Elevating Energy Expenditure with Workstation Alternatives

REVIEW Changing the way we work: elevating energy expenditure with workstation alternatives

C Tudor-Locke1, JM Schuna Jr1, LJ Frensham2 and M Proenca3,4

Emerging evidence supports the feasibility of raising daily energy expenditure (EE) by replacing office work-related sedentary behavior with low-intensity non-exercise physical activity (PA) via workstation alternatives to the traditional office chair and desktop computer-based combinations. The purpose of this review article is to introduce a simple taxonomy to facilitate classification and study of workstation alternatives, catalog the diversity of research undertaken to date related to energy balance, and present and summarize the gaps and opportunities for a research agenda for workstation alternatives moving forward. A PubMed search elicited 57 English language articles published since 2000; additional articles were identified by reviewing reference sections and contacting authors. Selection criteria ultimately focused on use of workstation alternatives during simulated or real work tasks. The EE of sitting on a stability ball or using sit–stand/standing desks is comparable to the traditional seated condition (D1.2 kcal min À 1). The treadmill and pedal desks (active workstation alternatives) offer the greatest promise in terms of EE (D2–4 kcal min À 1). Sitting on a stability ball or using sit–stand/standing desks does not impair task performance relative to the traditional seated condition. Some evidence of typing impairment is inconsistently reported with active workstation alternatives; the finer motor skills required for mouse-related tasks may be more affected. Little is known about learning or adaptation with practice. Users are generally accepting of workstation alternatives; however, there is evidence of less than optimal use. Active workstations (that is, treadmill desks and pedal desks) in particular represent a potential strategy for mitigating the diminished EE inherent to contemporary office-based workplaces, but only if they are scalable. The science supporting active workstations is young and heterogeneous; however, this means that there are many knowledge gaps and opportunities for research, including those focused on implementation issues related to optimizing both employers’ and workers’ uptake.

International Journal of Obesity (2014) 38, 755–765; doi:10.1038/ijo.2013.223 Keywords: occupation; sedentary behavior; sitting; workplace; physical activity

INTRODUCTION to stoic occupational sedentarism are intended to generate In 2007, Hamilton et al.1 presented compelling evidence that an innocuous additional EE that is: elevated above that asso- sedentary behaviors, such as those ubiquitous to contemporary ciated with sitting, tolerable for extended durations and minimally workplaces,2 may be contributing to the current obesity epidemic distracting from the primary work task. As such, workstation and corollary chronic diseases, including cardiovascular disease and alternatives can be contrasted with conventional worksite Type 2 diabetes. For example, Hu et al.3 estimated that each 2-hour wellness initiatives, including worksite fitness facilities and/or increase in occupational sitting time was associated with a 5–7% structured exercise programs12,13 that focus on participation increase in risk for obesity and Type 2 diabetes. Over the past 50 in structured and typically higher intensity activities scheduled years in the US, there has been a shift toward occupations that during non-working time. The novelty of workstation alternatives are largely composed of desk-based sedentary behavior is characterized by large gaps in scientific understanding that and this trend has been associated with population-level weight present great research opportunities. The purpose of this review gain.4 The direct5 and indirect (non-medical)6 costs of obesity article is to introduce a simple taxonomy to facilitate classification arehigh;foremployerstheseindirectcostsincludeabsenteeismand and study of workstation alternatives, catalog the diversity of disability6 and reduced productivity associated with presenteeism.7 research undertaken to date as it relates to energy balance- The prevailing ergonomic paradigm has advocated office related questions and summarize the research gaps and furniture and environmental design in favor of continually opportunities. As our primary focus will be related to energy reducing physical workloads, and thus energy expenditure (EE).8 balance, we organize the presentation by answering the following Recent research, however, has reported the feasibility of replacing questions about workstation alternatives: (1) Will they increase EE? work-related sedentary behavior with low-intensity non-exercise And if so, by how much? (2) Will they interfere with work? physical activity (PA) via workstation alternatives to the tradi- (3) Will workers use them? And how long will they use them? tional seated office chair and desktop computer-based (4) Are interventions designed around workstation alternatives combinations.9–11 These office environmental countermeasures successful?

1Preventive Medicine and Healthy Aging, Pennington Biomedical Research Center, Baton Rouge, LA, USA; 2Exercise for Health and Human Performance Research Group, University of South Australia, Adelaide, South Australia, Australia; 3Laborato´rio de Pesquisa em Fisioterapia Pulmonar (LFIP), Departamento de Fisioterapia, Universidade Estadual de Londrina (UEL), Londrina, Brazil and 4CAPES Foundation, Ministry of Education of Brazil, Brası´lia, Brazil. Correspondence: Dr C Tudor-Locke, Walking Behavior Laboratory, Pennington Biomedical Research Center, 6400 Perkins Road, Baton Rouge, LA 70808, USA. E-mail: [email protected] Received 19 September 2013; revised 18 November 2013; accepted 24 November 2013; accepted article preview online 28 November 2013; advance online publication, 24 December 2013 Energy expenditure of workstation alternatives C Tudor-Locke et al 756 MATERIALS AND METHODS height and adding a perpendicular treadmill to accommodate walking A PubMed search was updated on November 04/2013 using the following while working (a ‘treadmill desk’) and (4) incorporating an under-desk Boolean phrase: (workstation OR desk) AND (sitting OR standing OR pedaling/stepper/elliptical device (a ‘pedal desk’). The first two workstation sit–stand OR walking OR stepping OR pedaling OR cycling OR treadmill OR alternatives represent a sub-category best described as ‘static work- bicycle OR ball). Additional filters pinpointed articles published in English stations,’ where the opportunity for movement is incidental and sporadic since 2000 (the year that Levine et al.14 published a seminal article titled (for example, postural transitions, weight shifting) and otherwise largely ‘Energy expenditure of non-exercise activity’) and involving human adults contained to upper body movements associated with the primary work of typical full-time working age 19–64 years. Fifty-seven articles were task (for example, typing, mouse pointing, and so on). In contrast, the latter identified. Titles and abstracts were reviewed and selection criteria two workstation alternatives can be classified as ‘active workstations,’ required that the research focus on simulated or actual work-related tasks because they facilitate a user’s low-intensity rhythmic movement (for (for example, computer-based tasks, reading, and so on) while using the example, walking, pedaling) while simultaneously engaged in desk work. workstation alternative, or otherwise be intended for use in the workplace. Similar to the sit–stand/standing desks, treadmill desk designs may have Additional research was located by examining reference sections of fixed (for example, a treadmill desk that does not also accommodate the previously identified articles, corresponding with authors and extending option to sit) or height-adjustable (by mechanical or electronic means) the publication date to earlier years (for example, 198915) as needed. positions to accommodate intermittent bouts of walking/sitting behaviors. Ultimately, we identified nine controlled studies of EE (5 with comparable A pedal desk design is inherently adjustable because it can accommodate data presented in Figure 1), 14 studies focused on task performance both active pedaling and traditional seated work merely by ceasing to (Table 2) and 8 intervention studies focused on sit–stand desks (Table 3; 4 pedal. In addition to movement and position, user posture (sitting, upright, studies), treadmill desks (Table 4; 4 studies) and 2 multi-component studies or accommodation for both) is also determined by design. A stability (that also utilized one or two types of active workstations) directly relevant ball and a pedal desk provide seated posture and a standing desk and a to our energy balance-related questions listed above. Twenty-five studies treadmill desk accommodate upright posture; a sit–stand desk or an that provide insight into acceptability of workstation alternatives are adjustable treadmill desk can accommodate both types of posture. Finally, presented in the text below. There is overlap between topics; some studies access is a distribution choice and can be described as individually provide evidence related to more than one of the energy balance-related dedicated (one user with ad libitum access) or shared (multiple users with questions guiding this review. either as-needed or scheduled access). The following are examples of how the taxonomy may be used: an imaging center may choose to implement a shared access, fixed position, static standing desk for technologists to operate cameras and process images on an as-needed basis: a call Taxonomy of workstation alternatives center may provide a scheduled shared access, fixed position, active The primary alternatives to traditional office chair and desktop computer- workstation (treadmill desk) for workers to check email messages and based workstations that have been considered include (1) replacing an engage in on-line training; dedicated pedal desks may be provided to office chair with a stability/exercise ball, (2) changing the desk height cubicle-based administrative workers (with the option to pedal as desired to accommodate intermittent (for example, an adjustable ‘sit–stand desk’) throughout the day), and executives may prefer their own active or continuous standing (a fixed ‘standing desk’), (3) changing the desk workstation, specifically a dedicated, height-adjustable, treadmill desk. To summarize, the taxonomy classifies workstation alternatives by movement (static vs active), position (fix vs adjustable) and posture (seated vs upright) based on design, and access (shared vs dedicated) based on distribution choice. A tabular version of the taxonomy is presented in Table 1.

