The Effect of Three Alternative Keyboard Designs on ... The Effect of Three Alternative Keyboard Designs on Forearm Pronation, ... ed pressure inside the carpal tunnel and to strain

40 January/February 2006, Volume 60, Number 1

The Effect of Three Alternative Keyboard Designs on Forearm Pronation, Wrist Extension, and Ulnar Deviation:A Meta-Analysis

Nancy A. Baker, Erin L. Cidboy

The growth of computer keyboard use in the workplace is believed to be one important determinant of theincreased prevalence of work-related musculoskeletal disorders of the upper extremity (MSD-UE). One possi-ble contributing factor to the development of MSD-UE is the flat standard keyboard, which places the forearmand wrist in biomechanically awkward postures. This meta-analysis examines the efficacy of three alternativekeyboard designs, adjustable slope (AS), split fixed-angle (FA), and adjustable open-tented (AT), in reducingforearm pronation, wrist extension, and ulnar deviation. Analyses of pooled effect size from six studies indi-cated that the AT had a large effect on pronation (r = 0.85) and ulnar deviation whereas the FA had a large effectonly on ulnar deviation (r = 0.79). The AS was found to have a large effect (r = 0.66) on wrist extension. TheFA had a moderate effect on pronation (r = 0.33) and wrist extension (r = 0.30). None of these keyboards werefound to have a significant effect on all three postures. This meta-analysis has implications for clinicians byproviding objective information that may assist with the selection of an alternative keyboard that best reducesan identified problematic posture.

Baker, N. A., & Cidboy, E. L. (2006). The effect of three alternative keyboard designs on forearm pronation, wrist extension,and ulnar deviation: A meta-analysis. American Journal of Occupational Therapy, 60, 40–49.

In the past 2 decades, the United States has experienced rapid growth in computeruse in the workplace. Use of computers in the workplace has increased from 67.4

million (36%) in 1993 to 92.2 million (47.1%) in 1997 (Newburger, 1999). In2001, the Bureau of Labor Statistics reported that 72.3 million people at work useda computer, accounting for 53.5% of total employment (Bureau of Labor Statistics,2002a). Studies have suggested that there is an increased incidence of muscu-loskeletal disorders of the upper extremity (MSD-UE) in computer users (Bergqvist,Wolgast, Nilsson, & Voss, 1995; Gerr et al., 2002; Hales et al., 1994) although thenational incidence of MSD-UE among computer users is unknown. However, theoverall incidence of MSD-UE has been steadily rising and has been estimated to be70% of all occupational illnesses (Bureau of Labor Statistics, 2002b). This suggeststhat computer-related MSD-UE is likely to be on the rise as well.

This has significant connotations for the occupational therapy professional.The ability to use computers without discomfort or injury is important not onlyfor the completion of workplace tasks but also in meaningful leisure and dailytasks. One in eight adults used the Internet to perform job-related tasks at homeand 21% of children used the Internet to perform school-related tasks (Newburger,1999). Computer use can increase the independence of persons with disabilities.Persons who have difficulty leaving their home can use the computer to order gro-ceries, pay bills, shop, research health-related questions, obtain the news, and catchup with old friends and/or make new ones (Gordon, Capell, & Madhok, 2002;Kaye, 2000). Pain and irritation secondary to MSD-UE interferes with this taskcompletion, decreasing productivity in the workplace, limiting participation inleisure activity, and for some decreasing independence with daily task completion.

One hypothesized pathomechanism of MSD-UE, also referred to as cumula-tive trauma disorders or repetitive strain injuries, is repeated or sustained exertionsof the body while in biomechanically awkward postures. The postures of the fore-arm and wrist assumed during use of the flat, horizontal design of the conventional

Nancy A. Baker, ScD, OTR/L, is Assistant Professor,Department of Occupational Therapy, School of Health and Rehabilitation Sciences, University of Pittsburgh, 5023 Forbes Tower, Pittsburgh, Pennsylvania 15260;[email protected]

Erin L. Cidboy, MS, OTR/L, is Research Associate,Department of Occupational Therapy, School of Health and Rehabilitation Sciences, University of Pittsburgh,Pittsburgh, Pennsylvania.

Downloaded From: http://ajot.aota.org/pdfaccess.ashx?url=/data/journals/ajot/930164/ on 06/07/2018 Terms of Use: http://AOTA.org/terms

QWERTY keyboard is considered to be problematic bymany (Swanson, Galinsky, Cole, Pan, & Sauter, 1997;Zecevic, Miller, & Harburn, 2000). Simoneau and Marklin(2001) have reported that keyboard users maintain theirforearms in 60° of pronation and their wrists in 20° ofextension in order to keep their hands parallel to the surfaceof the flat conventional keyboard and ready to strike thekeys. Additionally, some keyboard users maintain theirhands close together on a standard keyboard while slightlyabducting their upper arms, leading to a posture exceeding20° of ulnar deviation (Serina, Tal, & Rempel, 1999).Marklin, Simoneau, and Monroe (1999) reported that stan-dard keyboard users assumed postures between 75.1° to52.6° forearm pronation, 0.4° to 25.8° ulnar deviation, and6.5° to 27.4° of wrist extension.

