Applied Ergonomics 45 (2014) 1240e1246
Contents lists available at ScienceDirect
Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo
A cycling workstation to facilitate physical activity in office settings Steven J. Elmer a, b, c, *, James C. Martin d a
Department of Exercise Science and STEM Education, University of Maine, 5740 Lengyel Hall, Orono, ME 04469, USA Department of Mechanical Engineering, University of Maine, Orono, ME, USA c Eastern Maine Medical Center, Bangor, ME, USA d Department of Exercise and Sport Science, University of Utah, Salt Lake City, UT, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 December 2013 Accepted 6 March 2014
Facilitating physical activity during the workday may help desk-bound workers reduce risks associated with sedentary behavior. We 1) evaluated the efficacy of a cycling workstation to increase energy expenditure while performing a typing task and 2) fabricated a power measurement system to determine the accuracy and reliability of an exercise cycle. Ten individuals performed 10 min trials of sitting while typing (SITtype) and pedaling while typing (PEDtype). Expired gases were recorded and typing performance was assessed. Metabolic cost during PEDtype was w2.5 greater compared to SITtype (255 14 vs. 100 11 kcal h1, P < 0.01). Typing time and number of typing errors did not differ between PEDtype and SITtype (7.7 1.5 vs. 7.6 1.6 min, P ¼ 0.51, 3.3 4.6 vs. 3.8 2.7 errors, P ¼ 0.80). The exercise cycle overestimated power by 14e138% compared to actual power but actual power was reliable (r ¼ 0.998, P < 0.01). A cycling workstation can facilitate physical activity without compromising typing performance. The exercise cycle’s inaccuracy could be misleading to users. Ó 2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.
Keywords: Workplace activity Energy expenditure Ergometer calibration
1. Introduction Thirty-five percent of adults in the United States are currently obese and at increased risk of developing several metabolic, cardiovascular, and psychological disorders (Haskell et al., 2007; Mutrie, 2001). Health care costs attributed to obesity are approximated to be $147 billion per year in the United States and those costs are predicted to increase to $200 billion per year over the next two decades (Finkelstein et al., 2009). Obesity results from a longterm excess of energy consumed vs. energy expended, a positive energy balance (Church et al., 2011). While overeating certainly contributes to a positive energy balance, the epidemic of obesity and related metabolic disorders is also driven by reductions in energy expended (Ford et al., 2012). With increased use of computers, office workers may remain seated and sedentary (Barnes et al., 2012) during which energy expenditure is minimal, making many offices obesogenic environments (Mummery et al., 2005). Sedentary behavior not only contributes to positive energy balance but is also an independent risk factor for diabetes, cardiovascular
* Corresponding author. Department of Exercise Science and STEM Education, University of Maine, 5740 Lengyel Hall, Orono, ME 04469, USA. Tel.: þ1 207 581 3314; fax: þ1 207 581 1206. E-mail address: [email protected] (S.J. Elmer). http://dx.doi.org/10.1016/j.apergo.2014.03.001 0003-6870/Ó 2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.