Will workstation alternatives increase energy expenditure? And by how much? A number of controlled studies have quantified objectively determined EE values associated with traditional seated working and various workstation alternatives.10,16–22 Studies of working seated in an office chair, seated on a stability ball, standing, seated pedaling and walking on a treadmill that could also be compared by mean kcal min À 1 are depicted in Figure 1. Two studies are conspicuously absent.16,22 Beers et al.16 presented results (for resting, sitting in an office chair, sitting on a stability ball and standing) graphically only, making it difficult to extract exact EE data. Cox et al.22 studied different speaking conditions while seated, standing and walking on a treadmill at 1.6 km h À 1 and reported mean oxygen consumption values (ml kg À 1 min À 1), but not expressly EE (dependent upon an Figure 1. EE during different working conditions. Note: values individual’s body mass). are presented as means. A, data presented by Carr et al.;32 B, Overall, there appears to be little variation in EE associated with working data presented by Reiff et al.;17 C, data presented by Speck and while sitting in an office chair, sitting on a stability ball, or even standing Schmitz;20 D, data presented by Swartz et al.;19 E, data presented (together averaging B1.2 kcal min À 1, but ranging from 0.99 to by Levine and Miller.18 1.46 kcal min À 1, both extremes measured in a traditional seated working

Table 1. Taxonomy of workstation alternative designs and their access determined by distribution choice

Workstation alternative Design Access

Movement Position Posture Distribution

Static Active Fixed Adjustable Seated Upright Dedicated Shared

Sitting on stability ball | — | — | — || Standing desk | — | —— ||| Sit–stand desk | —— || ||| Treadmill desk — | Either, but not both Upright only or accommodating both || Pedal desk — | — || — ||

International Journal of Obesity (2014) 755 – 765 & 2014 Macmillan Publishers Limited Energy expenditure of workstation alternatives C Tudor-Locke et al 757 condition), which suggests these particular workstation alternatives would condition was similar to the faster walking condition (3.2 km h À 1). have little impact on reducing body weight relative to the traditional Carr et al.32 also reported that heart rate, systolic blood pressure, oxygen seated condition. To be clear, differences in EE between seated work and consumption and leg muscle activation all increased significantly while either sitting on a stability ball or standing work are less than what can be using a pedal desk vs a seated working condition. described as even modest. Although reported as statistically significant, As evidence indicates that active workstations present the greatest 16 À 1 Beers et al. calculated the difference to be no more than 4.1 kcal h . opportunity for impacting workers’ energy balance, they occupy the focus Considering an 8-hour work day and relative to seated work, the increment of the remainder of this discourse. Other types of workstation alternatives in EE from these two types of workstation alternatives only net an extra are referred to primarily as comparative examples to place findings of À 1 32 kcal day . Based on the simplistic guidelines that a pound of fat is active workstations in a broader context. approximately equivalent to 3500 kcals,23 this difference indicates that over 100 full-time working days sitting on a stability ball or standing at a desk must be endured to reduce body fat by one pound! Although it is Will workstation alternatives interfere with work? possible that a standing individual may be more inclined to accumulate While using an active workstation may decrease prolonged sedentary random movement patterns (fidgeting,14 weight shifting associated with behavior and elevate EE typically associated with traditional seated postural sway,24 incidental steps25) than one confined to a chair, the direct workstations, it is not entirely clear how its use will affect performance on evidence suggests that it is unlikely that sit–stand/standing workstation work tasks and ultimately productivity. Real or perceived conflicts between alternatives could effectively counteract positive energy balances productivity and energy balance priorities will inevitably undermine or associated with traditional sedentary workstations in an any meaningful at least restrict the implementation of workstation alternatives in the way. The data summarized in Figure 1 indicate that efforts to decrease workforce, and thus the opportunity to alter workers’ EE. sitting behavior only by replacing it with standing behavior appear to 33 Using an active workstation is an example of a dual task performance: promise only a negligible difference in EE. That is not to say that their use walking, pedaling or stepping while performing other tasks such as may be associated with other potential health benefits of decreased sitting reading, thinking, speaking, typing or texting. Studies of ambulatory behaviors.1 For example, heart rate is significantly elevated in the standing behavior demonstrate that gait, for example, is not an automatic process working position vs the seated working position, suggesting enhanced 16,26 and requires considerable information processing, which may in turn blood flow. Regardless, herein we shall remain focused on questions reduce performance on a secondary task depending on the complexity of related to energy balance. 34 that task and the age of the person. The concern is that a worker using a In contrast to sit–stand/standing desks, treadmill desks appear to treadmill desk may need to divide attentional resources between walking present a greater potential impact for altering EE above that associated safely on a treadmill and performing office work, ultimately compromising with conventional seated workstations. Heart rate responses to using a productivity. There is even less understanding about how a non-weight- treadmill desk at 1.6 and 3.2 km h À 1 are elevated in a graded manner 27 18 bearing pedaling action with a more stabilized upper body and no threat relative to the seated working condition. Levine and Miller reported of tripping/falling may or may not affect productivity. that the EE of working while walking on a treadmill at a self-selected pace Productivity generally refers to both the quality of output (for example, of 1.77 km h À 1 was B3.2 kcal min À 1 (Figure 1). Although there is no final product) produced and the rate at which it is produced, all relative to directly measured corroborating evidence at this time, this increment worker input (taking into consideration time, technology and other represents a B2.7-fold increase over the EE expected of seated work resources, education level, managerial techniques, and so on). Measuring (assuming this average is 1.2 kcal min À 1 based on Figure 1). Treadmill productivity in the workforce, especially in the service sector typified by workstations could theoretically increase an individual’s EE over the seated 35 À 1 18 office-based occupations, is a well-known challenge. Considerable condition by 4100 kcal h (depending upon body mass), which could heterogeneity exists within and between industries, workplaces, and have more dramatic impacts on energy balance, for example, by even workers with similar job titles and responsibilities. It is not surprising potentially 4800 kcal day À 1 if workers could realistically tolerate it for a therefore that the majority of published studies of workstation alternatives full 8-hour workday. Still, even a single hour of daily use represents a have focused primarily on controlled study of simulated task performance quotidian EE increment that has been posited to prevent weight gain.28 29 (for example, typing, mouse pointing and clicking, reading, math Thompson et al. used an accelerometer to indirectly estimate that problems, speaking, and so on), which is more easily standardized and physicians burned an additional 197 kcal day À 1 walking and working an À 1 À 1 measureable (Table 2). Studies that have reported workers’ perceptions of average 90 min day at 1.6 km h on the treadmill desk. Select the impact on productivity are presented later, along with other user groups29 may average up to 90 min day À 1 and up to more than perceptions of acceptability. 100 min day À 1 in longer-term users.30 Other office-based workers more The single study that examined task performance (typing) between frequently average only 30 min day À 1 of use, even with individually 9 traditional seated working and that conducted while sitting on a stability dedicated treadmill desks. Although these authors suggested that this 16 À 1 ball showed no difference. Likewise, standing while being engaged in a was equivalent to an estimated 100 kcal day increase in EE over a range of work-related tasks does not appear to negatively affect traditional seated working condition, the prior objectively monitored performance related to a seated working condition.16,22,27,36,37 Studies of evidence indicates that this is likely closer to a 50 kcal day À 1 increase.18 the treadmill desk have shown no impact on typing (while walking at Although passive pedaling using chair/bedside devices is routinely used 15 22 38 À 1 15 31 10,32 1.4, 1.61, 2.25, and 2.8 km h , and reductions in typing speed in physical therapy, there are only two studies that have investigated 38 27,39 38 127 (while walking at 1.3, 1.6, 3.2, and 3.6 km h À ), all relative to the the EE associated with a pedaling/stepping device. Office work was not seated working condition. Mouse-related tasks appear to be affected by performed concurrently while using the stepping device in the study 27,39 10 walking on a treadmill desk. Transcription speed is negatively affected conducted by McAlpine et al., however, so it is not included in Figure 1. while walking on a treadmill, even with 4 hours of practice; however, in Regardless, the researchers in that study reported relatively large increases that study transcriptionists also had to learn a new system for starting and in EE (B4.81 kcal min À 1 over sitting) in participants ranging in size from 32 À 1 stopping transcription tapes that differed from the traditional foot pedal lean to obese. Carr et al. measured EE to be 2.14 kcal min while 40 approach. There are only two studies that have evaluated work-related working at a pedal desk and pedaling at a cadence of 45 r.p.m. (estimated 27,32 32 task performance while using a pedal desk. Carr et al. reported no 2.25 km h À 1) and a reported effort of 9 watts. This pace was set by the differences in tests of cognitive function, most mouse-related tasks researchers so the EE of self-selected pedaling while working at a pedal (dragging and menu selection) or typing speed/errors while working in a desk remains unknown. seated condition or at a pedal desk; however, there was a significant There may be a potential to realize increased EE with a pedal desk, again decrement in mouse pointing speed (0.13 s) in the pedal desk condition. depending whether or not the activity could be tolerated for long 27 21 Straker et al. reported decrements in typing speed and mouse-related durations while working. Peterman et al. also reported elevated EEs over tasks relative to the seated condition that were less in comparison to that seated rest with motor-driven passive pedaling (participants supplied observed while participants walked on a treadmill desk at 1.6 and no volitional muscle activation, rather their feet were ‘pulled’ around the 3.2 km h À 1. Mouse-related tasks represent finer motor tasks than typing À 1 À 1 cog) at both 60 r.p.m. (1.78 kcal min ) and 90 r.p.m. (2.51 kcal min ), but and may be more affected by extraneous movement associated with both again office work was not concurrently performed. Although not a direct types of active workstation alternatives.27 27 measure of EE, Straker et al. studied heart rate responses to different Sample sizes for studies evaluating performance while using workstation workstation alternatives, including pedaling at 5 and 30 watts. The slower alternatives have generally been small and participants are typically pedaling condition elicited a similar heart rate increase over a seated novices in terms of their previous exposure to working while using working condition as a standing workstation, and the faster pedaling workstation alternatives. Assessed tasks are diverse, as are use parameters

