In its Centennial Vision, the American Occupational Therapy Association (AOTA)

Transcription

1 Potentially Problematic Postures During Work Site Keyboard Use Nancy A. Baker, Mark Redfern KEY WORDS computers posture task performance and analysis workplace OBJECTIVE. We described the frequency and distribution of keyboard users potentially risky postural behaviors. METHOD. Forty-three participants keyboard postural behaviors were rated with the Keyboard Personal Computer Style instrument (Baker & Redfern, 2005). The frequency and distribution of keyboard postural behaviors and the associations and differences between the right and left sides were assessed. RESULTS. Generally, each static posture had one criterion that occurred frequently, whereas dynamic postures were distributed throughout the criteria. The right and left sides were significantly associated for shoulder flexion, elbow flexion, hand displacement, wrist extension, forearm rotation, isolated fifth digit, metacarpophalangeal hyperextension, and wrist support use and significantly different for hand displacement, isolated thumb, number of digits, and metacarpophalangeal hyperextension. CONCLUSION. Potentially problematic postural behaviors are common among keyboard users. Our results suggest that occupational therapists must systematically assess postures on both the right and the left sides to develop the most effective intervention strategies. Baker, N. A., & Redfern, M. (2009). Potentially problematic postures during work site keyboard use. American Journal of upational Therapy, 63, Nancy A. Baker, ScD, OTR/L, is Assistant Professor, Department of upational Therapy, 5012 Forbes Tower, University of Pittsburgh, Pittsburgh, PA 15260; Mark Redfern, PhD, is Professor, Department of Bioengineering, University of Pittsburgh. In its Centennial Vision, the American upational Therapy Association (AOTA) identified work and industry, particularly work injury prevention, as one of the important emerging areas of practice in occupational therapy (Baum, 2006). One aspect of industrial injury prevention is ergonomics. Ergonomics focuses on evaluating the fit between the work environment and the worker with the goal of adapting the environment to prevent injury and to facilitate work participation at the highest possible level. One work environment that has been identified as having the potential to cause injury is the computer workstation (Wahlstrom, 2005). upational therapists are frequently asked to observe keyboard users to assess the effect that their computer workstations have on discomfort and injury and to identify workstation interventions to reduce or prevent musculoskeletal disorders of the upper extremity (MSD UE). Unfortunately, only limited information exists concerning the distribution of potentially risky postural behaviors assumed by keyboard users during worksite keyboard use, and no means are available to accurately document the potentially risky postural behaviors at the work site. Although many studies have focused on the association between posture and MSD UE, they have had only limited consensus concerning which postures place a keyboard user most at risk for MSD UE. Researchers, however, have developed some general information on potential risky postural behaviors in computer users. In general, computer users with a head tilt <20 (Baker, Sussman, & Redfern, 2008; Hunting, Laubli, & Grandjean, 1981; Lueder, 1996; McAtamney & Corlett, 1993; Szeto, Straker, & Raine, 2002) and shoulder flexion <20 (Marcus et al., 386 July/August 2009, Volume 63, Number 4

2 2002; McAtamney & Corlett, 1993) appear to be at less risk for MSD UE than those with greater neck flexion postures. Some disagreement exists in the literature concerning a safe elbow angle; some studies have reported that an elbow angle between 80 and 120 is best (Faucett & Rempel, 1994; McAtamney & Corlett, 1993), whereas a recent longitudinal study by Marcus et al. (2002) reported that an elbow angle in excess of 120 was protective. The best wrist extension posture is also an area of debate; some researchers have advocated a neutral wrist (Hedge & Powers, 1995; Simoneau, Marklin, & Berman, 2003), whereas others commonly used postural assessment tools, such as the Rapid Upper Limb Assessment (McAtamney & Corlett, 1993) and the Strain Index (Moore & Garg, 1995), suggest that a cut-off point of 15 of wrist extension should be used to promote low-risk wrist extension postures. Wrist ulnar deviation of 20 or more has been associated with MSD UE in computer users (Hunting et al., 1981). Optimal finger positions during keyboard use have not been well researched. In their study on piano playing, Harding, Brandt, and Hillberry (1993) reported that a curved finger with a metacarpophalangeal (MCP) flexion angle of between 40 and 65 put the least stress on the MCP joint. They reported that a proximal interphalangeal (PIP) flexion angle of approximately 50 and distal interphalangeal (DIP) flexion angle of approximately 25 caused the least amount of force on the tendons and joints of the hand. Because piano playing is biomechanically similar to keyboarding, this research suggests that people using a keyboard should adopt a similar posture. Other keyboard postural phenomena that have been associated with MSD UE are the tendency for some keyboard operators to maintain their fifth finger, thumb, or both in hyperextension while typing (Pascarelli & Kella, 1993; Rose, 1991). Additional considerations in keyboard work are the use of an elbow, forearm, or wrist support during keyboarding. Research on the utility of supporting the arm, wrist, or both during keyboarding is equivocal. Studies have suggested that although arm and wrist supports can decrease electromyography