disease, and all-cause mortality (Wilmot et al., 2012). Therefore, increasing physical activity and/or decreasing sedentary time are paramount to reducing obesity and associated disorders. However, traditional strategies to increase physical activity are compromised by insufficient time for exercise, perception that exercise is boring, concern about appearance during exercise, and fatigue after the workday (Mayo Clinic, 2011). These barriers could be reduced by facilitating physical activity during working hours. Treadmill workstations have been promoted as means to increase physical activity during working hours. John et al. (2011) reported that office workers, provided with a treadmill workstation for nine months, increased their walking time by 38e 75 min day1, compared to baseline, and averaged a walking speed of 1.5 mph. This duration and speed suggests an energy expenditure of approximately 100e200 kcal day1, above baseline (Levine and Miller, 2007). Compared to baseline, the participants in this study exhibited significant decreases in waist and hip circumference as well as improvements in blood lipids and cholesterol. Similarly, Koepp et al. (2013) recently reported that office workers who were provided with a treadmill workstation for one year increased physical activity and decreased weight, while maintaining productivity (Koepp et al., 2013). This finding of unchanged productivity is important because other authors have reported that use of a treadmill workstation may compromise computer use accuracy (Funk et al., 2012; John et al., 2009) and/or work productivity
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
(Thompson and Levine, 2011). A potential limitation of treadmill workstation use is that individuals who are overweight and obese are at increased risk for knee osteoarthritis (Felson et al., 1988) and thus may have limited ability to walk. As a non-weight bearing activity, cycling could serve as an alternative modality for facilitating increased physical activity in desk-bound office workers. Indeed, Straker et al. (2009) evaluated the effects of a cycling workstation (computer desk and an upright exercise cycle) on physiological responses and computer operation in office workers. Specifically, these authors reported that heart rate values increased by w25% when pedaling and typing compared to normal sitting and typing. Additionally, these authors reported that typing performance while cycling was slightly compromised but was better than typing performance during treadmill walking. To the best of our knowledge, no previous authors have reported the actual rate of energy expenditure that individuals might self-select for pedaling during an office task such as typing. It is also important to note that the previous cycling workstation design used by Straker et al. (2009) consisted of an upright exercise cycle with a relatively small seat. This design may present some ergonomic challenges, especially when using it for extended durations, as several participants in that study (9 out of 30) reported some form of hip or gluteal discomfort related to the exercise cycle seat. Further, such bicycle seats are known to reduce penile oxygen pressure, reflecting perineal compression (Schwarzer et al., 2002) and prolonged sitting on such seats can lead to perineal numbness and erectile dysfunction (Dettori et al., 2004). Thus, a cycling workstation with a recumbent seat position that is more similar to a standard office chair may facilitate prolonged activity without risks associated with upright cycling. Our primary purposes for conducting this investigation were to 1) determine the metabolic cost associated with self-selected pedaling intensity while performing a standardized typing task and 2) assess the influence of pedaling on typing performance (typing time and number of typing errors). These responses were compared to those associated with normal sitting while typing. We hypothesized that compared to sitting while typing, pedaling while typing would increase metabolic cost without compromising typing performance. Additionally, because the accuracy and reliability of commercial exercise cycles to quantify power are not documented, our secondary purposes were to 1) fabricate a power measurement system to determine the actual power associated with pedaling the exercise cycle across a range of resistance levels and pedaling rates before and after the experimental trials and 2) compare the actual power, as determined by the power measurement system, to the power displayed by the exercise cycle console. This study could serve as an important first step to utilizing a cycling workstation to reduce sedentary time and increase physical activity in office settings. 2. Methods 2.1. Participants Ten healthy males volunteered to participate in this study (age: 32 8 yr, body mass: 72 8 kg, height: 1.78 0.07 m, body mass index: 23 2 kg m2). Participants were college students and faculty who spent considerable time sitting at a desk each day (6.3 2.4 h day1, self-reported). Participants were also recreationally active in a variety of sports and familiar with cycling exercise (both upright and recumbent cycling positions). Experimental procedures used in this investigation were reviewed by the University of Utah Institutional Review Board. The protocol and procedures were explained verbally and all participants provided written informed consent prior to testing.