These awkward postures are believed to lead to elevat-ed pressure inside the carpal tunnel and to strain musclestructures (Luchetti, Schoenhuber, & Nathan, 1998).Researchers have reported that pressure inside the carpaltunnel is elevated when the wrist is extended (Serina et al.,1999; Simoneau & Marklin, 2001), ulnarly deviated(Serina et al.), or the forearm is pronated past 45° (Rempel,Bach, Levinsohn, & Gordon, 1996). For example, pressureinside the carpal tunnel has been demonstrated to increasefrom an average of 29.8 mm Hg at 15° of wrist extensionto an average of 40.3 mm Hg at 30° of extension (Marklinet al.). When the wrist is ulnarly deviated to 10° or more,forces exerted by the carpal bones and carpal ligamentsagainst the flexor tendons increase, contributing to inflam-mation with repetitive use (Marklin et al., 1999). Theinflammation resulting from these deviated wrist posturescan cause increased pressure on nerves, blood vessels, andtendons passing through the carpal tunnel (Smith et al.,1998). This can cause ischemia and flexor tendon synovialtissue edema, which may lead to pressure on the mediannerve, nerve block, and to development of carpal tunnelsyndrome and other MSD-UE (Rempel et al., 1996;Werner & Andary, 2002).

In recent years, alterations to the standard keyboardthat reduce if not eliminate the need to assume awkwardpostures of the hand and forearms have been developed andare believed to increase comfort and decrease musculoskele-tal injury (Hedge, Morimoto, & McCrobie, 1999; Nelson,Treaster, & Marras, 2000; Simoneau & Marklin, 2001),although no studies, to date, have shown that they doreduce the occurrence of MSD-UE. These proposed designsinclude splitting the keyboard in half (Smith et al., 1998),laterally tenting keyboard halves (Zecevic et al., 2000), andaltering the tilt in the positive or negative direction (Hedge& Powers, 1995; Simoneau & Marklin, 2001). Single stud-ies using a randomized control design have been conducted

to describe the effect of alternative keyboards designs onawkward postures (Swanson et al., 1997; Tittiranonda,Rempel, Armstrong, & Burastero, 1999; Zecevic et al.).Although the results of these studies have reported favorableresults, no systematic review of these studies has been com-pleted to compare and contrast the keyboards efficacy inreducing different wrist postures. This meta-analysis ofthese studies provides therapists with a clear, objectivedescription of the overall results and clarifies the differencesand similarities between the keyboards.

The effect of alternative keyboards on posture is animportant clinical issue when determining if a client wouldbenefit from an alternate keyboard and which alternativekeyboard to prescribe. The object of this review was todetermine whether alternative keyboard designs promotebetter postures of the forearms and wrists during computerkeyboard use. This meta-analysis provides clinicians withinformation about the different effect that three commonalternative keyboards have on awkward postures of the wristand forearm, thus assisting them in making appropriatekeyboard recommendations to clients.

Method Criteria for Selecting Studies for Review

This meta-analysis examined the effect of alternate key-board use on the postures of clients without MSD-UE orother disorders of the upper extremity. Therefore, studiesincluding persons with MSD-UE, acute trauma, inflamma-tory disease, neurological disease, or neoplasm were exclud-ed. Studies were selected whose target population includedat least 10 total asymptomatic adult computer users. Bothlaboratory and clinically based randomized controlled trialscomparing a standard flat keyboard with (a) a split fixed-angle keyboard design (FA); (b) an adjustable open-tentedkeyboard (AT); and (c) an adjustable slope keyboard (AS),adjusted in the positive (keys tilted toward the user) andnegative (keys tilted away from the user) directions wereincluded (see Table 1). Outcome measures investigatedincluded ulnar deviation angle, wrist extension angle, andforearm pronation angle.

Search Strategy for Identification of Studies

An extensive search of MEDLINE (1966 to present),PEDro—the Physiotherapy Evidence Database, CINAHL(1982 to present), ISI Web of Knowledge, PubMed,OTSeeker, and Cochrane Database of Systematic Reviewswas undertaken. Reference lists of retrieved articles werealso searched for additional articles as well as a personal col-lection of articles. The literature search strategy was not limited to studies published in peer-reviewed journals. An

The American Journal of Occupational Therapy 41


in-depth search via Internet search engines to identifyunpublished work and/or ongoing projects of relevancefrom credible resources was used to identify if there wereunpublished studies that might be included. Main searchkey words entered for computerized database searchesincluded cumulative trauma disorders, repetitive straininjury, ergonomics, keyboards, posture, musculoskeletaldisorders and injuries, fatigue, pain, visual display units,computer terminal, and human engineering.

Methods of the Review

Titles and abstracts were identified by electronic databasesearch, and a reviewer scrutinized hand searches. Studiesthought relevant by the reviewer were retrieved for full-textreview. A total of 17 articles were identified and retrieved.Each article was examined and those that met the followinginclusion criteria were included: (1) asymptomatic samplepopulation greater than or equal to 10; (2) randomizationof participants to keyboard condition (standard or alterna-tive) or randomized order of keyboard presentation; (3)clear description of types of keyboard designs included; (4)comparison of the effect of an AS keyboard, FA keyboard,and/or AT keyboard to a standard flat keyboard for forearmpronation, wrist ulnar deviation, and/or wrist extension;and (5) adequate information for the calculation of an effectsize (r ). Ten studies met these inclusion criteria.

Methodological Quality Assessment

Studies meeting the above inclusion criteria were furtherassessed for methodological quality. Each article was ratedusing a criterion-based checklist by the second author(EC). Areas addressed on this checklist were those identi-fied in the literature on reporting quality. They includedoverall study design, sample descriptions such as size, inclu-sion or exclusion criteria, retention and equality of groups,clear definitions of instruments and interventions, andappropriate and detailed statistical analyses. Study design

was defined using the hierarchy proposed by Moore,McQuay, and Gray (1995) for critiquing quality of evi-dence. Descriptions of the sample relate to internal andexternal validity issues, and are included in many checklists(Greenhalgh, 1997; Maher, Sherrington, Herbert, Moseley,& Elkins, 2003; Moher, Schulz, & Altman, 2001; Sackett,Straus, Richardson, Rosenberg, & Haynes, 2000; Verhagenet al., 1998). Although not all checklists on research quali-ty examine the clarity of definition of outcome measuresand interventions, the reliability of measures and the abili-ty to replicate therapy interventions provides informationon the quality of the intervention as well as the reliabilityand validity of the measurement tools (Law et al., 1998).Statistical analyses have also been identified as an area ofconcern, particularly in the effort to interpret results. Theinclusion of central tendency and variability data in partic-ular has been highlighted as important to article quality(Maher et al., 2003; Verhagen et al.).