outputs at the shoulder (Aaras, Ro, Fostervold, Thoresen, & Larsen, 1997; Fernstrom, Ericson, & Malker, 1994; Visser, De Korte, Van der Kraan, & Kuijer, 2000), they do not necessarily improve posture (Hedge & Powers, 1995) and may not reduce the incidence of MSD UE (Bergquist, Wolgast, Nilsson, & Voss, 1995; Marcus et al., 2002; Rempel et al., 2006). Studies have also reported differences between rightand left-side postural behaviors. In particular, differences appear to occur in wrist and hand postures. For example, right- and left-wrist extension and ulnar deviation have been reported to be different (Simoneau, Marklin, & Monroe, 1999), although the differences have not always been significant (Serina, Tal, & Rempel, 1999). Baker, Cham, Cidboy, Cook, and Redfern (2007) reported significant differences between the right and left hands for hand displacement and for thumb postures. Thus, keyboard users may have a greater potential risk for MSD UE on one side than on the other. Much of the current literature on keyboard use has been done for the engineering community. Several studies have described the mean postures of keyboard users (Baker et al., 2007; Baker & Cidboy, 2006; Marklin, Simoneau, & Monroe, 1999; Rose, 1991; Simoneau et al., 1999; Sommerich, Marras, & Parnianpour, 1996; Zecevic, Miller, & Harburn, 2000) and have identified that on average, keyboard users position themselves in neutral, nonrisky postures. Studies have also suggested, however, that although little variability exists within the postural behaviors of a single keyboard user (Baker et al., 2007; Ortiz, Marcus, Gerr, Jones, & Cohen, 1997), a great deal of variability exists among keyboard users (Psihogios, Sommerich, Mirka, & Moon, 2001; Simoneau et al., 1999; Sommerich et al., 1996). This variability suggests that some keyboard users may assume potentially risky postures, but the degree to which potentially risky postures are assumed may not be easily identified in studies that use mean postures rather than frequency and distribution of postures as the outcome measure. Although research examining keyboard users mean postures has provided excellent information on general keyboard kinematics, they may not provide clinically applicable information about the frequency or distribution of postural behavior during keyboard use. Other aspects of keyboard kinematics studies also make them less clinically useful. Keyboard kinematics studies have generally used direct methods to obtain data (Li & Buckle, 1999). Direct methods, which measure kinematics by applying a measuring device, such as an electric goniometer or motion analysis device, to keyboard users extremities while they are typing, require a great deal of highly technical equipment. This requirement is problematic for occupational therapists for two reasons: (1) Most occupational therapists do not have access to this type of equipment, making it not feasible for them to use these methods to obtain data and, therefore, to directly compare their results to the literature, and (2) studies using direct methods are almost always completed in a laboratory, where the equipment can be easily set up and used. Laboratory setups tend to be standardized to participants anthropometrics and are therefore ideal workstation setups promoting neutral postures. Even if participants are instructed to set up the laboratory workstation to match their own workstation, few appear willing or able to manipulate The American Journal of upational Therapy 387

3 the laboratory office equipment to mimic their work site setup, which makes the results of these studies less applicable to real-world interpretations. The information generated by direct methods, therefore, is often not feasible or useful for occupational therapists attempting to evaluate and intervene with clients. Observational methods those that use observation of participants without the direct application of measuring equipment provide more familiar and clinically applicable results. Only one study (Pascarelli & Kella, 1993) has used observational methods to provide a description of the frequency of potentially risky postural behaviors in a group of keyboard users with MSD UE. Although this study provided insight into keyboard use, the observational methods used to collect the data were not from a valid and reliable rating instrument but were more from the researchers general observations. upational therapists need to have a systematic method to evaluate the whole body during keyboard use so that interventions can comprehensively address all risk factors. Until recently, no reliable and valid observational method has existed to assess and document keyboarding styles, and there has been a need for both a method to measure keyboarding style and data about the frequency and distribution of postural behaviors of keyboard users that can easily be translated to practice by occupational therapists. This study is the first to use a valid and reliable observational method, the Keyboard Personal Computer Style (K PeCS) instrument (Baker & Redfern, 2005), to measure postural behaviors hypothesized to be potential risk factors for MSD UE for keyboard users. Our purpose in this study was to describe the frequency and distribution of keyboard users postural behaviors while they worked at their own workstations. To further examine keyboard users postural behaviors, we calculated the associations and differences between their right and left sides during keyboard use. In addition to describing keyboarding postural behaviors, we provide readers with information about a reliable and valid observational method for measuring keyboarding postural behaviors. Method This descriptive correlational study involved videotaping participants