1241
2.2. Experimental protocol Participants reported to the laboratory dressed in typical school or office attire and were instructed to refrain from eating for at least 3 h prior to testing. During this visit resting baseline data were first recorded. Subsequently, participants performed two experimental trials (10 min) while physiological responses were recorded (described below): 1) Sitting while typing 2) Pedaling while typing For the sitting while typing condition, participants were positioned on the cycling workstation (but did not pedal) and performed a standardized typing task (described below). For the pedaling condition, participants were positioned on the cycling workstation and pedaled while they performed the same standardized typing task. The two experimental trials were presented in a random order. At least 5 min of recovery was provided between each trial to allow physiological responses to return to baseline levels. 2.3. Typing task To quantify typing performance, we measured the time required to transcribe the Gettysburg Address, a famous speech delivered by United States President Abraham Lincoln in 1863, which participants practiced typing twice prior to the experimental protocol. For this task, a copy of the Gettysburg Address was placed adjacent to the computer screen and participants were instructed to transcribe the Gettysburg Address at a comfortable typing speed, a speed at which they typically performed school and/or office-related typing. Participants were also instructed to correct any typing errors that might have occurred as they typed so that the overall quality of the document was equivalent to that of school and/or office work. Because all participants regularly used a word processing autocorrect spelling/grammar feature this setting was enabled during the standardized typing task. The total time required to transcribe the Gettysburg Address and correct any potential errors was recorded. In addition, an investigator carefully reviewed the document and recorded the actual number of typing errors. If participants completed the typing task before the 10 min period was over then they were simply asked to start the typing task again but the second round was not timed or completed. 2.4. Cycling workstation A cycling workstation was constructed using a commercially available recumbent exercise cycle (Model R3i, LifeSpan Fitness, Salt Lake City, UT, USA), a set of adjustable height table legs (Model UpLift 700, The Human Solution, Austin, TX, USA), and a custom made keyboard tray with integrated arm support (Fig. 1). The seat pan of the exercise cycle was 44 cm wide, 26 cm deep, and positioned at an inclination of w1 (positive to horizontal). The back support was 40 cm wide, 52 cm tall, and positioned at an inclination of w15 (backward of vertical). The standard exercise cycle cranks (170 mm) were replaced with shorter cranks (114 mm, Unicycle.com) to facilitate usable ergonomics. Participants selected a seat position that allowed a slight bend at the knee when the leg was in its most extended position. The desk height was then adjusted to be as low as possible while still providing clearance for the knees during pedaling. The cutout in the keyboard tray was approximately 24 cm deep and 41 cm wide (Fig. 1). This depth allowed the tray to extend to the anterior-posterior midline of the torso and thus provide support for the elbows when the upper arm
1242
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
Fig. 1. Overview of the cycling workstation which consisted of a 1) commercially available recumbent exercise cycle, 2) set of adjustable height table legs, and 3) a custom made keyboard tray with integrated arm support that extended to the anterior-posterior midline of the torso. Participants pedaled at a resistance level and pedaling rate that they could comfortably maintain while performing a standardized typing task.
was in a near vertical orientation. Forearm/elbow support provided by the custom keyboard tray was an integral aspect of the ergonomic solution intended to stabilize the arms and reduce the requirement for muscular elevation (see discussion for further explanation). After these adjustments were made, participants were given time to become familiar with pedaling the cycling workstation. That is, participants were instructed to select a resistance level that they could comfortably maintain for prolonged periods while performing computer tasks. During the experimental pedaling condition, the preferred resistance level was set on the exercise cycle console and participants pedaled at that resistance level for 10 min while they were free to modulate their pedaling rate. 2.5. Power measurement system The commercial exercise cycle used in this study was equipped with a console that displayed, and recorded resistance level, pedaling rate, and power. The accuracy and reliability of power determined by the console of commercial exercise cycles in general are not well known. Therefore, prior to experimental data collection we determined the actual power required to drive the cranks at a range of pedaling rates and resistance settings. To facilitate power measurement, we constructed a system consisting of an electric motor, variable frequency drive, and a power meter (Fig. 2). Specifically, a 1.1 kW, 100 rpm, three-phase, electric gear motor (Baldor Electric Co. Model Syncromotor, Ft. Smith, AR, USA) was used to drive the power meter (Schoberer Rad Messtechnik, SRM, Jülich, Germany) via a chain-drive system. The power meter was connected in series to a driveshaft which attached to the exercise cycle crank axle. Therefore, the electric motor drove the power meter which ultimately “pedaled” the exercise cycle. Motor speed and thus pedaling rate were controlled by a variable frequency drive (TECO/Westinghouse, Model FM50-101-C, Round Rock, TX, USA). Thus, power required to pedal the exercise cycle was quantified using the power meter, a system that serves as an accurate method
Fig. 2. Schematic of the power measurement system used to determine the power required to drive the exercise cycle at a range of pedaling rates and resistance settings. An electric motor was connected to a power meter (light gray circle) via a chain drive system (dashed line). The power meter was connected in series to a driveshaft that attached to the exercise cycle. Motor speed and pedaling rate were controlled using a variable frequency drive (dark gray box). Thus, the electric motor drove the power meter which ultimately “pedaled” the exercise cycle.