& 2014 Macmillan Publishers Limited International Journal of Obesity (2014) 755 – 765 758 nentoa ora fOeiy(04 5 765 – 755 (2014) Obesity of Journal International

Table 2. Studies of the impact that workstation alternatives have on performance

Reference; workstation Sample (size, gender, age, physical limitations, Study design Performance measure Signiﬁcant Findings alternative(s) studied work abilities, prior experience, all if reported) relative to seated condition

Edelson and Danoff15 1 man, 4 women; mean age 26 years; BMI 20 min intervals of typing while treadmill walking vs typing in seated Typing speed–accuracy No difference between conditions Treadmill desk not reported; with typing speeds over 50 condition; 5 2-hour sessions over a two-week period, first 3 sessions composite score words per min were the learning period, session 4 and 5 were experimental and control conditions, alternated among participants; 30 s break to nryepniueo oktto alternatives workstation of expenditure Energy reposition between intervals and 5 min break at session mid-point; participants permitted to listen to background music while they worked; self-selected speeds between 1.4 and 2.8 km h À 1 Ebara et al.36 12 men, 12 women; 12: 20–29 years, 12: 60– 150 min English transcription under counterbalanced standard seated Mean number of letters No difference between conditions Sit-to-stand desk 69 years; BMI not reported; experience using condition, high chair (chair with elevated work surface) condition and correctly transcribed every word process/spread sheet software; typing 10-min sitting combined with 5-min seated condition; X3days 10 min ability, right-handed between conditions Beers et al.16 13 men, 12 women; age 21–48 years; mean Counterbalanced comparison of 20 min of word-processing while Total number of words No difference between conditions Seated on exercise ball BMIo30 kg m À 2 ; employed sedentary seated in an office chair, on an exercise ball and standing typed during each of the Standing desk workers; no physical limitations; ability to conditions type and familiarity with word-processing software Drury et al.61 7 men and 5 women; mean age 36.2 years; Balanced cross-over comparison of simulated job performance tasks Counts of correct No difference between conditions Sitting on a high chair BMI not reported (for example, X-ray screening, threat detection) holding posture for responses and false alarms Tudor-Locke C standing 40 min-sessions Fidler et al.62 2 male radiologists in good health with a Participants re-interpreted 100 clinical computed tomographic (CT) Discrepant findings Significantly higher detection rates on treadmill desk Treadmill desk BMI of 23.3 kg m À 2 and 26.4 kg m À 2 , examinations previously interpreted (one-year interval earlier) while than the initial seated CT interpretations. respectively. walking at 1.6 km h À 1 at a treadmill desk. Husemann et al.63 60 men; ages 18 to 35 years; BMI not Randomized controlled trial (seated condition, sit–stand condition) of Data entry efficiency, A trend toward decreased data entry efficiency Sit–stand workstation reported; university students; no diseases or data entry. For each hour of work, the control group completed 45 min quantity of keystrokes, (difference detected in quantity of keystrokes but not

central nervous system problems. (75% of the time) of data entry in a sitting position and the sit–stand errors and error rates errors or error rate) in sit–stand condition that was not al et workstation group completed only 30 min (50% of the time) of data (per keystrokes) significant. entry in a sitting position, followed by a 15-minute interval during which participants entered data while standing. The remaining 15 min for both groups were occupied with a 10-min period of non-data entry office work and a 5-min break. This working cycle was repeated similarly each hour for 4 h day À 1 , for 5 consecutive days. John et al.39 11 men (mean 26.4 years), 9 women (mean Participants completed a battery of 5 tests under counterbalanced Tests assessed selective Mouse clicking, drag and drop speeds were faster, typing Treadmill desk 27.0 years); mean BMI 22.98 kg m À 2 ; traditional seated and treadmill walking (1.6 km h À 1 ) conditions on 2 attention and processing speed greater and math scores better (6–11%) in the university students; no experience with a separate days speed, cognitive function traditional seated condition. No other differences treadmill desk and fine motor movement between conditions. Straker et al.27 14 men, 16 women; ages 22 to 64 years; Standardized counterbalanced computer tasks (3-min typing, Typing: words per min and Relative to seated condition: Typing speedk6%, error Standing desk mean BMI 25 kg m À 2 ; variable typing skills unlimited time for mouse pointing to completion, 2-min combined error rate. Mouse pointing: ratem3% across walking speeds. No difference between Treadmill desk typing and mouse) in 6 conditions: seated, standing, walking at speed and error rate. walking speeds. Pedal desk 1.6 km h À 1 and 3.2 km h À 1 , and pedaling at 5 and 30 watts (users Mousek14%, error ratem106% across walking speeds. were required to maintain a set speed). ‘Percentage error increase appears large because of very low rate.’ No difference between walking speeds. Typing speedk3% at the slower pedaling speed, error ratemby 0.7%. No difference in typing speed or error rate at faster pedaling speed. Mousek5%, error ratem61% across pedaling speeds. No

& difference between pedaling speeds. k m

04McilnPbihr Limited Publishers Macmillan 2014 Combined tasks speed 15%, error rate 35% across walking speeds. No difference between walking speeds. Combined tasks speed k3% across pedaling speeds. Error rate across pedaling speeds similar to seated condition. No difference between pedaling speeds. No difference in any performance measure between sitting and standing conditions. Cox et al.22 9 men, 22 women; mean 37 years; BMI not Participants read silently, read aloud and spoke spontaneously during Expert listeners assessed No difference between conditions Standing desk reported 3 conditions: sitting, standing and treadmill walking at 1.61 km h À 1 ; randomized samples of Treadmill desk 5 min at each of the posture/task conditions participants’ speech; average number of syllables included in each speech Ohlinger et al.37 50 participants (gender not reported); mean Counterbalanced assessment of 2 cognitive and 1 motor task; three Divided attention, short No difference between conditions in cognitive tests. Standing desk age 43.2 years; mean BMI 28.5 kg m À 2 , not conditions performed in randomly assigned 1 of 6 orders: seated, term auditory verbal Fingers tapping scores were higher in sitting vs walking Treadmill desk limited in ability to walk at slow speeds; no standing, treadmill walking at 1.6 km h À 1 . All completed during a memory, selective and also higher in standing vs walking. No difference balance disorders single 75-min session. attention and a motor skill between sitting and standing conditions. (ﬁnger taps) Energy expenditure of workstation alternatives C Tudor-Locke et al 759 (for example, walking speed). Little is known about how continued use and practice affects performance, whether users eventually adapt to workstation alternatives, what factors may affect this process, and/or how long it may take.