as they used their work site computers. We used the videotapes to identify their keyboard-related postural behaviors. The study was approved by the University of Pittsburgh Institutional Review Board, and we obtained informed consent from all participants before study participation. Participants We recruited 43 participants from among University of Pittsburgh students, faculty, and staff. Participants had to be between ages 18 and 65, of either sex, and computer keyboard users. Participants were excluded if they had a fracture or traumatic injury that prevented them from using bilateral upper extremities. We did not exclude them, however, if they experienced musculoskeletal discomfort or had a diagnosed MSD UE. Participants were primarily female (84%) and White (84%) and had a mean age of 45.7 (±8.7) years. They had been using a computer at work for a mean of 14.2 (±6.6) years. Mean computer use was 6.3 (±2.0) hr per day and ranged from 3 to 12 hr of daily use. Participants reported that on average, they used their keyboard 60% of the time and their mouse 40% of the time. More than half reported having taken a touch-typing course (58%). When asked to rate their typing speed, 30% reported themselves as fast typists (>60 words per minute [wpm]), 40% reported themselves as moderate typists (40 60 wpm), 18% reported themselves as slow typists (<40 wpm), and 12% were not sure of their speed. Instruments Participants keyboarding styles were rated using the K PeCS, a 19-item, criterion-based observational tool that documents the frequency of stereotypical postural behaviors during keyboarding (Baker & Redfern, 2005). The items of the K PeCS have been divided into three general categories: (1) items of static posture (Items 1, 3, 4, and 5); (2) items of dynamic posture (Items 8, 10, 11, 12, 13, 14, 15, 16, 17, 18, and 19); and (3) items of tension and force (Items 2, 6, 7, and 9; see Table 1). Most items are measured separately for the right and left sides. K PeCS items are generally rated in one of three ways (see Figures 1 5 for rating criteria). Some items are rated as yes no (i.e., the behavior does or does not occur; Items 2, 6, 7, 12, 18, and 19). Some items are rated using frequency ratings, the most common of which are never (0% of the time), occasionally (1% 30% of the time), frequently (31% 99% of the time), and always (100% of the time; Items 10, 11, 13, 14, and 17). Item 8 s frequency rating was slightly modified to occasionally (0% 30% of the time), often (31% 75% of the time), and most of the time (>75% of the time). Other items are rated as one of several possible ranges (Items 1, 3, 4, 5, and 9). Two items have unique ratings: Item 15, for which the rater counts the number of digits used to activate the keys, and Item 16, for which the rater identifies whether the right thumb, right index finger, or some other means is used to activate the space bar. The K PeCS has been shown to have good inter- and intrarater reliability (interrater intraclass correlation coefficient [ICC] =.90, p <.001, and intrarater ICC =.92, p <.001; Baker, Cook, & Redfern, 2009). Most individual items on the K PeCS have good to excellent reliability, although 388 July/August 2009, Volume 63, Number 4

4 Table 1. Items and Outcomes Being Measured by the Keyboard Personal Computer Style Item Outcome Being Measured Items of static posture 1. Torso angle Generally, what is the angle of the keyboard user s torso to the horizontal plane? 3. Neck flexion angle Generally, what is the displacement angle and position of the head? 4. Shoulder flexion angle Generally, what is the flexion angle of the shoulders? 5. Elbow flexion angle Generally, what is the angle of the elbows? Items of dynamic posture 8. Hand displacement Does the keyboard user move his or her hands while typing? 10. Wrist ulnar angle Does the keyboard user exceed 20º of ulnar deviation? 11. Wrist extension angle Does the keyboard user exceed 15º of wrist extension? 12. Forearm rotation Does the keyboard user ever rotate his or her forearm (increase pronation or supination)? 13. Isolated fifth digit Does the keyboard user isolate the fifth digit? 14. Isolated thumb Does the keyboard user isolate the thumb? 15. Number of digits to type How many digits does the keyboard user use to strike the keys? 16. Space bar activation What finger does the keyboard user use to strike the space bar? 17. MCP hyperextension Does the keyboard user hyperextend the MCP joints? 18. PIP DIP curve Are the keyboard user s PIP DIP joints generally curved (>25 ) or generally straight (<25 )? 19. DIP hypermobility Do the DIP joints ever collapse when the fingers strike the keys (hypermobility)? Items of tension and force 2. Back rest use Does the keyboard user rest at least two-thirds of the back against the back rest while using the computer? 6. Forearm support use Does the keyboard user support his or her forearms or elbows on an arm rest or table? 7. Wrist support use Does the keyboard user support his or her wrist(s) on the wrist pad or table? 