Items on the list were scored ranging from zero to 2depending on how well the article met the methodologicalcriteria. The maximum score that could be achieved on theassessment was a 22. The criteria used in this methodolog-ical quality review were as follows: (1) clear description ofpurpose or aim of study; (2) explanation of pre-interventionscreening response such as the number of returned mailedquestionnaires compared to total questionnaires mailed; (3)clear description of statistical analysis; (4) randomized con-trolled study versus nonrandomized control study; (5) clearexclusion/inclusion criteria; (6) greater than 20 partici-pants; (7) groups equated at baseline; (8) a clear descriptionof keyboards examined; (9) blinding of participants and/orassessors; (10) attrition at follow-up(s); and (11) a clear def-inition and explanation of investigated outcome measures.Studies with a quality assessment score of 17 or higher wereincluded in the meta-analysis (see Table 2). A rating of 17ensured that the article included greater than 75% of thespecified criteria considered necessary for a high-quality


Table 1. Description of Keyboard DimensionsLateral Opening Frontal

inclination angle tiltType of keyboard Abbreviation Description (degrees) (degrees) (degrees)

Standard none Flat horizontally oriented keyboard with QWERTY arrangement. 0 0 0–15Split fixed-angle FA Keyboard is split down the “center.” The “center” keys are fixed

higher than the “outside” keys and the top row keys are fixed closer together at the center than the bottom row keys. 10 25 0

Adjustable open-tented AT Keyboard has a split down the “center” and is constructed to allow the height of the “center” keys to adjust upward and the bottom row keys to adjust outward (“tented”). 42 15 0

Adjustable slope AS Keyboard can tilt toward (+ direction) or away from (- direction) the keyboarder. 0 0 (-)15–(+)15

Note. Lateral inclination—tilt angle in the coronal plane corresponding to pronation/supination movement of forearm; opening angle—split angle between the twohalves of a keyboard in the transverse plane corresponding to ulnar and radial deviation movement of the wrist; frontal tilt—angle in the sagittal plane correspond-ing to flexion and extension movement of the wrist.



research article. Initially 10 studies were identified for fur-ther review, after quality assessment, 6 studies remained (seereference list for studies excluded after quality assessment).This provided an overall population of 182 participants.

Data Analysis Descriptive statistics for all the studies were extracted. Theseincluded both article characteristics (e.g., date of publica-tion, language) and participant characteristics. In order tocompare the effects between alternative keyboards from theselected studies, the appropriate statistics (means, t statistic,F ratio, X 2) were extracted from the selected articles andconverted into correlation effect sizes (r ) (Hedge & Powers,1995; Marklin et al., 1999; Simoneau & Marklin, 2001;Smith et al., 1998; Szeto & Ng, 2000; Zecevic et al., 2000).The correlation effect size r was then transformed into itscorresponding Fisher-r. The Fisher-r for each correlationeffects size was used to calculate all mean effect sizes. Eachmean correlation effect size was adjusted based on the sam-ple size of each study by transforming each correlationeffect size into its corresponding Fisher-r, multiplying theFisher-r of each study by the inverse of the variance of eachstudy, summing those results together, and dividing by thesum of the inverse of the variances of each study. Theweighted correlation effect size was translated back to anunbiased effect size r for reporting results and this weightedcorrelation was also used to calculate a 95% confidenceinterval (CI) around the mean to provide an estimate of thevariability of r and to test whether r was statistically differ-ent from zero. The calculated effect sizes were consideredsmall if .10 ≥ r ≥ .29, moderate if .30 ≥ r ≥ .49, and large ifr ≥ .50 (Cohen, 1988). A homogeneity statistic, Q, was alsocalculated for the overall effect of each alternative keyboardon forearm pronation, wrist ulnar deviation, and wristextension. The homogeneity statistic has an approximate

chi-square distribution with k-1 degrees of freedom, wherek is the number of effect sizes and is used to determinewhether the various effect sizes included in the analysis thatare averaged into a mean value all estimate the same popu-lation effect size (Lipsey & Wilson, 2000). If Q was greaterthan the critical value of chi-square, then the variability ofthe effect size was larger than would be expected from sam-pling error and therefore each effect size does not approxi-mate a common population mean (Lipsey & Wilson).Some studies examined more than one keyboard or postureof interest.

Results Publication Characteristics

Table 2 presents the characteristics of the selected articles.In the six studies, 93% of the participants were female.Mean age of all participants was 27.7 years. The majority ofthe participants were employed in positions requiringextensive use of a computer keyboard and were consideredto be expert computer keyboard operators. The averagenumber of hours of computer use was 3.6 hours per day. Inall six studies, participants were excluded if they were symp-tomatic of MSD-UE.

Intervention Characteristics

In all six studies the order of keyboard presentation was ran-domized allowing participants to act as their own control(Table 2). Content of the keyboarding task varied fromstrictly alphabetic text to a combination of alphabetic andalphanumeric. Two studies employed the use of TypingTutor 6.0 software to provide typing text as well as to recordkeyboarding performance information (Marklin et al., 1999;Simoneau & Marklin, 2001). Information pertaining toacclimation period, rest period duration, and data collec-tion instruments is presented in Table 2.

