly, improved design can save lives and money. Mantey and Teorey ( 1988) provide a more detailed analysis of the potential benefits of introducing ergonomic considerations into the software development life-cycle. New prototyping tools make it possible to present users with mock-ups of a new system very early on in the design process and with sufficient fidelity to give them a genuine "feel" for what the system will be like to use. The costs of including ergonomics in the design stage of the product life cycle include: • Salaries of additional staff, consultant's fees etc. • The cost of running focus groups • The cost of building prototypes or mock-ups (story boards, ·'empty" interfaces, etc.) • The cost of prototyping tools • The cost of modifying prototypes • Costs associated with running user trials • Usability laboratory costs • Survey and questionnaire design, implementation and analysis The benefits of utilizing an ergonomic approach include: • A reduction in learning times and a direct saving on Journal of Clinical Engineering • May/June 1998

183

user salary and trainer salary expenditure • Fewer errors when the system is up and mnning • Lower maintenance costs • Possible avoidance of costs due to retrofitting/ upgrading of the system after implementation • Faster processing of information at the humanmachine level Mantei and Teorey give estimates of actual cost savings of $280,000 in the first year for a system with 32,000 delivered source instc·Jctions to be used interactively by 250 employees. They emphasize that there are numerous intangible benefits that arise due to increased usability. Also, the main benefit of the iterative approach to system design is to shift design changes to an earlier stage in the product lifecycle where they will be easier and cheaper to implement. However, with the emergence of popular software solutions that become "industry standards'· comes the risk for smaller or less-established manufacturers to be caught "off-side" in the intellectual property war; this could extend to medical equipment manufacturers. Any visual similarity to an existing program makes software vulnerable to litigation by a competitor. For example, firms like Lotus Development and Apple Computer have been suing competitors over similar user-interface features (Dossick 1994). The same concerns apply to health service organizations where life-saving equipment may not be operational because of system (user-interface) design- please refer to case study quoted in the section entitled "Some Global Trends". In this situation, the hospital may itself become liable for not ensuring that staff are properly trained in the use of such equipment. Medical device designs as a whole may also minimize (or increase) the potential for human error in the application of the device (Vredenburgh, Saifer and Cohen, 1995; Hutchinson 1995).

Technology Transfer and User-Interface Design There is a growing amount of literature on technology transfer, its potential benefits and problems, and the precautions that have to be taken if it is to succeed. Ergonomic aspects of technology transfer have been discussed by numerous authors. Abeysekera and Shahnavaz (1989) discussed the impact of body-size variability between people of developed and developing countries, and its impact on the use of imported goods. The physical differences between people of different countries can have significant impacts; for example, a product designed to fit between the fifth and 95th percentiles of British users would fit 90% of German and U.S. users, 70% of Italians, 59% of Egyptians, 43% of Japanese, 22% of Indians and 13% of Vietnamese users. Further information on some of the physical differences between users in developed and developing countries can be found in Wisner (1989). In a study of private hospitals in Cape Town, Botha and 184


Bridger (in press) found that a significant amount of the usability problems and musculoskeletal pain experienced by nurses at work was related to physical mismatches between nurse and equipment dimensions. Neither the designers nor the management of the hospitals concerned were aware of the importance of achieving good fit between nurses and their work environment. For improved usability and reduced stress, the authors recommended: • More stepladders to be provided in storage areas • All plugs and sockets to be raised and made accessible • Monitors to be installed no higher than fifth percentile eye height • Objects to be placed no higher than fifth percentile overhead reach height • Heavy objects to be stacked between knee and shoulder height • Space for manual handling to be increased in all patient bathrooms and toilets • Height-adjustable beds and equipment • Nurses' stations to be designed in accordance with nurse anthropometry and task requirements Many of the changes were implemented and the rest were filed for use in the design of a new facility. Users in developed and developing countries also differ in relation to their educational experiences and exposure to technology, such that assumptions made by designers in developed countries (assumptions which are often inappropriate even in their own countries) are entirely inappropriate in other contexts (see Meshkati 1986). The environment in which people grow up as well as the formal and informal educational processes to which they are exposed have a major influence on the cognitive stmctures they develop. To merely state that people from different cultures think differently is trivial - much of human behavior and biology is clearly an adaptation to the particular surroundings. However, it becomes important when people whose cognitive stmctures have developed in a particular socio-technical milieu are exposed to new technologies or industries. This can happen when technologies or work systems from developed countries are transferred to developing countries without taking into account the knowledge and beliefs of workers in the recipient country (McKie 1990). There may be mismatches between the knowledge and cognitive styles of the users and the requirements of the technology being introduced. In industrially developing countries, people's formal exposure to technology, as well as to the infrastructure that makes it possible to own and use technology, may be lacking. Potential users of new technologies have therefore developed in an environment in which they have not been able to internalize many key concepts about how technology, and more generally, technological society, actually works. A common user coping Bridger & Poluta