Will workers use workstation alternatives? How long will they use them? An acceptable workstation alternative can generally be described as one that is perceived as tolerable by the worker (not being unduly and between seated typing and burdensome, inconvenient, inaccessible or uncomfortable) and innocuous 1

À (non-distracting from the task at hand, not tiring). Assessment of acceptability is frequently a part of formative evaluation before larger scale study or implementation.41 There is no standardized approach to the . No other differences between conditions. 1 measurement of acceptability and researchers frequently report À exploratory findings derived from their own constructed and delivered questionnaires, interviews and/or focus groups. For example, Thompson 9 3.2 km h No difference between conditions Speed of using amake mouse to a move selection a wascondition; cursor slower to no a (0.13 other s) target differences in and between the conditions. pedal desk No difference on theseated accuracy or of using transcription theComplete while treadmill actual desk. dictations werewalking. 16% slower while Significant differences in typingtyping speed and between 1.3 km seated h et al. reported acceptability results and provided the questions they constructed based on a Likert scale. In general, reported study findings may reflect study-specific instrument design; if there are no questions solicited about negative perceptions, then none are captured. Confronted with no easy way to summarize these diverse findings, we only offer here a brief summary of those reported for the different types of workstation alternatives. The few studies that have been conducted comparing sitting on an office chair with sitting on a stability ball have generally reported that sitting on a ball was associated with increased discomfort.42–44 Cognitive tests assessed response time, response accuracy Cognitive tests, fine motor/ computer performance tests (typing and mousing) Speed and accuracy (number of errors) of transcription Adjusted words per min (measure of typing speed after exclusion of errors) An exception is the study by Beers et al.16 that reported no difference in sensations of comfort or of ‘liking’ between the two seated conditions. or

1 Both users (workers) and purchasers (employers) have been generally À positive about the introduction of sit–stand desks into workplaces.45 Study participants have reported that sit–stand/standing workstation alternatives were easy to use,46 enjoyable,46 comfortable,46,47 did not negatively affect 9,47 to 4.0 km h

, reported at productivity or were perceived as having improved productivity by 1 1 46 À

À at least some of the workers sampled, and improved perceptions of energy,47 health,47 happiness47 and stress.47 Open-ended comments solicited by Pronk et al.47 at the end of a sit–stand desk intervention while transcribing; half of included perceptions of less low back pain and shoulder tension, posture 1

À improvement, decreased wrist and elbow pain and increased comfort. Others have also noted perceived beneﬁts for good posture and potential for reduced musculoskeletal soreness associated with all-day sitting.26 There is some evidence that a sit–stand desk may be preferred to the traditional seated workstation.27,46 In particular, Alkhajah et al.46 reported that 83% of the workers engaged in their study of sit–stand desks indicated that they disagreed or strongly disagreed with a statement indicating that they would prefer to return to their original workstation set- up. The ability to alternate freely between sitting and standing postures, . 2 min rests between conditions. 1

À rather than being confined to either posture, is associated with the reduced discomfort and increased preference.48 On the negative side, there have been reports that it is less comfortable than traditional seated 16,26,36 a seated control condition, separated by 48 h. 9 watts) or in a seated condition. subjects were randomly assignedfirst to be 8 h, seated transcribing followedthe during by other the 8 half h transcribed transcribingthe using using the second the treadmill 8 treadmill h desk for desk, transcribed the while while first seated. 8 h and Two counterbalanced sessions of walking 0.8 km h Two counterbalanced sessions of computer-basedthe tasks pedal performed desk on (40 r.p.m. or an estimated 2.3 km h 4-min standardized typing testscounterbalanced were conditions: seated performed and under while3.2 4 km walking h at 1.3, 2.25 and Installation of a treadmilla desk special consisted keyboard of to a treadmill, walk a at monitor 1.6 km and h Study design Performance measure Significant Findings work, less comfortable and less ‘likable’ than working while sitting on a stability ball,16 more fatiguing,16 hands and wrists are insufficiently supported,46 there is insufficient room to navigate a mouse,46 there is leg discomfort,27 and the need to change footwear to comfortably use the sit–stand desk is a burden.46 Alkhajah et al.46 also reported that some workers experiencing the sit–stand desk perceived that it reduced their productivity. Edelson and Danoff15 reported that typists felt less stressed while ; undergraduate

2 working on a treadmill desk vs a traditional seated condition, but À otherwise there were no statistical differences in perceived arousal or body ; no ambulatory/exercise

2 complaints. Positive user perceptions have also included the ability to À ‘break up the day,’ and that use ‘may help creativity.’27 Thompson et al.9 reported that workers believed treadmill desks were not too noisy, did not negatively affect productivity and did not induce undue fatigue. The

40 words per min; no balance disorders authors reported that a question regarding fatigue at the end of the day work abilities, prior experience, all if reported) students; healthy, not diagnosedattention with deficit hyperactivity disordernot and currently using stimulants limitations, overt cardiovascular, metabolic, respiratory, or neurological diseases. Sample (size, gender, age, physical limitations, 11 women (age andmedical BMI transcriptionists; not not reported); familiartreadmill with desk or special keyboard. 9 men, 15 women; meannot age reported; 23.2 not years; familiar BMI desk; with variable treadmill typing skills;X ability to type 27 men, 39 women;mean mean BMI age 24.7 kg 21.06 m years; 1 man, 17 women:university mean office-based age employees, 43.4 mean BMI years; 28.8 kg m generated the greatest disagreement among participants; some felt less tired while an equal number felt more tired. They also reported that participants with back pain indicated that it was improved while using the treadmill desk. Straker et al.27 reported that some study participants 64 perceived decreases in both typing and mouse speeds and increases in (Continued ) et al. 38 associated error rates, regardless of walking speed tested (1.6 and 32 3.2 km h À 1), relative to a seated condition. Specific concerns included 40 et al. et al. diminished fine motor coordination, a dizzying effect of the head movement relative to the computer monitor, slight wrist and/or leg Alderman Carr Funk Table 2. Reference; workstation alternative(s) studied relative to seated condition Treadmill desk Treadmill desk Treadmill desk Pedal desk Thompson and Levine NB: apparent differences in descriptive information reflect underlying discrepancies in reporting conventions used in original studies. discomfort, and a perception of an increased need for concentration on

& 2014 Macmillan Publishers Limited International Journal of Obesity (2014) 755 – 765 760 nentoa ora fOeiy(04 5 765 – 755 (2014) Obesity of Journal International

Table 3. Sit-to-stand interventions/standing desk interventions

Reference Study design; sample (size, gender, age, Intervention Outcome measures Significant findings occupation/workplace, BMI, salient inclusion/exclusion criteria, all if reported) nryepniueo oktto alternatives workstation of expenditure Energy Husemann et al.63 Randomized controlled trial (seated For each hour of work, the control group Self-reported psychological and No significant difference in condition, sit–stand condition) of data completed 45 min (75% of the time) of data physical well-being psychological well-being by condition. entry; 60 men; ages 18–35 years; BMI entry in a sitting position and the sit–stand (musculoskeletal complaints); There were significantly fewer not reported; university students; no workstation group completed only 30 min data entry efficiency (presented musculoskeletal complaints after the diseases or central nervous system (50% of the time) of data entry in a sitting previously in Table 2) working day in the sit–stand condition. problems. position, followed by a 15-minute interval during which participants entered data while standing. The remaining 15 min for both groups were occupied with a 10-min

period of non-data entry ofﬁce work and a Tudor-Locke C 5-min break. This working cycle was repeated similarly each hour for 4 h day À 1 , for 5 consecutive days. Alkhajah et al.46 Quasi-experiment comparing Installation of individually dedicated At all three time points: 1 week: sitting time reduced intervention group (1 man, 7 women, commercially available sit–stand objectively monitored (activPAL) 143 min day À 1 at workplace, À 2 À 1 mean BMI 22.6 kg m ) and workstation (Ergotron WorkFit-S Single LD time spent sitting, standing, 97 min day during waking time al et comparison group (2 men, 12 women, sit–stand Workstation, E$US380) stepping. 3-month sitting time reduced mean BMI 21.5) from separate Self-reported fatigue, eyestrain, 137 min day À 1 at workplace, workplaces; aged 20–65 years; baseline, self-rated work performance. 78 min during waking time 1-week and 3-month measurements At baseline and 3-months: Sitting time primarily replaced by fasting total cholesterol, HDL standing time, minimal stepping time. cholesterol, triglycerides, HDL cholesterol increased glucose levels. Self-reported headaches, digestion, sleep problems, musculoskeletal health and absenteeism. Gilson et al.50 Single group (4 men, 7 women, mean Installation of shared height-adjustable Objectively monitored NS differences BMI 25.9 kg m À 2 ) pretest posttest desks (equipped with necessary computer (Sensewear armband) time in design; baseline week and 1 week of technology) in open ofﬁce plan; employees sedentary, light and moderate intervention; encouraged to use any desk and stand and intensity categories. work as often as possible, costs not

& provided. 47

04McilnPbihr Limited Publishers Macmillan 2014 Pronk et al. Nonrandomized interrupted time Installation of individually dedicated Experience-sampling methods Frequency of sitting time decreased series design comparing intervention commercially available sit–stand (frequent sampling by cell during intervention. group (1 man, 22 women) and workstation (Ergotron WorkFit-S or WorkFit phone text messages) to Upper back and neck pain improved. comparison group (10 women), BMI not C, sit–stand Workstation); costs not monitoring frequency of sitting Improved mood states. reported, over 7 weeks; 1-week provided. behavior at work throughout 7 Removal of device negated most baseline, 4-week intervention, 2-week week-project. improvements. post-intervention Self-reported health-related outcomes, mood states and ofﬁce behaviors. NB: apparent differences in descriptive information reﬂect underlying discrepancies in reporting conventions used in original studies. & 04McilnPbihr iie nentoa ora fOeiy(04 5 765 – 755 (2014) Obesity of Journal International Limited Publishers Macmillan 2014

Table 4. Treadmill desk interventions

Reference Study design; sample (size, gender, age, Intervention Outcome measures Signiﬁcant ﬁndings occupation/workplace, BMI, salient inclusion/exclusion criteria, all if reported)