Table 2. Characteristics of Included StudiesQuality Keyboard(s) Order of Rest Analysis

Study N rating examined keyboards Acclimation period Length of task between task instrument(s)

Marklin et al., 1999 87 20 FA, AT, AS Counterbalanced, 10 hrs at workplace 4 sess of 8 mins 2 mins Goniometricrandomized monitors

Hedge & Powers, 1995 12 17 AS Counterbalanced, 5 mins 50 mins — Video analysisrandomized with computer-

based programSimoneau et al., 2001 27 20 AS Randomized 5 mins 48 mins 2 mins Goniometric

monitorsZecevic et al., 2000 16 19 FA, AT Randomized — 30–60 mins — Retro-reflective

circular markersSzeto & Ng, 2000 10 21 FA Randomized — 60 mins 10 mins Retro-reflective

circular markers,EMG

Smith et al., 1998 24 19 AT Randomized 10 mins 4, 1 hr sess 10 mins Video cameraanalysis

Note. Unreported value in source study is denoted by —. AS = adjustable slope; AT = adjustable open-tented; EMG = electromyography test; FA = split fixed-angle.



Effect of Alternative Keyboards on Forearm Pronation

Two studies measured the effect of two alternative key-boards (FA and AT keyboards) on forearm pronation(Marklin et al., 1999; Zecevic et al., 2000). A single studyexamined forearm pronation posture resulting from the useof an FA keyboard (Zecevic et al.). Although a moderatecorrelation effect size (r = 0.33) was found, this result wasnonsignificant (95% CI = -0.21, 0.87, p < 0.08) (see Table3). Two studies examined forearm pronation while key-boarding on an AT keyboard (Marklin et al., 1999; Zecevicet al.). The AT keyboard had a large effect on reducing fore-arm pronation (r = 0.85) (95% CI = 0.65, 1.05, p < 0.05)when compared to the standard flat keyboard (see Table 2).When the overall effect of both the FA keyboard and theAT keyboard on forearm pronation were pooled and ana-lyzed a large effect (r = 0.82) (95% CI = 0.64, 1.00) (Q =0.10) on decreasing forearm pronation angle was revealed(p < 0.05) (see Table 3).

Effect of Alternative Keyboards on Ulnar Deviation Posture

Ulnar deviation postures during alternative keyboard use ascompared to the standard flat keyboard were examined infive studies (Marklin et al., 1999; Simoneau & Marklin,2001; Smith et al., 1998; Szeto & Ng, 2000; Zecevic et al.,2000). Three studies measured the effect of an FA keyboardon ulnar deviation compared to a standard keyboard

(Marklin et al.; Szeto & Ng; Zecevic et al.). Examination ofthe combined effect reveals that use of the FA keyboard hada large effect in reducing ulnar deviation posture (r = 0.79)(95% CI = 0.60, 0.98) as compared to the flat standard key-board (see Table 3).

The effect of an AT keyboard versus a standard flat key-board on ulnar deviation was analyzed in three studies(Marklin et al., 1999; Smith et al., 1998; Zecevic et al.,2000). As displayed in Table 3, the AT keyboard had a largeeffect on reducing ulnar deviation angle (r = .77) (95% CI= 0.59, 0.95) as compared to the standard flat keyboard.

Only one study investigated the effect of altering theslope of a keyboard on ulnar deviation posture compared toa standard keyboard (Simoneau & Marklin, 2001). Ulnardeviation of the wrists was examined with the AS keyboardset first at a positive tilt angle of 7.5° and then with the ASkeyboard set at a positive tilt angle of 15°. Ulnar deviation ofthe wrists was also examined with the AS keyboard adjustedto two predetermined negative tilt angles of first, a negative7.5° and second, a negative 15°. When the AS keyboard wasset at an inclined positive tilt angle of 7.5° or 15°, the ulnardeviation angle of the wrist was closer to neutral than whenthe AS keyboard was set at either of the two negative inclinedtilt angles. Statistical analysis revealed that an increasing pos-itive tilt of a keyboard had a small effect on ulnar deviation(r = 0.04) compared to the standard flat keyboard. This resultwas not significant (95% CI = –0.23, 0.32) (see Figure 1).When the tilt of the AS keyboard was decreased in a negative

Table 3. Effect of Alternative Keyboards on Pronation, Ulnar Deviation, and Wrist ExtensionKeyboard Study Study ES (r ) Total ES (r) 95% CI p value

PronationFA Zecevic et al., 2000 0.33 0.33 –0.21, 0.87 0.08AT Marklin et al., 1999 0.86

Zecevic et al., 2000 0.80 0.85 0.65, 1.07 0.0003Total effect of FA and AT keyboard **0.82 0.64, 1.00 0.05

Ulnar deviationAS (+ direction) Simoneau et al., 2001 0.05 0.04 –0.23, 0.31 0.0003AS (- direction) Simoneau et al., 2001 0.17 0.16 –0.11, 0.44 0.0003FA Marklin et al., 1999 0.85

Szeto & Ng, 2000 0.44 0.79 0.60, 0.98 0.0003Zecevic et al., 2000 0.18

AT Marklin et al., 1999 0.85Smith et al., 1998 0.34 0.77 0.59, 0.95 0.0003Zecevic et al., 2000 0.51

Total effect of AS, FA, and AT keyboards **0.42 0.26, 0.58 0.05

Wrist extensionAS (+ direction) Simoneau et al., 2001 0.51 0.51 0.11, 0.91 0.0025AS (- direction) Simoneau et al., 2001 0.65

Hedge & Powers, 1995 0.90 0.74 0.41, 1.07 0.00087AT Zecevic et al., 2000 0.19

Marklin et al., 1999 0.05.07 –0.13, 0.27

FA Marklin et al., 1999 0.26Szeto & Ng, 2000 0.39 0.30 0.11, 0.49 0.0113Zecevic et al., 2000 0.48

Total effect of AS and FA keyboards **0.52 0.17, 0.87 0.05

Note. **Indicates pooled effect of the listed alternative keyboards; AS = adjustable slope; AT = adjustable open-tented; ES = effect size; FA = split fixed-angle.


direction, ulnar deviation angle assumed by the subjectincreased 16% (r = 0.16) when compared to ulnar deviationpostures present during the use of a standard keyboard.However, like the result for increasing the tilt of the AS key-board in the positive direction, this result was nonsignificant(95% CI = –0.11, 0.44) (see Figure 1). Overall, the pooledeffect of the three alternative keyboards on ulnar deviationsuggests that the use of an alternative keyboard has a largeeffect on reducing ulnar deviation (r = .42) (95% CI = 0.26,0.58) (homogeneity, p < 0.05) compared to the standard key-

board. Statistical homogeneity was observed among theresults for the effect of the FA, AT, and AS keyboard on ulnardeviation of the wrists (Q = 0.00).