9. Force Generally, what kind of force does the keyboard user use to strike the keys? Note. MCP = metacarpophalangeal; PIP = proximal interphalangeal; DIP = distal interphalangeal. five items fell below an ICC of.75. The K PeCS has also been found to have content- and criterion-based validity (Baker et al., 2009; Baker & Redfern, 2005). Procedures Forty-three participants were videotaped in their workplace at their own computer workstation while typing a standard document that was long enough to require 10 min of continuous typing. Each workstation was unique to the individual participant. No equipment was constant in all workstations, and no attempt was made to set up the work environment to reduce potentially risky postures, because the purpose of the study was to describe the frequency and distribution of postural behaviors of typical keyboard users while they used their workstations as usual. The workers hands were recorded using three digital video cameras for approximately 10 min; one camera focused on a lateral view of the right hand, the second camera focused on a lateral view of the left hand, and the third camera focused on both hands from overhead. Additionally, still camera photographs of the full left and right body were taken to help in rating Items 1 through 6 (see Table 1). Each participant s recordings were processed into a 3- min clip consisting of 1 min each of overhead, right, and left data. Data from approximately the same time frame were used for each view so that the clip presented differing aspects of the same minute of time. The same rater, experienced in using the K PeCS, rated the clips and the still photographs twice. We calculated intrarater reliability using ICCs and obtained good to excellent reliability. To improve accuracy, we assessed all data between the two rating sessions for agreement. Where there was disagreement between the ratings for an item, we reassessed the participant s video recordings and completed a final determination of the correct rating. In some cases, it was impossible to obtain both right and left still photographs because of space constraints; most workstations allowed access to at least one side, but only 13 allowed sufficient access to obtain complete data for both right and left full-body postures. When we did not have enough space to obtain a photograph, we obtained live K PeCS ratings while the keyboard user was videotaped. These live ratings were all completed by one expert rater. Data Analysis We statistically analyzed data using SPSS 14.0 (SPSS, Inc., Chicago). Descriptive frequency statistics for the K PeCS were completed to provide information on the distribution and frequency of each item. Where appropriate, we compared The American Journal of upational Therapy 389

5 K PeCS Item Elbow flex angle Shld flex angle Neck flex angle Torso position Leaning forward Upright Reclined > <10 > > <79 Single Right Left % of Participants Rated for Each Criterion Figure 1. The distribution of ratings on the Keyboard Personal Computer Style (K PeCS) for items of static posture (Items 1, 3, 4, 5). Note. Shld = shoulder; flex = flexion. K PeCS Item Forearm rot Wrst ext >15 Uln dev >20 Hand displace Often Most of time No Yes Right Left % of Sample Rated for Each Criterion Figure 2. The distribution of the ratings of the Keyboard Personal Computer Style (K PeCS) for items of dynamic posture (Items 8, 10, 11, 12). Note. Displace = displacement; = occasionally; = frequently; Uln dev = ulnar deviation; Wrst ext = wrist extension; rot = rotation. 390 July/August 2009, Volume 63, Number 4

6 K PeCS Item Space act Isolated thumb Isolated 5th digit R 1st digit R 2nd digit Other Single Right Left 1 # of digits DIP hypermob No Yes % of Sample Rated for Each Criterion Figure 3. The distribution of ratings of the Keyboard Personal Computer Style (K PeCS) for items of dynamic posture (Items 13, 14, 15, 16, 19). Note. = occasionally; = frequently; Space act = spacebar activation; R = right; DIP hypermob = distal interphalangeal hypermobility. K PeCS Item PIP DIP position MCP hyperextension 5th 4th 3rd 5th 4th 3rd Straight Curved Straight Curved Straight Curved Right Left % of Participants Rated for Each Criterion Figure 4. The distribution of the ratings of the Keyboard Personal Computer Style (K PeCS) for items of dynamic posture (Items 17, 18). Note. = occasionally; = frequently; MCP = metacarpophalangeal; PIP DIP = proximal interphalangeal distal interphalangeal. The American Journal of upational Therapy 391

7 K PeCS Item Force Wrst support Forearm support Backrest use No Yes No No lft unsupp No rt unsupp Yes No During pauses Yes Low Mod Single Right Left High % of Sample Rated for Each Criterion Figure 5. The distribution of the ratings of the Keyboard Personal Computer Style (K PeCS) for items of tension and force (Items 2, 6, 7, 9). Note. No lft unsupp = no, left wrist unsupported; No rt unsupp = no, right wrist unsupported; Wrst = wrist; Mod = moderate. data for the right and left sides, using the nonparametric McNemar test of marginal homogeneity test (Portney & Watkins, 2008), which calculates a chi-square critical value, to examine differences in the distribution of the K PeCS score between the right and left sides and Spearman s rank correlations rho to examine the linear relationship between the right and left sides. We did not use Bonferroni s correction even though multiple tests were used; we were interested not in testing the universal hypothesis but rather in testing the differences within each test (Perneger, 1998). To help control for a Type I error, we set alpha at a more stringent.01. Results Most items were well populated: Ratings were distributed in all criteria (Figures 1 5). Five items had a criterion for which no participant ratings occurred three of the items had ratings for that item on one side of the body but not on the other. The items for which no participant was rated as performing a criterion were torso angle; supine; and number of digits used to type, both one digit and two digits. Items of static body postures had one criterion that occurred most frequently (Figure 1). Most keyboard users sat in an upright position (67%), with shoulder flexion between 0 and 20 (left [L] = 67%; right [R] = 63%) and elbow flexion between 80 and 120 (L = 74%; R = 72%). Neck flexion postures were generally either <10 (33%) or between 11 and 20 (40%). Dynamic postures rated using a frequency score were generally distributed throughout the