Thompson et al.9 Single group (n ¼ 25, gender and BMI (Assumed) installation of individually Objectively monitored (StepWatch Activity Relative to baseline, steps taken not reported); nurses, clinical assistants, dedicated treadmill (Pacemaster) and Monitor) steps taken during working hours. during working hours increased secretaries and appointment built-in accommodations for keyboard during acclimation period and secretaries; neither age nor gender and computer screen. Participants used regular use period. reported; 2-week baseline monitoring, as they wished. 2-week monitoring during acclimation and 2-week monitoring during regular use of workstation John et al.53 Single group; 5 men and 7 women; Installation of individually dedicated Objectively monitored (activPAL) steps per 3-month: steps per day increased mean age 46.2 years; mean BMI treadmill (T1450, Vision Fitness) and day and time spent sitting/lying, standing, 9-month: steps per day increased. 33.9 kg m À 2 , pretest posttest design; separately purchased height-adjustable stepping. Time spent sitting/lying decreased. measurements during baseline week, at table. Costs not provided. No Anthropometric measurements (height, Time spent standing, stepping 3-month and at 9-month recommendations given on frequency, weight, waist, hip). Bodpod-derived body increased. duration or speed of walking bouts. composition. Truncal fat mass (DXA). Waist and hip decreased. Resting heart rate, blood pressure. LDL and total cholesterol decreased. Lipid and metabolic profiles. HbA1c decreased. Dietary intake (24-hour recalls). Koepp et al.30 Single group; 11 men and 25 women; Installation of individually dedicated Objectively monitored (Actical) PA, 6-month: PA increased. mean age 42 years; mean BMI height-adjustable treadmill desk sedentary behavior. Sedentary behavior decreased. 29 kg m À 2 ; sedentary job; pretest (SteelCase,E$US3000–4000) Anthropometric measurements (height, 12-month: PA increased Tudor-Locke alternatives C workstation of expenditure Energy posttest design; measurements at weight). Bodpod-derived body Sedentary behavior decreased. baseline, at 6-month and at 12-month composition. Weight and waist circumference Lipid and metabolic profiles. decreased. Resting blood pressure. HDL increased. EE (indirect calorimetry) at rest, sitting, walking at three different speeds (not clear al et whether work was performed simultaneously). Self-reported and supervisor-reported work performance. Thompson et al.29 Randomized cross-over trial; 17/20 Exercise counseling, accelerometer with Objectively monitored PA, fitness (VO2 EE extrapolated from accelerometer physicians completed the study, 14 or without feedback and installation of max). Body composition (BMI, waist data increased by 197 kcal day À 1 men and 3 women; aged 25–70 years; individually dedicated treadmill desk to circumference, weight and body fat by over seated condition, 1.85 kg BMI425 kg m À 2 ; sedentary; able to walk at 1.6 km h À 1 while dictating, dual energy X-ray absorbtiometry). Cardiac weight loss and 1.9% body fat loss perform an exercise treadmill test. typing, talking on the phone and risk factors (cardiac c-reactive protein, using the treadmill desk. responding to e-mail. fasting lipid profile, fasting glucose, fasting No differences in metabolic or well- insulin, HA1c, blood pressure). being measures. Quality of Life Measures (NCHS General Well Being Schedule, Center for Epidemiologic Studies Depression Scale, the Linear Analog Self-Assessment Scale and Physician Survey). NB: apparent differences in descriptive information reflect underlying discrepancies in reporting conventions used in original studies. 761 Energy expenditure of workstation alternatives C Tudor-Locke et al 762 tasks. Funk et al.38 reported that of three treadmill speeds tested, 31 min day À 1 on 22.6 days used (37.7% compliance) as part of a 12- 17% of study participants preferred the 1.3 km h À 1 speed for typing, 46% week behaviorally enhanced multi-component intervention that also preferred 2.25 km h À 1 and 37% preferred 3.2 km h À 1. Straker et al.27 noted prescribed daily increases in pedometer-monitored steps per day. that some users commented that a 1.5 km h À 1 walking condition was In these latter two studies it appears that usage was mostly confined to ‘uncomfortably slow.’ Thirteen of 17 physicians completing a treadmill breaks (that is, lunch time) rather than concurrent use while working, desk intervention indicated that they would be interested in continuing to which may also explain the relatively lower compliance and use patterns 29 use this workstation alternative if it were available. compared with at least some of the treadmill desk studies. The tendency Workers experiencing an under-desk pedaling device reported that it towards 30 min day À 1 duration suggests that workers may be disposed to was easy to use, did not affect productivity or quality of work and was an adopting a usage pattern reminiscent of public health recommendations for acceptable alternative to activity during bad weather.11 More recently, Carr 32 exercise, regardless of the clear differences in intended purpose and et al. reported that pedal desk users believed that they could be easily executed intensity. used in a typical office work setting, would reduce sedentary behavior at Although merely replacing sitting with standing is believed to engender work and improve health. Participants also reported that they felt a new set of musculoskeletal complaints,51 no studies have reported comfortable maintaining a professional phone call, typing on a keyboard acute/overuse injuries or other reportable events associated with the use and using a computer mouse. Some pedal desk users have noted that it is of active workstation alternatives. This should not be surprising an improvement over the treadmill desk in terms of balance and as the typical use patterns9,11 cannot be described as physically stressful producing less upper body movement than what occurs during (biomechanically or physiologically demanding8), especially considering walking.27 However, some users have also reported minor leg or gluteal 27 the low-intensity nature of the PA enacted using active workstation discomfort related to the seat and others have reported that having their alternatives in real-world workplaces. knees hit against the underside of a conventional desk was problematic.49 Both of these concerns seem to be remediable design issues. There has only been one study thus far that has compared worker Are interventions designed around workstation alternatives perceptions across seated, standing, treadmill and pedal desk workstation 27 successful? alternatives. The study design required participants to walk at At this time there are four peer-reviewed standing/sit–stand desk speeds controlled externally by treadmill settings during that condition; interventions (Table 3), four treadmill desk interventions (Table 4) and however, they needed to actively and consciously maintain pedaling two multi-component interventions32,52 that have included active speeds required for the research application. This latter requisite was workstations. As cataloged in Tables 3 and 4 and below, interventions considered to be distracting by some, but it is unlikely that real workplace have varied considerably on design, duration, outcome measures implementation would continue such a rigid requirement. Straker et al.27 and reported significant findings. There has been only one randomized did note that a number of participants suggested that a self-selected cross-over controlled trial.29 The two multi-component interventions32,52 speed would be more comfortable than the set speeds used to standardize have been randomized controlled trials, but the actual effect of the active the investigation. Compared with the traditional seated working condition, workstations included in these studies is obscured within the intervention however, the standing workstation was ranked highest in terms of design. Across interventions, there have been 304 participants, of whom perceived feasibility (83%) for workplace implementation, followed by the 114 (37.5%) have been men and 190 (62.5%) have been women. pedal desk (63% thought it feasible, 13% thought it may be feasible) and 46,47 Focusing on the standing desk/sit–stand interventions, two have finally the treadmill desk (50% thought it feasible, 13% thought it may be studied individually dedicated sit–stand workstations and described feasible). In terms of overall preference, the standing desk and pedal desk significant findings in objectively monitored behavior and other outcomes. were equally preferred, and the treadmill desk was least frequently The third50 studied access to shared height-adjustable desks and noted no preferred. observed significant differences in objectively monitored behavior (no Evidence of adherence to use can also be interpreted as an indicator of 16 other outcomes measured). acceptability of workstation alternatives. Beers et al. reported that having 9,30,53 All four treadmill desk interventions have studied installation of been required to perform word-processing tasks for 20 min each while individually dedicated workstations and documented significant increases seated in an office chair, seated on a stability ball or while standing, in objectively monitored PA (variably measured and defined) and/or participants preferentially chose to do additional word-processing tasks in decreased time spent in sedentary behaviors. Somewhat confusingly, the traditional condition, followed by the stability ball condition and lastly Thompson et al.29 reported that physicians averaged B90 min day À 1 in the standing condition. Self-reported use of shared standing 1 walking on the treadmill at 1.6 km h À (which is considered light desk ranges from high frequencies (for example, daily) and durations 54 activity ); however, accelerometer data indicated that moderate (up to 3 hours at a time), to infrequent and short durations (0.2 h), to not þ intensity activity increased (exact amount not reported), which suggests at all.50 Wilks et al.45 reported low usage patterns of sit–stand desks after either a measurement error or a change in behavior beyond the use of the their introduction to four different companies, with modest increases in treadmill desk. In this randomized cross-over controlled study, Thompson utilization with education. et al.54 also showed that physician use of treadmill desks produced a small In a study of unrestricted access to individually dedicated treadmill but significant weight reduction of 1.85 kg over 12 weeks. They further desks,30 accelerometer-determined walking during working hours À 1 reported a significant weight gain (exact amount not reported) over the increased 128 min day over baseline values at 6 months and same time period in those physicians who ceased to use the treadmill desk decreased somewhat to 109 min day À 1 over baselines values at 9 as part of the study cross-over design. A significant difference in body fat 12 months (suggesting some waning of behavior). Thompson et al. percent between treadmill desk and control (usual seated work) behaviors reported that even the most avid user did not use the treadmill desk more À 1 was attributed primarily to a gain in body fat percent during the latter than 4 h day ; as indicated above, the modal duration of use was 30 min 53 À 1 32 condition. The two studies of longer term use (9 months and 12 day . Carr et al. asked novice pedal desk users following 30 min of use months30) have also tracked and reported changes in indicators of body to estimate their potential workplace usage under dedicated use or shared habitus, namely, weight ( À 1.4 kg on average)30 waist ( À 4 to 5.5 cm)30,53 use paradigms; 95.6% reported they would use it daily if they could have and hip ( À 4.8 cm)53 circumferences. John et al.53 reported significant one in their private office whereas only 11.1% reported they would use it improvements in LDL ( À 8mgdlÀ 1), total cholesterol ( À 11 mg dl À 1) and daily if access was shared in a common area. Despite this optimism, HbA1c ( À 0.4%), whereas Koepp et al.30 reported significant improvements however, in a 4-week study of workers’ use of individually dedicated in HDL ( þ 4mgdlÀ 1). 11 under-desk pedaling devices, participants only pedaled an average of All of the interventions presented in Tables 3 and 4 are focused just on 12.2 days (range 2–20 days) out of a possible 20 working days (61% providing access (primarily dedicated) to workstation alternatives. Straker compliance), and an average of 23.4 min at a ‘somewhat hard’ intensity on et al.55 have stated that access may be insufficient to encourage optimal days that they did use the device. The researchers reported that although use; education and behavioral support is likely also needed. Robertson average pedal time was maintained over the course of the study, the et al.56 demonstrated that workers assigned to ergonomics training with number of participants who actually used the devices declined over the their sit–stand desks reported fewer musculoskeletal and visual discomfort same time frame. Participants also perceived that their sedentary time at and had significantly higher performance scores (quantity and quality of work decreased as a result of using the device; however, there were no faxes sent) over a 15-day period than workers provided sit–stand desks associated significant differences in self-reported time spent sitting, with minimal training. One study reported that participants received standing or walking. More recently, Carr et al.49 reported usage of behavioral counseling and accelerometer-based behavioral feedback, but the same type of dedicated under-desk pedaling device averaging few other details were provided.29 None of the studies presented in