Effect of Alternative Keyboards on Wrist Extension

Wrist extension angle on alternative keyboards was mea-sured in five studies (Hedge & Powers, 1995; Marklin et al.,1999; Simoneau & Marklin, 2001; Szeto & Ng, 2000;Zecevic et al., 2000). Simoneau and Marklin specificallyexamined the effect of a positively tilt keyboard on wrist

Figure 1. Effect size (r) (central dot) and 95% confidence interval (heavy line) for the effect of AS keyboard, FA keyboard, and AT keyboard onforearm pronation, ulnar deviation, and wrist extension. The lines are numbered from 1 to 10. Those numbers correspond to each keyboardcondition as listed in the following key. The letters in the key refer to the studies.

Key1—FA keyboard on pronation (A*) 6—AT keyboard on ulnar deviation (A, B, F*)2—AT keyboard on pronation (B, A*) 7—AS keyboard (negative tilt) on wrist extension angle (C, E*)3—AS keyboard (positive tilt) on ulnar deviation (C*) 8—AS keyboard (positive tilt) on wrist extension angle (C*)4—AS keyboard (negative tilt) on ulnar deviation (C*) 9—AT keyboard on wrist extension angle (A, B*)5—FA keyboard on ulnar deviation (A, B, D*) 10—FA keyboard on wrist extension angle (A, B, D*)

*Note. A = Zecevic et al., 2000; B = Marklin et al., 1999; C = Simoneau et al., 2001; D = Szeto & Ng, 2000; E = Hedge & Powers, 1999; Smith et al.,1998. AS = adjustable slope keyboard; AT = adjustable open-tented keyboard; FA = split fixed-angle keyboard.

Effect Size (r)1—1



extension angle compared to a standard flat keyboard.Statistical analysis of the results of this study demonstratesthat wrist extension increases when the AS keyboard tilt wasincreased from 0° to a positive 7.5° and from 0° to a posi-tive 15° (r = .54) (95% CI = 0.27, 0.81) (see Table 3).

The effect of increasing the negative tilt of the AS key-board was evaluated by two studies (Hedge & Powers,1995; Simoneau & Marklin, 2001). When the tilt angle ofthe keyboard was decreased from 0° to a tilt of negative 7.5°and a tilt of negative 15°, there was a large decrease in wristextension angle (r = .77) (95% CI = 0.41, 1.13) comparedto the standard flat keyboard (see Table 3).

Three studies measured the effect of an FA keyboard onwrist extension versus a standard keyboard (Marklin et al.,1999; Szeto & Ng, 2000; Zecevic et al., 2000). Evaluationof the effect of an FA keyboard revealed a moderate decreasein wrist extension angle (r = 0.30) (95% CI = 0.11, 0.49)compared to a standard flat keyboard (see Table 3). Theseresults indicated that slight rotation and angulations of key-board halves, as characteristic of the FA keyboard, have amoderate effect on wrist extension posture.

Two studies investigated the effect of an AT keyboardon reducing wrist extension compared to a standard flatkeyboard (Marklin et al., 1999; Zecevic et al., 2000). TheAT keyboard was found to decrease wrist extension (r =0.07), however this result was small and lacked statisticalsignificance (95% CI = –0.13, 0.27).

Overall, the alternative keyboards were found to have alarge effect on wrist extension posture, generally positioningthe wrist in a more neutral position (r = 0.52) (95 % CI =0.17, 0.87). Statistical homogeneity of the results wasobserved (Q = –0.04) (p < 0.05) indicating the dispersion ofeffect sizes from the studies estimate the same populationeffect size.

Discussion The results of this meta-analysis provide evidence to sub-stantiate claims that alternative design keyboards promoteforearm and wrist postures closer to neutral position thanstandard flat keyboards. No previously identified work hasprovided a comprehensive review of the literature on theeffect of keyboard design on forearm position. In this studythree alternative keyboard designs were specifically investi-gated to determine their effect on wrist and forearm posturecompared to a standard keyboard. The three keyboarddesigns included AS keyboards, FA keyboards, and AT key-boards. Figure 2 presents the overall reduction in each pos-ture based on computer keyboard type.

Compared to the standard keyboard, the FA keyboardand the AT keyboard substantially reduced mean forearm

pronation angle (r = 0.72, p < 0.05). Though both the FAkeyboard and AT keyboard reduced forearm pronation, theAT keyboard was shown to be superior to the FA keyboard.The AT keyboard alone had a large effect (r = 0.83) ondecreasing forearm pronation angle compared to the stan-dard keyboard. The reason for a significant reduction inforearm pronation angle observed on the AT keyboard canbe attributed to a fixed lateral inclination of the keyboardhalves. This fixed lateral inclination, reported by Zecevic etal. (2000) to be approximately 42°, physically prevents key-boardists from pronating the forearm past 45°. The modestlateral inclination of 10° (Zecevic et al.) characteristic of theFA keyboard also decreased forearm pronation, but to alesser extent than did the AT keyboard. As pressure insidethe carpal tunnel increases as the forearm is pronated past45° (Rempel et al., 1995) and keyboard users have beenobserved to sustain forearm posture of around 60° of prona-tion while using a standard flat keyboard (Simoneau &Marklin, 2001), the use of an AT keyboard and, to a lesserextent, the FA keyboard, may be effective at reducing awk-ward postures that potentially could lead to increases incarpal tunnel pressure and development of a MSD-UE.