strategy when faced with a mismatch of this nature is to learn by rote without developing any form of "mental model'' or high-level representation of technology, the context in which it operates or the environment which supports it. This is a workable strategy under the circumstances, because it enables the individual to function and to interact with machines on a routine basis. It is a strategy which will be predicted to break down whenever the individual is presented with novel situations or unprogrammed events beyond the scope of the rote-learned behaviors. This is one of the reasons why technology transferred from developed to industrially developing countries may fail, often spectacularly, and as a result of behavior that seems bizarre or inconceivable to someone with appropriately internalized concepts. In this sense, the individuals supposedly responsible for the accident are not responsible - they are just incompatible with the system. In these cases, there is a need to" fit the worker to the job" by means of appropriate training and upliftment via exposure to enriching technological experiences. Recent studies of injury and accident in our own hospital show an unexpectedly high number of injuries to nurses when adjusting the cot sides of imported hospital beds. The cot sides are designed with a ··safety" feature to prevent their being lowered accidentally. Three knobs have to be operated in sequence before the cot side will drop. Poorly educated nurses frequently suffer foot injuries from falling cot sides as a result of incorrect operation.

WORK ENVIRONMENT AND PSYCHOSOCIAL ISSUES A major source of problems in the healthcare setting is the physically dysfunctional work environment of hospitals and its negative effects on staff, particularly nurses. Compatibility is often lacking between the equipment provided, the clinical setting and the needs of caregivers.

Musculoskeletal Disorders Occupational musculoskeletal disorders constitute a major source of debility in nurses, more so than in other hospital workers and with industrial workers (Cust et a!., 1972). Ljungberg eta!. (1989) compared nurses and warehouse workers and found more injuries amongst nurses. Baty and Stubbs (1987) found that nurses spend 22% of daily work time in a stooped posture. Since back injury is a process rather than an event (Pheasant, 1995), the daily postural stress on the nurse's back has a perniciously weakening effect over the years and eventually, often as an immediate result of sudden or unexpected loading, the trunk fails catastrophically. Garg and Owen (1992) evaluated the effectiveness of an ergonomic intervention program in reducing back injuries amongst personnel in a nursing home.They identified the tasks perceived to be the most stressful by careBridger & Poluta

givers and then carried out an ergonomic evaluation of these tasks. Patient handling devices were selected and personnel were trained in their use. Mechanical hoists for lifting patients out of bed were introduced, as were "walking belts" (a method of "putting handles" on patients). Modifications were made to toilets and showers to improve patient handling. The back injury rate dropped from 87 injuries per 200,000 work hours to 47 injuries per 200,000 hours. Ljungberg (1989) found that with spacious premises, electric hoists and better work organization, patient lifting could be reduced by 50%, with a consequent reduction in risk. Musculoskeletal problems are not restricted to nurses, however. They are also common in radiographers due to the nature of their work and clothing (e.g., heavy aprons) and the manual handling of x-ray cassettes. May et a!. (1994) report particular problems amongst radiographers involved in mass screening programs for breast cancer. Particular care should be taken when designing facilities for mass screening procedures because of the repetitive nature of the work. Crombie and Graves (1996) investigated the ergonomics of new surgical instruments tor keyhole surgery (specifically, Laparoscopic Cholycystectomy). Keyhole surgery is often cited as being more patient-friendly and cost-effective than conventional surgery, but according to the authors, ergonomics has been overlooked because of the rapid development and uptake of the procedure. 64% of surgeons interviewed reported musculoskeletal problems associated with the use of specific instruments, for example: Grasping forceps: a handle/tool angle of 115° normalizes wrist posture and minimizes strain when the tool is worked within one plane, but exacerbates postural problems when the tool is used in many planes, as was observed in this study. Electrosurgical hook: Hand fatigue was caused by the adoption of a precision grip (first two fingers and thumb) due to the handle being too short. The weak finger muscles fatigue rapidly when they work at a mechanical disadvantage. Knee pain was caused by having to adopt a twisted leg posture due to poor placement of foot pedals. Suction/irrigation device: Control buttons were placed directly on the stem of the instrument instead of on a properly designed handle, resulting in the exertion of excessive grip forces during use. General risks: The operative procedure required approximately 30 changes of instrument, removing each from the cannula, reinserting it into the instrument holder and grasping the next instrument and reinserting it into the cannula. These problems may appear to be minor but since keyhole surgery requires fine control, the precision of the surgeon's actions must not be degraded by the need to exert excessive forces or by avoidable muscle fatigue. Journal of Clinical Engineering • May/June 1998