International Journal of Obesity (2014) 755 – 765 & 2014 Macmillan Publishers Limited Energy expenditure of workstation alternatives C Tudor-Locke et al 763 Tables 3 and 4 have rigorously evaluated methods of behavioral is necessary to operate treadmill workstations, and although this is counseling, automated prompts or managed behavior tracking. arguably trivial for individual use, it can become a cost and resource Two studies49,52 of multi-component interventions including active management consideration for larger-scale adoption in corporate work- workstations have been conducted. Parry et al.52 took a participative places. Gender-related preference for active workstations cannot be approach, taking into consideration physical and psychosocial features inferred from intervention participation at this time; apparent imbalances of organizations, to engage government clerical, call center and data may only reflect study recruitment strategies (for example, targeting processing workers in the design, implementation and evaluation of three traditionally stereo-typical male or female occupations). The need for different workplace interventions. One intervention included either appropriate footwear (that is, athletic or other type of comfortable shoe) treadmill desks or pedal desks and promotion of incidental office for extended standing and/or treadmill desk use (both in terms of comfort activity, another used pedometer challenges to increase activity outside but also noise control on treadmills) may also seem frivolous but is likely a of work and a third focused on re-designing office ergonomics and real barrier to some people’s uptake.46 The need for such footwear while breaking up computer tasks. Participants included in the final analysis (31 using a pedal desk has not been explicitly evaluated, but may be men, 31 women; mean age 43.5 years; mean BMI 28 kg m À 2) experienced unnecessary as it is a non-weight-bearing activity. For women, the decision significant reductions in sedentary behavior (both on work days and to wear a skirt or dress while using an active workstation likely remains a during working hours), increased number of breaks per sedentary hour personal choice within their unique workplace culture. A layered upper (again both on work days and during working hours), increases in light body (for example, shirt and sweater and/or jacket) will provide the activity during works hours and increases in moderate-to-vigorous PA on opportunity to quickly adjust body temperature as needed, depending work days. No significant differences were observed between groups; on workplace air conditioning or heating conventions and personal however, the authors noted that the participatory nature results in tolerance/preference. If active workstations are used as intended overlapping strategies between intervention groups. (for example, the top end of the SteelCase WalkStation is 3.2 km h À 1 Carr et al.49 provided 3 men and 20 women (mean age 42.6 years; mean and the top end of the LifeBalance Station is 65 r.p.m.), then sweating BMI 31.7 kg m À 2) access to a motivational website and sent email prompts should not be a problem.57 Cox et al.22 reported that the oxygen uptake and messages to encourage use of an under-desk pedal device as part of a of using a treadmill desk at 1.6 km h À 1 was 7–8 ml kgÀ 1 min À 1 and Carr 12-week multi-component study that also included a pedometer-based et al.32 reported that using a pedal desk at 40 r.p.m. was almost program. Objectively measured sedentary behavior significantly increased 6mlkgÀ 1 min À 1. As 1 MET (metabolic equivalent of resting oxygen within the waiting list control condition (1 man, 16 women; mean age consumption) is 3.5 ml kg À 1 min À 1, then these studies indicate that active 47.6 years, mean BMI 33.2 kg m À 2) and significantly decreased in the workstations represent approximately 2–2.5 MET activities, considered light multi-component intervention, producing a significant between-group intensity.54 None of the studies reviewed herein mentioned any concerns difference. There was also a significant between-group difference in waist with sweating. circumference change in favor of the multi-component intervention. No Although no study has reported injury related to use at this time, other within or between-group differences in objectively measured PA tripping/falling while using a treadmill desk remains a plausible low risk (any intensity) or cardiometabolic parameters were apparent. It is difficult and therefore employers must consider and weigh the potential for to attribute the differences in sedentary behavior and waist circumference insurance and litigation issues. Walking on a treadmill is likely a common to the pedaling device in this multi-component intervention, especially enough experience for most people to be able to judge their personal given the somewhat limited compliance ascribed specifically to this comfort and safety. The SteelCase WalkStation has an emergency stopping specific component (described above). feature in the event of a trip/fall. If advanced age and/or disease/ disability compromise gait and increase the potential for tripping/falling, then a treadmill desk may not be a prudent choice. Obviously the risk of Limitations and future directions tripping/falling would be less for pedal desk users. However, as all active workstation alternatives have moving parts, lifetime maintenance and Changing the way we work by incorporating active workstation replacement of the equipment must be considered. John et al.53 pointed alternatives presents an opportunity to re-balance the EE equation that out that some commercially available treadmills were not designed has been insidiously disrupted by modern conventions. There are, however, limitations and potential barriers to consider. Foremost is the for slow walking over prolonged durations and that such limited motors fact that although active workstations uniquely provide the opportunity to can become overloaded and lead to mechanical problems. raise EE relative to seated or standing-based workstations, they are also At least one of the solutions to some of the access/cost-based limitations is to provide shared desks, that is, allowing multiple users access generally more expensive. Although the internet is rife with homemade to a centrally located workstation alternative. Although shared access in a versions, many commercially available treadmill desks (for example, common area may introduce perceived embarrassment/discomfort using SteelCase WalkStation, approximately $4000) and elliptical pedal desks the active workstation in the presence of co-workers, this was directly (for example, LifeBalance Station, originally $8000, now described as $2795 queried and ultimately not considered a concern by novice pedal desk for the pedaling device and $900–1500 for an accompanying height- 49 adjustable desk32) are too expensive for wide-spread worksite adoption users. However, shared access to active workstations does present other unique logistical and human resource management problems including beyond niche individual use. These units also take up a large ‘footprint’ personal inconvenience, equity of scheduling and shared responsibility of that may or may not fit into the confined individual space allocations of maintenance. For example, a worker may shut down their usual desktop existing office designs other than more spacious executive suites. A lack of computer to maintain confidential files and move to the workstation portability is also a barrier to managing physical resources by easily or quickly moving such workstation alternatives between offices, meeting alternative only to find another worker there. Attempting to maximize use by rooms and other workplace facilities. These cost and convenience factors allocating scheduled use requires active management and also may be too likely underlie the fact that all of the interventions studied to date have rigid for workers who must respond to others’ (including clients’) impromptu requests. Supervisors may or may not view time away from dedicated work necessarily been focused on relatively small sample sizes. For example, 53 areas to use the workstation alternatives as productive time for individual John et al. divulged that their sample size of 12 was determined by the workers, which may lead to workplace friction. The repeated frustration of number of individual treadmill desks they could afford for their study. such scenarios could undermine use entirely. Finally, even the quietest active Further, their presence does not necessarily eliminate the need for a separate seated working area located in the same office space that is also workstations will likely make some degree of noise. Shared multiple units fully supported by requisite work equipment. Under such requirements, may or may not present a distraction in an open office design. With regard to future directions, more studies of EE are needed directly executives, eccentrics, self-employed workers and those with occupations comparing the different types of workstation alternatives currently that allow a great degree of autonomy and disposable income appear to be the primary commercial target for users. available, and especially under conditions of real use, including at self- Besides expense, the purchase of multiple units presents a human selected speeds. In addition, studies are needed to compare the relative resource management challenge to employers who must then plan how to acceptability of workstation alternatives in real-world workplaces; this equitably distribute such workstation alternatives among employees. should include tracked measures of actual use as an estimate of workers’ Equipment may also come with user size limitations, including weight tolerance for the behavior. As sit–stand/standing desks do not appear to restrictions, and thus employers must also consider this in terms of offer an advantage over traditional seated work, at least in terms of EE, workplace equity. The need to safely ambulate precludes some use by then their impact on other health outcomes must be candidly evaluated to workers who are susceptible to motion-related imbalance or have determine their relative utility and set the stage for comparative cost- orthopedic limitations or joint pain. Access to an electrical power source effectiveness studies.