Compared to the standard flat keyboard, the FA (r =0.56) and AT (r = 0.61) keyboards profoundly reducedulnar deviation posture, while the AS keyboard demon-strated a small (r = 0.10) but significant effect on ulnar devi-ation posture. When the tilt of the AS keyboard wasincreased in the positive direction, wrist ulnar deviation sig-nificantly decreased whereas the opposite effect wasobserved when the tilt of the AS keyboard was adjusted toan increased negative tilt. Both the FA and AT keyboardallow the hands to be spaced at a greater distance from eachother during keyboarding and consequently promoting lessabduction of the upper arms and allowing a more neutralwrist posture to be assumed by the keyboard user. The use


Figure 2. Overall effect of three alternative keyboards on wrist andforearm postures. AS = adjustable slope; AT = adjustable open-tented; FA = split fixed-angle.


of alternative keyboards that significantly decrease ulnardeviation angles, such as the FA and AT keyboard, mayassist in reducing the forces acting on the biomechanicalstructures of the wrist and fingers as well as reducing thepressure in the carpal tunnel during keyboarding tasks.

The design of the flat standard keyboard may cause theuser to assume a posture of excessive wrist extension in orderto position the fingers for striking the appropriate keys.Results from the current study suggest the FA keyboardreduced wrist extension. The results also suggest that the ASkeyboard positioned in the negative direction reduce wristextension. When the tilt of the AS keyboard is increased ina positive direction, however, wrist extension increases alongwith the risk of MSD-UE. Clearly these results indicate thattilting the keyboard to a negative angle positions the wristcloser to a neutral extension posture. Compared to the stan-dard keyboard the AT keyboard was found to have a small,nonsignificant effect on wrist extension (r = 0.11). The twohalves of an AT keyboard are tented and rotated in a hori-zontal plane. Rather than specifically decreasing wrist exten-sion angle, the set up of the AT keyboard changes the orien-tation of the of the hand compared to the standard flatkeyboard (Zecevic et al., 2000) allowing for the hands towork in a more neutral wrist position but not requiring aneutral posture to use the keyboard. Therefore, some sub-jects may have maintained an excessive wrist extended pos-ture while using the AT keyboard. Overall, the three alter-native keyboard designs place the hand and forearm closer toa neutral extension position than the standard keyboard.However, the FA keyboard and AS keyboard adjusted in thenegative direction place the hand in a more neutral wristextension position than did the AT keyboard.

It is also important to note that in the study completedby Marklin et al. (1999) there were 10 hours of acclimationfor the assigned alternative keyboard whereas the otherstudies analyzed allowed only 5 to 10 minutes of acclima-tion. The longer acclimation period would allow users suf-ficient time to overcome the unfamiliarity of wrist and fore-arm posture on the alternative keyboard and thereforeresume prior level of typing performance on the standardkeyboard. However, the increased familiarity with the alter-native keyboard should not have had a substantial effect onwrist and forearm postures assumed when typing on thealternative keyboard compared to studies employing short-er periods of acclimation.

Clinical Significance This current meta-analysis provides evidence that alterna-tive keyboard designs decrease potentially harmful awkwardpostures typically assumed on a standard flat keyboard. If

an occupational therapy practitioner is concerned aboutreducing a specific awkward posture, this study indicatesthe keyboard type that best reduces that posture. It isimportant for clinicians to understand the priority of eachclient. No single keyboard was superior for reducing allthree awkward postures, so it is important that occupation-al therapy practitioners identify which posture appears to becausing the problem and select the keyboard that bestreduces it. This knowledge could be used to identify whichtype of keyboard may increase participation in computertasks that are limited by postural limitations secondary todisability. For example, a client may have reduced pronationsecondary to a radial head fracture. Based on the results ofthis analysis, his or her therapist would probably recom-mend an adjustable open-tented angle keyboard as thisdesign is the most successful for reducing the need topronate while using a keyboard (see Figure 2).

Despite the ability of these keyboards to promoteimproved postures, it is important to note that these stud-ies offer no evidence of a link between a specific keyboarddesign and alleviation of a specific MSD-UE condition orcomplaint. In other words, the use of an alternate keyboardmay reduce or eliminate risky postures and may still notalleviate, reduce, or prevent symptoms of MSD-UE inclients. A comprehensive program that addresses additionalelements of the computer workstation as well as the psy-chosocial environment may be necessary to help eliminateMSD-UE in computer users. However, armed with thisinformation, occupational therapy practitioners can provideclients with evidence-based information to make choices asto which keyboard may be the best for that client, thusdecreasing work and leisure time that might otherwise belost to discomfort or disability.

Limitations and Future Research This study provides information on only three alternativedesigns. Although additional alternative keyboard designshave been introduced on the market, they were not includ-ed in the present study as there was no literature availableon their effects on reducing potentially harmful postures.These other alternative designs may provide benefits to thecomputer user that is not provided by the three investigat-ed keyboards. A second limitation of the current study is itsrestriction to the effect of these three keyboard designs ononly three potentially harmful postures; forearm pronation,ulnar deviation, and wrist extension. Associations may existbetween MSD-UE and postures not included in the anal-ysis such as arm abduction and neck angle. The presentanalysis is limited to a small number of studies and a lim-ited number of effect sizes, however these effect sizes are



relatively large providing clinicians with some indication ofthe benefit of alternative keyboards.