range (Figures 2 4). Items rated as occurring or not occurring clustered around the no criterion: forearm rotation (no: L = 95%; R = 84%) and DIP hypermobility (no: L = 86%; R = 95%; Figures 2 and 3). Most keyboard users activated the spacebar with the right thumb (79%). All participants using at least three fingers on each hand during keyboarding, and most used four or five. Many participants maintained the third to fifth digits in positions of tension, such as MCP hyperextension or fifthdigit isolation (Figure 6a). Almost all participants isolated the fifth digit to some degree (Figure 6a; never isolated: L = 12%; R = 21% [Figure 3]). Variability also occurred in finger postures: About half the participants hyperextended the right and left fourth MCP joints and the right fifth MCP joints (Figure 6b) at least occasionally, whereas 72% of participants hyperextended the left MCP joints at least occasionally (Figure 4). Participants did not generally maintain digits in a straight posture, although 33% had a straight right fifth digit. Thumb isolation was less common than fifth-digit 392 July/August 2009, Volume 63, Number 4

8 Table 2. Linear Correlations and Difference Between the Right- and Left-Side Items During Keyboard Use A. B. Figure 6. Examples of common postures. A. Hyperextension of the fourth and fifth MCP joints (Item 17). B. Thumb and fifth-digit isolation (Items 13 and 14). Item Right to Left χ 2, Right vs. Left (p) 1. Torso angle 3. Neck flexion angle 4. Shoulder flexion angle.67 < Elbow flexion angle.68 < Hand displacement.58 < Wrist ulnar angle Wrist extension angle Forearm rotation Isolated fifth digit Isolated thumb Number of digits used to type Space bar activation 17. MCP hyperextension Second Third Fourth Fifth.56 <.001 < PIP DIP curve Second Third Fourth Fifth DIP hypermobility Back rest use 6. Forearm support use 7. Wrist support use.67 < Force Note. MCP = metacarpophalangeal; PIP = proximal interphalangeal; DIP = distal interphalangeal. = item is not measured for right and left, so no comparison was made. isolation, particularly on the right side (never: L = 56%; R = 81%; Figure 3). Some variability occurred across participants for items of tension and force that rated support use (backrest support, forearm support, and wrist support). Participants about equally did or did not use a backrest (yes: 47%; Figure 5). Although most participants generally did not use forearm support (70%), they were more likely to use a wrist rest on the left side than on the right (L = 58%; R = 47%). About half of the participants used moderate force while keyboarding (47%). In a secondary analysis, we examined the associations and differences between the right and left sides across participants. For items of static posture, we found large significant correlations between the right and left elbow and right and left shoulder postures (Table 2). We also found large significant correlations between the right and left sides for wrist support use (Table 2). For items of dynamic posture, we found significant moderate correlations between the right and left sides for hand displacement, wrist extension, forearm rotation, and isolated fifth digit and smaller but still significant correlations between the sides for MCP joint hyperextension for all but the third digit. The degree of association increased progressively from the fourth to fifth digits (Table 2). Associations between right and left side for wrist ulnar angle, isolated thumb, number of digits used, third-digit MCP hyperextension, third- and fourth-digits PIP DIP curve, and DIP hypermobility were all nonsignificant (Table 2). The right and left side differed significantly on hand displacement, number of digits, isolated thumb, and MCP joint hyperextension for the fourth and fifth digits (Table 2). The right hand tended to displace more frequently than the left, and more participants used five digits when keyboarding with the right hand. Participants were more likely to isolate the left thumb than the right and to hyperextend the left MCP joints more than the right. In summary, the keyboard users in this study generally assumed a neutral static posture of the neck, body, and arm but tended to assume a variety of dynamic postures along the spectrum of neutral postures to postures that were potential risk factors for MSD UE. About half of these keyboard users used supports (backrest, forearm, or wrist), and more ρ p The American Journal of upational Therapy 393

9 than half used minimal to moderate force to strike the keys. Postures between the right and left side were significantly correlated except for ulnar angle, isolated thumb, number of digits used, third-digit MCP hyperextension, third- and fourth-digits PIP DIP curve, and DIP hypermobility. Keyboard users were significantly different in their right and left postures for hand displacement, number of digits, isolated thumb, and MCP joint hyperextension for fourth and fifth digits. Discussion The distribution of postures identified by this method suggests that for some items, most participants worked in positions that placed them in neutral, nonrisky postures. These item distributions were concentrated in one criterion (Items 1, 4, 5, 6, 9, 12, 14, 15, 16, 18, and 19 in Table 1). The common measures were seated upright (Item 1), shoulder flexion <20 (Item 4), and elbow postures between 80 and 120 (Item 5). Most participants did not use any forearm support (Item 6). They used the keyboard with moderate force (Item 9), did not change their forearm rotation angle (Item 12), did not isolate their thumb (Item 14), and used four digits to activate the keys with the left hand and five digits to activate the keys with the right hand (Item 15); this last difference was probably