& 2014 Macmillan Publishers Limited International Journal of Obesity (2014) 755 – 765 Energy expenditure of workstation alternatives C Tudor-Locke et al 764 Exercise as recommended by public health authorities58 should be of at (3.2 km h À 1 as indicated by device display, no incline; wearing business attire least moderate intensity. None of the workstation alternatives described and athletic shoes) and a pedal desk (52–56 r.p.m. and 40–44 watts as indicated herein meets that threshold, nor are they intended to do this. Active by Garmin Vector Power Meter Pedals, no resistance; wearing business attire workstations are not intended to replace exercise; they are intended to and a range of footwear including high heels). She says she doesn’t sweat. replace sedentary behavior. To emphasize by inculcation, although 30 min (Thompson et al. 9) or even 90 min (Thompson et al.29) of daily walking on a treadmill desk at 1.6 km h À 1 or even 3.2 km h À 127 should increase EE over the seated working condition,18 it does not obviate public health REFERENCES recommendations for PA that is of at least moderate intensity. The 1 Hamilton MT, Hamilton DG, Zderic TW. Role of low energy expenditure and sitting benefits of a physically active lifestyle extend beyond energy balance. in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Therefore, researchers should consider the impact of workstation Diabetes 2007; 56: 2655–2667. alternatives on other PA and sedentary behaviors performed (for 2 Thorp AA, Healy GN, Winkler E, Clark BK, Gardiner PA, Owen N et al. Prolonged example, priming or compensating) both at work and also in non-work sedentary time and physical activity in workplace and non-work contexts: a time. As active workstations do elevate EE, however, and thus represent a cross-sectional study of office, customer service and call centre employees. feasible strategy for addressing obesity (and related chronic diseases Int J Behav Nutr Phys Act 2012; 9: 128. including cardiovascular disease and type 2 diabetes1), mechanistic studies 3 Hu FB, Li TY, Colditz GA, Willett WC, Manson JE. Television watching and other of the impact of their use on heart health indicators and glycemic control sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in are also needed. Finally, implementation and process-oriented research is women. JAMA 2003; 289: 1785–1791. desperately needed in addition to outcome-based research conducted 4 Church TS, Thomas DM, Tudor-Locke C, Katzmarzyk PT, Earnest CP, under real-world conditions to better inform employers and workers about Rodarte RQ et al. Trends over 5 decades in U.S. occupation-related physical what works best for whom and under what conditions.41 activity and their associations with obesity. PLoS One 2011; 6: e19657. 5 Tsai AG, Williamson DF, Glick HA. Direct medical cost of overweight and obesity in the USA: a quantitative systematic review. Obes Rev 2011; 12: 50–61. CONCLUSIONS 6 Trogdon JG, Finkelstein EA, Hylands T, Dellea PS, Kamal-Bahl SJ. Indirect costs of Sedentary behavior is now recognized as a potential health obesity: a review of the current literature. Obes Rev 2008; 9: 489–500. threat.51,59 Although there have long been occupations 7 Cancelliere C, Cassidy JD, Ammendolia C, Cote P. Are workplace health promotion recognized as sedentary in nature (for example, seamstress/ programs effective at improving presenteeism in workers? A systematic review seamster, accountants, and so on),60 modern technological and best evidence synthesis of the literature. BMC Public Health 2011; 11: 395. advances (for example, computer hardware), task automation 8 Straker L, Mathiassen SE. Increased physical work loads in modern work--a necessity for better health and performance? Ergonomics 2009; 52: facilitated by increasingly sophisticated software, increasing forms 1215–1225. and numbers of electronic communication modes, individually 9 Thompson WG, Foster RC, Eide DS, Levine JA. Feasibility of a walking dedicated office equipment (for example, desktop printers) and workstation to increase daily walking. Br J Sports Med 2008; 42: 225–228, office furniture (for example, ergonomically comfortable chairs (discussion 228). 15 that actually increase sedentary behavior ) and design trends (for 10 McAlpine DA, Manohar CU, McCrady SK, Hensrud D, Levine JA. An office-place example, cubicle-based pods supported by their own kitchenettes, stepping device to promote workplace physical activity. Br J Sports Med 2007; 41: restrooms) have each likely contributed to the increasingly 903–907. sedentary nature of more and more types of occupations over 11 Carr LJ, Walaska KA, Marcus BH. Feasibility of a portable pedal exercise the last 50 years.4 Workstation alternatives that include sitting on a machine for reducing sedentary time in the workplace. Br J Sports Med 2012; 46: stability ball or engaging in standing-based work technically 430–435. 12 Osilla KC, Van Busum K, Schnyer C, Larkin JW, Eibner C, Mattke S. Systematic reduce sitting behaviors associated with traditional office work; review of the impact of worksite wellness programs. Am J Manag. Care 2012; 18: however, the evidence clearly shows that they do not generate a e68–e81. meaningful difference in EE. As the work day takes up a large 13 Proper KI, Koning M, van der Beek AJ, Hildebrandt VH, Bosscher RJ, amount of a full-time employed individual’s day, the health-based van Mechelen W. The effectiveness of worksite physical activity programs on need to obtain sufficient amounts of regular PA has shifted almost physical activity, physical fitness, and health. Clin J Sport Med 2003; 13: 106–117. entirely to the non-work committed time of the day, which is 14 Levine JA, Schleusner SJ, Jensen MD. Energy expenditure of nonexercise activity. already strewn with other competing personal, family, home and Am J Clin Nutr 2000; 72: 1451–1454. community obligations. Confronted with this sedentary behavior- 15 Edelson N, Danoff J. Walking on an electric treadmill while perfroming VDT office prompting and supporting milieu, it makes sense to enrich or work. SIGCHI Bulletin 1989; 21: 72–77. fortify the contemporary workplace with multiple opportunities 16 Beers EA, Roemmich JN, Epstein LH, Horvath PJ. Increasing passive energy expenditure during clerical work. Eur J Appl Physiol 2008; 103: 353–360. to replace sedentary behavior with low-intensity non-exercise 17 Reiff C, Marlatt K, Dengel DR. Difference in caloric expenditure in sitting versus PA that does not distract the workers from their primary task. standing desks. J Phys Act Health 2012; 9: 1009–1011. Active workstations are potentially important strategies for 18 Levine JA, Miller JM. The energy expenditure of using a ‘walk-and-work’ desk for mitigating the diminished EE inherent to contemporary office- office workers with obesity. Br J Sports Med 2007; 41: 558–561. based workplaces, but only if use is optimized and devices are 19 Swartz AM, Squires L, Strath SJ. Energy expenditure of interruptions to sedentary affordable and thus scalable. The science supporting active behavior. Int J Behav Nutr Phys Act 2011; 8:69. workstations is young and heterogeneous; this means that there 20 Speck RM, Schmitz KH. Energy expenditure comparison: a pilot study of standing are many research gaps and opportunities, including those instead of sitting at work for obesity prevention. Prev Med 2011; 52: 283–284. focused on implementation issues related to optimizing both 21 Peterman JE, Kram R, Byrnes WC. Factors affecting the increased energy expen- employers’ and workers’ uptake. diture during passive cycling. Eur J Appl Physiol 2012; 112: 3341–3348. 22 Cox RH, Guth J, Siekemeyer L, Kellems B, Brehm SB, Ohlinger CM. Metabolic cost and speech quality while using an active workstation. J Phys Act Health 2011; 8: CONFLICT OF INTEREST 332–339. 23 Hall KD. What is the required energy deficit per unit weight loss? Int J Obes (Lond) Dr Tudor-Locke and her husband, Mr Gerald R Locke, co-invented and thus own 2008; 32: 573–576. intellectual property for a workstation alternative device that is not presented in this 24 Pline KM, Madigan ML, Nussbaum MA. Influence of fatigue time and level on review of published literature. We are currently conducting research with the device. increases in postural sway. Ergonomics 2006; 49: 1639–1648. The remaining authors declare no conflict of interest. 25 Orendurff MS, Schoen JA, Bernatz GC, Segal AV, Klute GK. How humans walk: bout duration, steps per bout, and rest duration. J Rehabil Res Dev 2008; 45: 1077–1090. AUTHOR CONTRIBUTIONS 26 Straker LM, Abbott RA, Piek JP, Pollock CM, Davies PS, Smith AJ. Rationale, design and methods for a randomised and controlled trial to investigate whether home All authors contributed to this manuscript and approved its final content. CT-L access to electronic games decreases children’s physical activity. BMC Public led the writing of this manuscript while alternating between a treadmill desk Health 2009; 9: 212–212.