When considering the information regarding alterna-tive keyboards provided by the present meta-analysis, cau-tion should taken due to limitations in the methodsemployed to gather and analyze the data. One limitation ofthis meta-analysis is there are no reliability values for therating of the articles. A single rater reviewed articles on threeseparate occasions. Both authors then discussed the ratingsof each article and a final decision was made concerninginclusion or exclusion of each article. No information onthe degree of intrarater or interrater reliability was recorded.A more rigorous and systematic process of rating the articlesmight have enhanced article selection. In addition, theauthors were not blinded to the results of each study priorto making the decision to accept or reject an article forinclusion. Removing the result section would help to ensurethat the decision to include the article was based solely onthe design and methods rather than the results.

Although this meta-analysis provides initial evidencethat alternative keyboards reduce awkward postures, it, andthe studies that comprise it, are only the first step in under-standing and reducing MSD-UE related to computer use.These studies do not demonstrate that reducing awkwardposture will actually reduce the incidence of MSD-UE ina computer-using population, nor do they examine theinteraction between alternative keyboards and other envi-ronmental interventions, such as chair configuration.Further research is needed to understand how the environ-ment, both physical and psychosocial, can be modified toenhance work performance and prevent work injury forcomputer users.

Conclusions The rise in computer use has affected the complex relation-ships among individuals, the environments in which theyfunction, and the occupations with which they becomeinvolved. Not only are computers increasingly common inthe workplace, they are also a venue for work, school, andleisure to enter the home. The use of the Internet to fulfilldaily occupations of many varied populations has led toincreased use of the computer keyboard and potentially anincrease in the occurrence of MSD-UE. Not only doesMSD-UE decrease worker productivity but also for thosewho rely on the computer to perform important daily tasks,MSD-UE may make use of the computer painful and irri-tating and result in a decrease in independence and overallquality of life.

This study examined three alternative keyboardsdesigns for their ability to reduce awkward postures associ-

ated with MSD-UE: forearm pronation, wrist extension,and ulnar deviation. Although each alternative keyboarddemonstrated a moderate to large effect on at least one ofthe three investigated postures, none of the keyboardsaffected all three postures (see Figure 2). Although the ATkeyboard had a large effect on reducing pronation of theforearm and wrist ulnar deviation, it had a small, non-significant effect on wrist extension, thus reducing two pos-tural risk factors while maintaining another. Although theFA keyboard was shown to have a large effect on ulnar devi-ation and a moderate effect on wrist extension, a nonsignif-icant effect on pronation was observed. Wrist extension wasdecreased on the AS keyboard adjusted to a negative anglehowever the concurrent with this decreased wrist extensionwas an increased ulnar deviation.

Studies indicating favorable results from the use of asingle alternative keyboard on forearm pronation, wristextension, and ulnar deviation are lacking in the literature.The high degree of variability from person to person whileusing a keyboard also makes it difficult to decrease awkwardforearm pronation, wrist extension, and ulnar deviationsolely through the selection of an alternative keyboarddesign. Additional research is needed to determine how spe-cific keyboard features such as slope, tilt, and split angleinteract with hand size, wrist and finger position, as well askeyboarding style to decrease biomechanically awkwardpostures while keyboarding. However, by decreasingbiomechanically awkward postures while keyboardingthrough the prescription of an alternative keyboard, occu-pational therapy practitioners may play an important rolein assisting computer users to maintain productivity, andpromote functional independence. s

Acknowledgments We would like to acknowledge the University of PittsburghCentral Research Development Fund and the CompetitiveMedical Research Fund of the UPMC Health System forthe grants that funded this research. The work on this meta-analysis was also supported by the University of Pittsburgh,Department of Occupational Therapy.

References Bergqvist, U., Wolgast, E., Nilsson, B., & Voss, M. (1995).

Musculoskeletal disorders among visual display terminalworkers: Individual, ergonomic, and work organizationalfactors. Ergonomics, 38, 763–776.

Bureau of Labor Statistics. (2002a). Computer and internet use atwork in 2001. U.S. Bureau of Labor Statistics. RetrievedOctober 30, 2002, from http://www.bls.gov/news.release/ciuaw.nr0.htm)



Bureau of Labor Statistics. (2002b). Over half of workers used acomputer in 2001. Retrieved March 15, 2004, from http://www.bls.gov/opub/ted/2002/oct/wk3/art04.htm

Cohen, J. (1988). Statistical power analysis for the behavioral sci-ences (2nd ed.). Hillsdale, NJ: Earlbaum.

Gerr, F., Marcus, M., Ensor, C., Kleinbaum, D., Cohen, S.,Edwards, A., et al. (2002). A prospective study of computerusers: I. Study design and incidence of musculoskeletalsymptoms and disorders. American Journal of IndustrialMedicine, 41, 221–235.

Gordon, M., Capell, H. A., & Madhok, R. (2002). The use of theInternet as a resource for health patients attending a rheuma-tology clinic. Rheumatology, 41, 1402–1405.

Greenhalgh, T. (1997). How to read a paper: Assessing the meth-odological quality of published papers. BMJ, 315, 305–308.

Hales, T. R., Sauter, S. L., Peterson, M. R., Fine, L. J., Putz-Anderson, V., Schleifer, L. R., et al. (1994). Musculoskeletaldisorders among visual display terminal users in a telecom-munications company. Ergonomics, 37, 1603–1621.

Hedge, A., & Powers, J. R. (1995). Wrist postures while key-boarding: Effects of a negative slope keyboard system and fullmotion arm supports. Ergonomics, 38(3), 508–517.

Hedge, A., Morimoto, S., & McCrobie, D. (1999). Effects ofkeyboard tray geometry on upper body posture and comfort.Ergonomics, 42(10), 1333–1349.