because they activated the space bar with their right thumb (Item 16). They usually typed with curved PIP and DIP joints (Item 18) and rarely demonstrated hypermobility of the fifth digit (Item 19). The use of a backrest (Item 2) was one item that did not follow this pattern of data concentrating under one criterion for an item: This item was essentially equally distributed between those who did and did not use one. Other items (Items 7, 8, 10, 11, 15, 13, 14, and 17 in Table 1) demonstrated relatively equal distributions of participants engaging in that behavior. Examples are hand displacement (Item 8) or ulnar deviation past 20 (Item 10). These results suggest that these items, many of which have been identified as potential risk factors for MSD UE, are well distributed throughout the computer-using population during work site keyboard use. Many postures and behaviors were significantly positively associated between the right and left sides, a phenomenon described in other studies (Baker et al., 2007; Marklin et al., 1999; Simoneau et al., 1999). The only items for which we did not find significant associations were ulnar angle, isolated thumb, number of fingers used to type, thirddigit MCP hyperextension, third- and fourth-digit PIP DIP curve, and DIP hypermobility. The low association between the right and left side for isolated thumb was not unexpected (see Figure 6a for an example of thumb isolation). Of participants, 79% used their right thumb to activate the space bar and did not use the left thumb at all. Thus, the left and right thumbs are not engaged in symmetrical typing activities. The low association between the number of digits used between the right and left sides may also be related to space bar activation. Most people used four digits on the left hand (81%) and four or five digits on the right hand (35% and 58%, respectively). The association between the right and left sides was not significant for ulnar deviation. This nonsignificant association is supported by other literature on keyboard use that has reported that the right and left wrists are often asymmetrical for ulnar deviation (Simoneau et al., 1999). The lack of symmetry between the right and left for DIP hypermobility is probably representative of the rarity of this phenomenon. Only 5% of the sample demonstrated DIP hypermobility on the right, and only 14% of the sample demonstrated DIP hypermobility on the left. We not only found significant moderate positive associations between the right and left sides for many items but also found significant differences between the right and left sides for many of the same items. An item could be both significantly associated and significantly different because of the distribution of criteria for the right and left side of each item. To clarify this concept, we present examples of the right and left distribution of criteria for hand displacement and fourth- and fifth-digit MCP hyperextension in Table 3. In the distributions, most of the pairs fell along the diagonal of each table (the concordant squares that indicate a match between the left and right sides). Those pairs that fall off the diagonal (the discordant square that indicates no match between the left and right sides), however, have a different distribution on the right and left. For hand displacement, for example, the discordant pairs tend to cluster in the occasional column for the left side, whereas the discordant pairs tend to cluster in the most of the time row for the right. That most of the data are in the concordant squares makes the data significantly linearly associated, whereas the different distribution of the right and left sides in the discordant squares makes the data significantly different. Note that the significant differences in body postures and actions found in this study were not for the items related to the body or even motions of the wrists, as indicated by Simoneau and colleagues (1999). Instead, most differences occurred in the actions of the fingers and hands. For all participants, the keyboard was placed on a horizontal surface. To access the keyboard with the hands, the body and arms would have to assume a symmetrical posture. The difference noted between the right and left sides for hand use are probably related to the job tasks required for each hand during typing (Dennerlein & Johnson, 2006). The right hand is used not only to activate the letter keys but also, generally, to activate the space bar (Item 16) and the enter, backspace, 394 July/August 2009, Volume 63, Number 4

10 Table 3. Items on the Keyboard Personal Computer Style That Were Both Significantly Linearly Associated (Spearman s Rho) and Significantly Different (Chi-Square) Item and Rating Right-hand displacement Left-Hand Displacement asionally Often Most Total asionally Often Most Total Right fourth-digit MCP hyperextension Left Fourth-Digit MCP Hyperextension asionally uently Total asionally uently Total Right fifth-digit MCP hyperextension Left Fifth-Digit MCP Hyperextension asionally uently Total asionally uently Total Note. Both ps.01. MCP = metacarpophalangeal. and delete keys. It therefore tends to have a greater amount of hand displacement (Item 8). Because the right side is used to activate the space bar, more people use all five fingers on the right hand when using the keyboard, and fewer people isolate their right thumb as they are positioning it over the space bar in preparation to strike it. The differences in the right and left hands for MCP hyperextension is less easy to explain because it is not clear why people hyperextend their MCP joints while using a keyboard (Figure 6b). Some degree of MCP hyperextension appears to be a common action during keyboard use (Figure 4). Rose (1991) suggested