185

Psychosocial Issues Affecting Quality of Care Persson et al. (1993) investigated perceptions of quality of care and job concerns amongst a sample of hospital staff from seven different departments. Perceived quality of care was influenced by a number of factors in the total work environment. In order of importance: 1. Social support 2.Work pace 3. Demands on attention 4.Training on equipment 5. Professional Development 6. Stimulation. mental variety 7. Time pressure 8. Smoking 9.Adaptation of environment and equipment Strong social support. a work pace which was not too rushed and low demands on attention were associated with reports of high-quality care. Better adaptation of environment and equipment and better training were also linked to reports of higher quality of care. Participation in equipment decisions, level of skill and issues of information and ergonomics were not linked to perceptions of better quality care. Better adaptation and training were associated with lower anxiety about equipment usage. Work role insecurity, a lack of discretion and work constraints were also associated with high anxiety. whereas issues concerning skill, leadership. work pace and participation on equipment decisions were not. These findings highlight the importance of achieving compatibility at the psychosocial, physical and physiological levels.

job Aids and Training Goemans et al. (1995) developed a remote control system for using visual reinforcement audiometry to test the hearing of children. The test was originally carried out by an audiologist with an assistant to help with the presentation of the visual reinforcements every time the child detected the auditory stimulus. A remote-control box was developed to enable the audiologist to control both the audiometer and the visual reinforcement system from inside the testing room. Due to close work with audiologists and the creator's understanding of the requirements of the task, the new system had the following benefits: • One audiologist could perform the tests without assistance- a saving in personnel time. • By working with the child in the testing room, the audiologist could better gauge the optimum time to present a stimulus. bearing in mind the child's readiness to respond. • The audiologist could better judge whether the child had or had not heard the stimulus and better control the presentation of reintiJrcement. • Speech discrimination tests were easier to perform on very young children. 186


• Audiologists reported a reduction in workload and better control over the testing procedure. Cowley and Gale (1996) investigated human error in the diagnosis of breast cancer by radiologists. Over 90% of breast screening radiologists from the U.K. participated. The researchers used signal detection theory (Green and Swets, 1974) to classifY diagnostic behavior and identity the individual radiological features that radiologists had problems detecting. The goal of the program, which was based on selt~assessment and retraining, was to minimize the number of false-negative errors. Retraining was found to improve performance in the detection of certain types of breast cancer in the short-term.

SOME GLOBAL TRENDS Development of human-machine systems is now influenced by the demand for increased efficiency at many levels. Both the U.S. and Europe have begun to draft and/or implement regulations governing various aspects of the work environment. NIOSH has updated its equation used to evaluate manual handling tasks (Waters et al., 1993) and attempts have been made to introduce an "ergonomic standard" in the U.S. In Europe, there are now regulations governing the design of manual handling tasks and of visual display terminal (VDD workstations in addition to national design standards such as BSI 7179 (and its U.S. counterpart ANSI/HFS 100-1988). These regulatory trends represent an attempt to formalize aspects of equipment design and the work environment which is stimulated by the need for increased efficiency both in the workplace in general and in the product development process in particular. Europe now has legal requirements covering the design of interactive software (see Bridger, 1995. for further discussion of these trends). The 62A Committee of the International Electrotechnical Commission has recently (late 1996) circulated proposals for new standards relating to graphical symbols for use on general medical equipment, medical monitoring equipment and radiologic equipment. The proposed standards are, in part, a response to the trend towards common international standards in user-interface design. While this provides undoubted benefits (in terms of production and cost-saving) to medical equipment manufacturers. there are potential hazards in moving away from text-based controls and interfaces. Users of equipment which relies heavily on graphic symbols user-interfaces need to be properly trained on an ongoing basis to cover staff rotation, particularly in hospital wards/units with life-support and life-saving equipment. However, even training will not prevent the situation reported to have arisen in a South African hospital some years ago: during an operating theater emergency, staff discovered that their defibrillator was faulty. A defibrillator from the adjoining theater was brought in; however, this defibrillator was of a different make and Bridger & Poluta

Ergonomics in Design ,July: 18-21. relied heavily on graphic symbols in its user-interface. The theater staff, not recognizing the symbol for the Standby mode, were unable to operate the defibrillator in time and the patient died. Another difficulty commonly encountered in developing countries is the language barrier that occurs when equipment installation, operating or service documentation is in a language which is not understood locally. This may be particularly true of equipment supplied via donor aid mechanisms. The authors have seen a sophisticated operating theater table still not installed six months after delivery. The reason offered was that nobody understood the installation instructions, which were in a foreign language and interestingly did not make use of any illustrations.This raises the issue of documentation design which, one could argue, forms part of the extended user-interface.