Kaye, S. (2000). Computer and internet use among people withdisabilities. In Disability Statistics Report (13) (pp. 1–13).Washington, DC: U.S. Department of Education, NationalInstitute on Disability and Rehabilitation Research.

Law, M., Stewart, D., Pollock, N., Letts, L., Bosch, J., &Westmorland, M. (1998). Guidelines for critical reviewform. Retrieved March 18, 2005, from http://fhs.mcmas-ter.ca/rehab/ebp/pdf/quanguidelines.pdf*

Lipsey, M. W., & Wilson, D. B. (Eds.). (2000). Interpreting andusing meta-analysis results. Practical Meta-Analysis (pp.146–171). London: Sage.

Luchetti, R., Schoenhuber, R., & Nathan, P. (1998). Correlationof segmental carpal tunnel pressures with changes in handand wrist positions in patients with carpal tunnel syndromeand controls. Journal of Hand Surgery, 23B, 598–602.

Maher, C. G., Sherrington, C., Herbert, R. D., Moseley, A. M.,& Elkins, M. (2003). Reliability of the PEDro Scale for rat-ing quality of randomized controlled trials. Physical Therapy,83, 713–721.

Marklin, R. W., Simoneau, G. G., & Monroe, J. F. (1999). Wristand forearm posture from typing on a split and verticallyinclined computer keyboards. Human Factors, 41(4), 559–569.

Moher, D., Schulz, K. F., & Altman, D. G. (2001). The CON-SORT statement: Revised recommendations for improvingthe quality of reports of parallel-group randomized trials.Lancet, 357, 657–662.

Moore, A., McQuay, H., & Gray, J. A. M. (Eds.). (1995).Evidence-based everything. Bandolier, 1(12), 1.

Nelson, J. E., Treaster, D. E., & Marras, W. S. (2000). Fingermotion, wrist motion, and tendon travel as a function ofkeyboard angles. Clinical Biomechanics, 15, 489–498.

Newburger, E. C. (1999). Computer use in the United States:Population characteristics. Current Population Reports.Retrieved March 15, 2004, from http://www.census.gov/prod/99pubs/p20-522.pdf

Rempel, D., Bach, J. M., Levinsohn, D. G., & Gordon, L.(1996). Effect of pronation/supination on carpal tunnel pres-sure. Paper presented at the Marconi Input Device ResearchConference, Marshall, CA.

Sackett, D. L., Straus, S. E., Richardson, W. S., Rosenberg, W., &Haynes, R. B. (2000). Evidence-based medicine: How to practiceand teach EBM (2nd ed.). New York: Churchill Livingstone.

Serina, E. R., Tal, R., & Rempel, D. (1999). Wrist and forearmpostures and motions during typing. Ergonomics, 42(7),938–951.

Simoneau, G. G., & Marklin, R. W. (2001). Effect of computerkeyboard slope and height on wrist extension angle. HumanFactors, 43(2), 287–298.

Smith, M. J., Karsh, B., Conway, F. T., Cohen, W. J., Morgan, J.J., Sanders, K., et al. (1998). Effects of a split keyboarddesign and wrist rest on performance, posture, and comfort.Human Factors, 40(2), 324–336.

Swanson, N. G., Galinsky, T. L., Cole, L. L., Pan, C. S., & Sauter,S. L. (1997). The impact of keyboard design on comfort andproductivity in a text-entry task. Applied Ergonomics, 28(1),9–16.

Szeto, G. P., & Ng, J. K. F. (2000). A comparison of wrist postureand forearm muscle activities while using an alternative key-board and a standard keyboard. Journal of OccupationalMedicine, 10(3), 189–197.

Tittiranonda, P., Rempel, E., Armstrong, T., & Burastero, S.(1999). Effect of four computer keyboards in computer userswith upper extremity musculoskeletal disorders. AmericanJournal of Industrial Medicine, 35, 647–661.

Verhagen, A. P., de Vet, H. C. W., de Bie, R. A., Kessels, A. G. H.,Boers, M., Bouter, L. M., et al. (1998). The Delphi list: Acriteria list for quality assessment of randomized clinical tri-als for conducting systematic reviews developed by DephiConsensus. Journal of Clinical Epidemiology, 51, 1235–1241.

Werner, R. A., & Andary, M. (2002). Carpal tunnel syndrome:Pathophysiology and clinical neurophysiology. ClinicalNeurophysiology, 113, 1373–1381.

Zecevic, A., Miller, D. I., & Harburn, K. (2000). An evaluationof the ergonomics of three computer keyboards. Ergonomics,43(1), 55–72.

References—Excluded Studies Fernstrom, E., Ericson, M. O., & Malker, H. (1994).

Electromyographic activity during typewriter and keyboarduse. Ergonomics, 37(3), 477–484.

Marklin, R. W., & Simoneau, G. G. (2001). Effect of setup con-figurations of split computer keyboards on wrist angle.Physical Therapy, 81(4), 1038–1048.

Monroe, J. (2003). Computer keyboard ergonomics. RetrievedOctober 28, 2003, from http://www.eng.mu.edu/~weber/pdf/monroe

Swanson, N. G., Galinsky, T. L., Cole, L. L., Pan, C. S., & Sauter,S. L. (1997). The impact of keyboard design on comfort andproductivity in a text-entry task. Applied Ergonomics, 28(1),9–16.

Tittiranonda, P., Rempel, E., Armstrong, T., & Burastero, S.(1999). Effect of four computer keyboards in computer userswith upper extremity musculoskeletal disorders. AmericanJournal of Industrial Medicine, 35, 647–661.



Documents

The Effect of Three Alternative Keyboard Designs on ... The Effect of Three Alternative Keyboard Designs on Forearm Pronation, ... ed pressure inside the carpal tunnel and to strain