that people using a flat keyboard must extend their third, fourth, and fifth digits up to 75% of their range to position these digits on the same plane as the thumb and therefore in line with the keyboard. This need to be on the same plane for keyboarding tasks, however, does not explain why the two hands are not symmetrical for MCP hyperextension. Further study of this postural behavior is important to understand why people hyperextend their MCP joints and why they may hyperextend more frequently on the left. The K PeCS instrument appears to be a sensitive tool with which to document computer keyboard style. On average, an assessment could be completed in <10 min, and the K PeCS was easy to use for both live and videotaped images. The data generated by the K PeCS have been shown to be reliable and valid (Baker et al., 2009), to be easily interpreted by occupational therapists, and to provide clinically relevant results that could help occupational therapists select intervention strategies for clients at risk. Limitations Because this study is descriptive of a relatively small sample of keyboard users, it may not generalize well to other populations outside of this university environment. We cannot suggest that any of the postural behaviors cause, or even aggravate, MSD UE only that in this sample, certain postural behaviors were more common than others and that the epidemiological literature has identified those postural behaviors as potential risk factors for MSD UE. Because the participants were examined at only a single time point, we cannot determine whether their postural behaviors changed over time as a result of factors such as fatigue. In addition, we did not analyze the data to identify underlying causes, such as the environmental setup or participants anthropometrics of any postural behavior. We could not obtain photographs of both the right and left sides of many participants to assess static body postures The American Journal of upational Therapy 395

11 of both the right and the left shoulder and arms. Static posture ratings for work sites in which still photographs could not be taken were completed live, and their reliability was not checked. Although the K PeCS has been found to be a reliable tool, certain items have lower intrarater reliability. Force, isolated thumb, wrist ulnar angle, and PIP DIP angles have demonstrated only moderate rather than good to excellent intrarater reliability. We attempted to control for this lower reliability by double-rating items and then coming to a consensus for items for which ratings did not match. Implications for Practice The participants in this study commonly demonstrated potentially risky keyboard postural behaviors, particularly in the hands and fingers. On the basis of these results, we recommend that occupational therapists systematically assess body, arm, wrist, and hand postures on both the left and the right sides to develop the most effective intervention for each client. Other postural behaviors, such as extremes of posture for the neck and torso, are uncommon, suggesting that occupational therapists must be vigilant to identify whether these occur. Although most research has focused on body and wrist postures, this research indicates that potentially risky finger and hand postural behaviors occur frequently. upational therapists evaluating keyboard users should examine not only body postures but also wrist and finger postures to determine whether and how often a client assumes a potentially risky postural behavior. We found that keyboard users tended to be symmetrical in their body and arm postures and less symmetrical in how they used their hands and fingers to operate the keyboard. upational therapists need to assess both the left and the right sides of the body while evaluating a client using a keyboard to determine variations in performance by side. Performances that differ on the right and left sides may require different interventions. The AOTA Centennial Vision explicitly identifies the need for a science-driven... and evidence-based profession (Baum, 2006, p. 610). Until now, occupational therapists who have engaged in ergonomic assessment and interventions for computer keyboard workstations have had no scientific method to assess keyboard use and no information about the frequency and distribution of potentially risky postural behaviors in keyboard users. This study provides occupational therapists with information about a clinically useful tool that can help document the occurrence of types of postural behaviors before implementing workstation redesign. Through accurate identification of client-specific problem areas, the occupational therapist can more efficiently and effectively implement interventions that will reduce potential risk factors. In addition, occupational therapists can reassess the success of their interventions by reevaluating keyboard users postural behaviors after intervention has been completed. The use of a standardized observational method to ascertain and document keyboard style has the potential to improve the overall practice of office work site intervention. This study also provides occupational therapists with baseline information on the prevalence of certain types of postural behaviors during computer use, which will help them to be more sensitive to the variety of computer-related postures and more vigilant in identifying them for intervention. s Acknowledgments We acknowledge the support of National Institute for upational Safety and Health Grant K01 OH and the University of Pittsburgh Central Research Development Fund. We also thank Jack Dennerlein, Rakie Cham, Caroline Sommerich, Erin Hale, Norman Gustafson, and Emily Eckel for their assistance in developing