CONCLUSIONS A lack of ergonomics in system design costs money and risks lives. There is now, however, a convincing body of evidence to show that the inclusion of ergonomic principles can reduce these costs and provide a safer environment for both hospital patients and employees.

REFERENCES Abeysekera,).D.A. and Shahnavaz, H. (1989). Body size variability between people in developed and developing countries and its impact on the use of imported goods. Int.] of Industrial Ergonomics, 4:139-149. Batey, D. and Stubbs, D.A. (1987). Postural stress in geriatric nursing. Int.] of Nursing Studies, 24:339-344. Botha,WE. and Bridger, R.S.Anthropometric Variability. Equipment Usability and Musculoskeletal Pain in a Group of :'-lurses in the Western Cape. Applied Ergonomics, In press. Bridger, R.S. (1995). Introduction to Ergonomics. McGraw-Hill Inc, New York. Chapanis,A. (1991). The business case for human factors in informatics. In: Human Factors for Informatics Usability, B Shackel and S Richardson (eds.), Cambridge University Press. Chapanis,A. Lindbaum, L.E. (1959).A reaction time study of four display/control linkages. Human Factors. 1: 1-17. Cowley H Gale A. 1996. Minimizing human error in the detection of breast cancer. ContemporaiJ' Ergonomics, 1996. Edited by S.A. Robertson, Taylor and Francis London: 379-384. Cox, K. and Walker, D. (1993). User Interface Design. Simon and Schuster (Asia) 2nd edition. Crombie. N.A.M. and Graves, R.). (1996). Ergonomics of keyhole surgical instruments - patient friendly, but surgeon unfriendly' In. Contemporary Ergonomics, 1996. edited by S.A. Robertson, Taylor and Francis, London:385-390.

Garg,A. Owen, B. (1992). Reducing back stress to nursing personnel: an ergonomic intervention in a nursing home. Ergonomics, 35:1.353-1375. Goemans, B. Bunn, A. E. and Bridger, R.S. (1995). Ergonomics and hearing tests for children. Ergonomics in Design, 20-25. Green, D.M. and Swets, ).A. 19...,4. Signal Detection Theory• and Psychophysics. Wiley, New York). Hutchinson. G. (1995). Taking the guesswork out of medical device design. Ergonomics in Design ,April issue, 21-26. Leamon, TB. (1980). The organization of industrial ergonomics: a human-machine model. Applied Ergonomics. 11:223-226. Ljungberg,A.S. Kilbom,A. Hagg. G.M .. (l989). Occupational lifting by nursing aids and warehouse workers. Ergonomics, 32:59-78. Maddox, M.E. and Allen, D.M. (1993). What have you done to my interface? Transitioning from prototype to implementation. Ergonomics in Design,]anuary issue. 13-19. Malone, R.B. (1986). The centered, high-mounted brake light: a human factors success story. Human Factors Socie(v Bulletin, 29:1-3. Mantei, M.M. and Teorey,T.). (1988). Cost/benefit analysis for incorporating human factors in the software lifecycle. Communications of the ACM, 31:428-439. May, ).L. Gale, A. G. Haslegrave. C.M. Casteldine,). Wilson A.R.M. 1994. Musculoskeletal problems in breast screening radiographers. Contemporary Ergonomics, 1994, edited by S.A. Robertson. Taylor and Francis, London. McKie) (1990). Management of medical technology in developing countries.] Biomed. Eng., 12:259-261. Meshkati, N. Robertson, M.M. (1986).The effects of human factors on the success of technology transfer projects to industrially developing countries: A review of representative case studies. In: 0 Brown jr H.W Hendrick (eds.) Human Factors in Organizational Design and Management, North Holland, Amsterdam. Meshkati, N. ( 1986) . .\1ajor human factors considerations in technology transfer to industrially developing countries:An analysis and proposed model. In: 0 Brown jr H.W Hendrick (eds.) Human Factors in Organizational Design and J1anagement, North Holland, Amsterdam. Miller, M.E. and Beaton, R.). (1994).The alarming sounds of silence. Ergonomics in Design,)an: 21-23. Mosenkis, R. (1994). Human factors in design. In: International Perspectives on Health and Safety, C.WD. van Gruting (ed.), Elsevier Science BV Norman, D.A. (1988). The Psychology of Everyday Things. Basic Books, New York. Norman, D.A. (1993). Toward Human-Centered Technology Review ,July issue, 47-53.