International Journal of Obesity (2014) 755 – 765 & 2014 Macmillan Publishers Limited Energy expenditure of workstation alternatives C Tudor-Locke et al 765 27 Straker L, Levine J, Campbell A. The effects of walking and cycling computer 47 Pronk NP, Katz AS, Lowry M, Payfer JR. Reducing Occupational sitting time and workstations on keyboard and mouse performance. Hum Factors 2009; 51: improving worker health: the Take-a-Stand Project, 2011. Prev Chronic Dis 2012; 9: 831–844. E154. 28 Hill JO, Wyatt HR, Reed GW, Peters JC. Obesity and the environment: where do we 48 Roelofs A, Straker L. The experience of musculoskeletal discomfort amongst go from here? Science 2003; 299: 853–855. bank tellers who just sit, just stand or sit and stant at work. Ergonomics 2002; 14: 29 Thompson WG, Koepp GA, Levine JA. Increasing physician activity with treadmill 11–29. desks. Work 2013; e-pub ahead of print 4 September 2013. 49 Carr LJ, Karvinen K, Peavler M, Smith R, Cangelosi K. Multicomponent intervention 30 Koepp GA, Manohar CU, McCrady-Spitzer SK, Ben-Ner A, Hamann DJ, to reduce daily sedentary time: a randomised controlled trial. BMJ Open 2013; 3: Runge CF et al. Treadmill desks: a 1-year prospective trial. Obesity (Silver Spring) e003261. 2013; 21: 705–711. 50 Gilson ND, Suppini A, Ryde GC, Brown HE, Brown WJ. Does the use of standing 31 Spruit MA, Singh SJ, Garvey C, Zuwallack R, Nici L, Rochester C et al. An official ’hot’ desks change sedentary work time in an open plan office? Prev. Med 2012; American Thoracic Society/European Respiratory Society Statement: key concepts 54: 65–67. and advances in pulmonary rehabiliation. Am J Respir Crit Care Med 2013; 188: 51 Gilson N, Straker L, Parry S. Occupational sitting: practitioner perceptions of e13–e64 in press. health risks, intervention strategies and influences. Health Promot J Austr 2012; 23: 32 Carr LJ, Maeda H, Luther B, Rider P, Tucker S, Leonhard C. Acceptability and effects 208–212. of a seated active workstation during sedentary work: A proof of concept study. 52 Parry SP, Straker L, Gilson ND, Smith AJ. Participatory workplace interventions can Inl J Workplace Health Manag (in press). reduce sedentary time for office workers—a randomized controlled trial. Plos One 33 Beurskens R, Bock O. Age-related deficits of dual-task walking: a review. Neural 2013; 8: e78957. Plast 2012; 2012: 131608. 53 John D, Thompson DL, Raynor H, Bielak K, Rider B, Bassett DR. Treadmill work- 34 Woollacott M, Shumway-Cook A. Attention and the control of posture and gait: a stations: a worksite physical activity intervention in overweight and obese office review of an emerging area of research. Gait Posture 2002; 16: 1–14. workers. J Phys Act Health 2011; 8: 1034–1043. 35 Mark JA. Measuring productivity in service industries. Mon Labor Rev 1982; 1982: 54 Ainsworth BE, Haskell WL, Herrmann SD, Meckes N, Bassett Jr DR, Tudor-Locke C 3–8. et al. 2011 Compendium of Physical Activities: a second update of codes and MET 36 Ebara T, Kubo T, Inoue T, Murasaki GI, Takeyama H, Sato T et al. Effects of values. Med Sci Sports Exerc 2011; 43: 1575–1581. adjustable sit-stand VDT workstations on workers’ musculoskeletal discomfort, 55 Straker L, Abbott RA, Heiden M, Mathiassen SE, Toomingas A. Sit-stand desks in alertness and performance. Ind Health 2008; 46: 497–505. call centres: associations of use and ergonomics awareness with sedentary 37 Ohlinger CM, Horn TS, Berg WP, Cox RH. The effect of active workstation behavior. Appl Ergon 2013; 44: 517–522. use on measures of cognition, attention, and motor skill. J Phys Act Health 2011; 8: 56 Robertson MM, Ciriello VM, Garabet AM. Office ergonomics training and a sit- 119–125. stand workstation: effects on musculoskeletal and visual symptoms and perfor- 38 Funk RE, Taylor ML, Creekmur CC, Ohlinger CM, Cox RH, Berg WP. Effect of walking mance of office workers. App. Ergon 2013; 44: 73–85. speed on typing performance using an active workstation. Percept Mot Skills 2012; 57 Montain SJ, Latzka WA, Sawka MN. Control of thermoregulatory sweating is 115: 309–318. altered by hydration level and exercise intensity. J Appl Physiol (1985) 1995; 79: 39 John D, Bassett D, Thompson D, Fairbrother J, Baldwin D. Effect of using a 1434–1439. treadmill workstation on performance of simulated office work tasks. JPhysAct 58 US Department of Health and Human Services. 2008 Physical Activity Guidelines for Health 2009; 6: 617–624. Americans: Be Active. Healthy, and Happy!: Washington, DC, USA, 2008, Report no.: 40 Thompson WG, Levine JA. Productivity of transcriptionists using a treadmill desk. ODPHP Publication No. U0036. Work 2011; 40: 473–477. 59 Katzmarzyk PT, Church TS, Craig CL, Bouchard C. Sitting time and mortality 41 Myers AM. Program Evaluation for Exercise Leaders. Human Kinetics: Champaign, from all causes, cardiovascular disease, and cancer. Med Sci Sports Exerc 2009; 41: IL, USA, 1999. 998–1005. 42 McGill SM, Kavcic NS, Harvey E. Sitting on a chair or an exercise ball: 60 Tudor-Locke C, Ainsworth BE, Washington TL, Troiano RP. Assigning metabolic various perspectives to guide decision making. Clin Biomech (Bristol, Avon) 2006; equivalent (MET) values to the 2002 Census Occupational Classification System. 21: 353–360. J Phys Act Health 2011; 8: 581–586. 43 Kingma I, van Dieen JH. Static and dynamic postural loadings during computer 61 Drury CG, Hsiao YL, Joseph C, Joshi S, Lapp J, Pennathur PR. Posture and perfor- work in females: sitting on an office chair versus sitting on an exercise ball. mance: sitting vs standing for security screening. Ergonomics 2008; 51: 290–307. Appl Ergon 2009; 40: 199–205. 62 Fidler JL, MacCarty RL, Swensen SJ, Huprich JE, Thompson WG, Hoskin TL et al. 44 Gregory DE, Dunk NM, Callaghan JP. Stability ball versus office chair: comparison Feasibility of using a walking workstation during CT image interpretation. JAm of muscle activation and lumbar spine posture during prolonged sitting. Hum Coll Radiol 2008; 5: 1130–1136. Factors 2006; 48: 142–153. 63 Husemann B, Von Mach CY, Borsotto D, Zepf KI, Scharnbacher J. Comparisons of 45 Wilks S, Mortimer M, Nylen P. The introduction of sit-stand worktables; aspects of musculoskeletal complaints and data entry between a sitting and a sit-stand attitudes, compliance and satisfaction. Appl Ergon 2006; 37: 359–365. workstation paradigm. Hum Factors 2009; 51: 310–320. 46 Alkhajah TA, Reeves MM, Eakin EG, Winkler EA, Owen N, Healy GN. Sit-stand 64 Alderman BL, Olson RL, Mattina DM. Cognitive function during low intensity workstations: a pilot intervention to reduce office sitting time. Am J Prev Med walking: a test of the treadmill workstation. J Phys Act Health 2013; e-pub ahead of 2012; 43: 298–303. print 14 March 2013.

& 2014 Macmillan Publishers Limited International Journal of Obesity (2014) 755 – 765