the K PeCS. References Aaras, A., Ro, O., Fostervold, K. I., Thoresen, M., & Larsen, S. (1997). Postural load during VDU work: A comparison between various work postures. Ergonomics, 40, Baker, N. A., Cham, R., Cidboy, E., Cook, J., & Redfern, M. (2007). Kinematics of the fingers and hands during computer keyboard use. Clinical Biomechanics, 22, Baker, N. A., & Cidboy, E. (2006). The effect of three alternative keyboard designs on forearm pronation, wrist extension, and ulnar deviation: A meta-analysis. American Journal of upational Therapy, 60, Baker, N. A., Cook, J., & Redfern, M. (2009). Rater reliability and criterion validity of the Keyboard Personal Computer Style instrument (K PeCS). Applied Ergonomics, 40, Baker, N. A., & Redfern, M. (2005). Developing an observational instrument to evaluate personal computer keyboarding style. Applied Ergonomics, 36, Baker, N. A., Sussman, N. B., & Redfern, M. (2008). Discriminating between individuals with and without musculoskeletal disorders of the upper extremity by means of items related to computer keyboard use. Journal of upational Rehabilitation, 18, Baum, M. C. (2006). Presidential Address, 2006: Centennial challenges, millennium opportunities. American Journal of upational Therapy, 60, Bergquist, U., Wolgast, E., Nilsson, B., & Voss, M. (1995). The influence of VDT work on musculoskeletal disorders. Ergonomics, 36, Dennerlein, J. T., & Johnson, P. W. (2006). Different computer tasks affect the exposure of the upper extremity to biomechanical risk factors. Ergonomics, 49, July/August 2009, Volume 63, Number 4

12 Faucett, J., & Rempel, D. M. (1994). VDT-related musculoskeletal symptoms: Interactions between work posture and psychosocial work factors. American Journal of Industrial Medicine, 26, Fernstrom, E., Ericson, M. O., & Malker, H. (1994). Electromyographic activity during typewriter and keyboard use. Ergonomics, 37, Harding, D. C., Brandt, K. D., & Hillberry, B. M. (1993). Finger joint force minimization in pianists using optimization techniques. Journal of Biomechanics, 26, Hedge, A., & Powers, J. R. (1995). Wrist postures while keyboarding: Effects of a negative slope keyboard system and full motion forearm supports. Ergonomics, 38, Hunting, W., Laubli, T., & Grandjean, E. (1981). Postural and visual loads at VDT workplaces. I: Constrained postures. Ergonomics, 24, Li, G., & Buckle, P. (1999). Current techniques for assessing physical exposure to work-related musculoskeletal risks, with emphasis on posture-based methods. Ergonomics, 42, Lueder, R. (1996, August 8 9). A proposed RULA for computer users. Paper presented at the Proceedings of the Ergonomics Summer Workshop, San Francisco. Marcus, M., Gerr, F., Monteilh, C., Ortiz, D. J., Gentry, E., Cohen, S., et al. (2002). A prospective study of computer users: II. Postural risk factors for musculoskeletal symptoms and disorders. American Journal of Industrial Medicine, 41, Marklin, R. W., Simoneau, G. G., & Monroe, J. F. (1999). Wrist and forearm posture from typing on split and vertically inclined computer keyboards. Human Factors, 41, McAtamney, L., & Corlett, E. N. (1993). RULA: A survey method for investigation of work-related upper limb disorders. Applied Ergonomics, 24, Moore, J. S., & Garg, A. (1995). The Strain Index: A proposed method to analyze jobs for risk of distal upper extremity disorders. American Industrial Hygiene Association Journal, 56, Ortiz, D. J., Marcus, M., Gerr, F., Jones, W., & Cohen, S. (1997). Measurement variability in upper extremity posture among VDT users. Applied Ergonomics, 28, Pascarelli, E. F., & Kella, J. J. (1993). Soft-tissue injuries related to use of the computer keyboard. Journal of upational Medicine, 35, Perneger, T. V. (1998). What s wrong with Bonferroni adjustments? British Medical Journal, 316, Portney, L. G., & Watkins, M. P. (2008). Foundations of clinical research. Upper Saddle River, NJ: Pearson/Prentice Hall. Psihogios, J. P., Sommerich, C. M., Mirka, G. A., & Moon, S. D. (2001). A field evaluation of monitor placement effects in VDT users. Applied Ergonomics, 32, Rempel, D. M., Krause, N., Goldberg, R., Benner, D., Hudes, M., & Goldner, G. U. (2006). A randomised controlled trial evaluating the effects of two workstation interventions on upper body pain and incident musculoskeletal disorders among computer operators. upational and Environmental Medicine, 63, Rose, M. J. (1991). Keyboard operating posture and actuation force: Implications for muscle over-use. Applied Ergonomics, 22, Serina, E., Tal, R., & Rempel, D. (1999). Wrist and forearm postures and motions during typing. Ergonomics, 42, Simoneau, G. G., Marklin, R. W., & Berman, J. E. (2003). Effect of computer keyboard slope on wrist position and forearm electromyography of typists without musculoskeletal disorders. Physical Therapy, 83, Simoneau, G., Marklin, R., & Monroe, J. (1999). Wrist and forearm postures of users of conventional computer keyboards. Human Factors, 41, Sommerich, C. M., Marras, W. S., & Parnianpour, M. (1996). A quantitative description of typing biomechanics. Journal of upational Rehabilitation, 6, Szeto, G. P., Straker, L., & Raine, S. (2002). A field comparison of neck and shoulder postures in symptomatic and asymptomatic office workers. Applied Ergonomics, 33, Visser, B., De Korte, E., Van der Kraan, I., & Kuijer, P. (2000). The effect of arm and wrist supports on the load of the upper extremity during VDU work. Clinical Biomechanics, 15(Suppl. 1), S34 S38. Wahlstrom, J. (2005). Ergonomics, musculoskeletal disorders, and computer work. upational Medicine, 55, Zecevic, A., Miller, D. I., & Harburn, K. (2000). An evaluation of the ergonomics of three computer keyboards. Ergonomics, 43, The American Journal of upational Therapy 397