Design.

Oxenburgh, .\1. (1991). Increasing Productiz•i()• and Profit Througb Healtb and Safe~y. CCH International,Australia.

Cust, G. and Pearson,).C.G. Mair,A. (19...,2).The prevalence of low back pain in nurses.Int.Nursing Reuieu•, 19:169-179.

Pheasant, S.T. (1995).A foreseeable risk of injury. Contemporary Ergonomics, 1995, edited by S.A. Robertson, Taylor and Francis, London.

Dossick, S. (1994) Cser-interface copyrights: obstacles to innovation. IEEE Technology and Society Magazine, Fall Issue, 23-27.

Persson ) Ekberg K Linden M. 1993. Work organization, u•ork environment and tbe use of medical equipment: a suruey of tbe impact on quali()• and safe(v i\JBEC.)an:20-24.

Gardner-Bonneau, D.). 199.3. What is Iatrogenics and why don't ergonomists know? An interview with Dr Lowell Levin,

Posner, M. (1980). Orienting of attention. Quarterly journal of Experimental Psycbology, 32:2-25.

Bridger & Poluta

journal of Clinical Engineering • May/June 1998

187

Preece ,.f. ( 1992).A Guide to Usabilit)'. Open Cniversity,Addison Weslcv. Reason,J. ( 1990)./Iuman Error. Cambridge llniversity Press. Rey, P Bousquet, A. (199S). Compensation for occupational injuries and diseases: its effect upon prevention in the workplace. Ergonomics, .)8:-±~S-486. Shelton,].K. Mann-Janosi,J. (1992). l'nhealthy health care costs.] ofJ!edicine and PbilosojJby. 1-:--19. Stamper. .\l.T. (198-). Good health is not for sale. t)gonomics. .10:199-206. Thimble by, H. ( 1991 ). Can humans think? The Ergonomics Society Lecture, 1991. Ergonomics, .14:1269-128"''. YredenburghA.G., Saifer A. G., Cohen H.H. (199S). You're going to do what with that thing' Examining the causes of human error during use of a medical device. Erg(momics in Design. April issue, 16-20. Waters, T.R. Putz-Anderson. Y. Garg. A. Fine, L.J. ( 199.1 ). Revised NIOSH equation for the design and evaluation of manual lifting tasks. Ergonomics . .)6:749-7~6 .. Wisner, A. (1989). Variety of physical characteristics in industrially developing countries - ergonomic consider.Jtions. Int.] of Industrial El~r_;olwmics. -±: 117-1_18.

BIOGRAPHIES Robert S. Bridger is head of the Ergonomics Group at the Department of Biomedical Engineering, l-niversity of Cape Town/Groote Schuur Hospital. He has BSc(Hons) and MSc (Ergonomics) degrees from the l-niversity of London and l niversity College London. respectively. From 1981 to 1984 he was a senior research oftlcer at the Chamber of Mines Research Organization in Johannesburg, where he worked on the evalua-

188


tion of prototype machines for the mechanization of gold mining operations. Since 1984, he has been a member of the biomedical engineering department at UCT. where he lectures on ergonomics and human-machine interaction. He was awarded a PhD from l 'CT in 1991 for his work on the functional anatomy of working posture. His current research interests are in cost-based methods in ergonomics and in occupational low-back injury. Apart from lecturing and research he acts as an ergonomics consultant to a wide r.1nge of local companies. and is author of the textbook "Introduction to Ergonomics" published by McGraw-Hill Inc (New York), 1995 . Mladen A. Poluta is Coordinator of the Health Technology Management Progr.1mme and Head of the Bioelectric Signal Processing Group at the Department of Biomedical Engineering, University of Cape Town/Groote Schuur Hospital. He has a BSc (Eng) degree from the llniversity of the Witwatersrand and is currently working on his PhD. After four years in the medical equipment industry. he joined the .\ledical Physics Department at the Johannesburg Hospital. In 198~, he moved to the Department of Biomedical Engineering at UCT/GSH, where he has lectured on medical instrumentation, medical equipment engineering and management, and biomedical signal processing. His current research interests include the development of appropriate and sustainable health technology management interventions for developing countries. and the Sub-Saharan region in particular. He is currently serving on the executive councils of the International Federation for Medical and Biological Engineering, the African Federation for Technology in Healthcare and the South African national biomedical and clinical engineering societies.

Address all correspondence and reprint requests to the authors at: Department of Biomedical Engineering, UCT .\1edical School, Observatory 792S, South Africa: Fax: + 2~ 21 448 .1291: email: bridger@anat. uct.ac.za.

Bridger & Poluta

I