to quantify cycling power (Abbiss et al., 2009; Gardner et al., 2004; Martin et al., 1998). Prior to this power evaluation, the power meter itself was calibrated using static calibration procedures (Wooles et al., 2005). With this power measurement system we evaluated the power required to pedal the exercise cycle at the lowest seven resistance levels at 40, 50, 60, 70, and 80 rpm. From these data, powerpedaling rate relationships were determined for each resistance level using linear regression. During the experimental pedaling condition, pedaling rate was measured using the exercise cycle console and actual power produced by the participant was calculated using the power-pedaling rate regression equations. We also determined the relationship between the actual power determined by our power measurement system (i.e., gold standard) and the power determined by the exercise cycle console, as this could be important for other researchers using commercial exercise cycles. Finally, this procedure was again performed after the completion of experimental pedaling trial in order to assess the test-retest reliability of the exercise cycle. 2.6. Metabolic, cardiorespiratory, and perceptual measures Gas exchange data were measured during each experimental condition using open circuit spirometry (True Max 2400, Parvo Medics, Sandy, UT, USA). The metabolic system was calibrated with a 3 L calibration syringe (Hans Rudolph, Kansas City, MO, USA) and medical gases of known concentrations (16.00% O2, 4.02% CO2, _ Þ, carbon dioxide probalanced N2). Oxygen consumption ðVO 2 _ _ duction ðVCO 2 Þ, minute ventilation ðVE Þ, respiratory exchange ratio (RER), and metabolic equivalents (METS) data were recorded every _ 15 s and averaged over min 5e10 of each condition. Absolute VO 2 1 and RER data were used to calculate metabolic cost (kcal h ) using the regression equation reported by Zuntz (1901) which is based upon the thermal equivalent of O2 for nonprotein respiratory equivalent. Heart rate was measured using a Timex monitor (Timex Ironman, Timex Group USA, Middlebury, CT, USA) and averaged
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
over min 5e10 of each trial. Finally, overall rating of perceived exertion (RPE) was assessed during min 4e5 and 9e10 using a Borg 6e20 scale (Borg, 1970).
Power-pedaling rate relationships determined from the power measurement system were highly linear for each resistance level evaluated (Fig. 5). Relationships between the actual power and power displayed by the exercise cycle console indicated that the exercise cycle console grossly overestimated power (Fig. 5 and Table 1). Specifically, mean power differences at the lowest resistance setting ranged from 3 to 25 W (14e47%) across the pedaling rates evaluated (40, 50, 60, 70, 80 rpm) and mean differences at the highest resistance setting ranged from 42 to 145 W (108e127%) across pedaling rates (Table 1). Power displayed by the exercise cycle console for each resistance setting and pedaling rate was identical before and after the experimental pedaling trial,
Power (W)
40 20
2
4
6
8
10
1.2 1.0 -1
*
0.8 0.6 0.4 0.2 0.0 0
2
4
6
8
10
30 25
VE (L∙min )
.
-1
3.2. Power measurements
60
0
*
20 15 10 5 0 0
2
4
6
8
10
100 -1
Heart Rate (b∙min )
All ten participants completed both the pedaling while typing and sitting while typing experimental trials. However, one participant accidently missed transcribing a phrase during the typing trial. Thus, values for typing time and number of typing errors were reported in only nine participants. Participants selected to pedal at a resistance level of 4 1 on the exercise cycle. Mean pedaling rate and power produced during the pedaling while typing trial were 51 14 rpm and 38 14 W (actual power based on power-pedaling rate relationships described below), respectively. Oxygen consumption, V_ E , and heart rate were substantially greater during pedaling while typing compared to sitting while typing (all P < 0.01, Fig. 3). Similarly, metabolic cost and METs were w2.5 times greater during pedaling while typing compared to sitting while typing (both P < 0.01, Fig. 4). Despite these differences, the time required to complete the standardized typing task and number of typing errors during pedaling while typing did not differ from that during sitting while typing (7.7 1.5 vs. 7.6 1.6 min, P ¼ 0.51; 3.3 4.6 vs. 3.8 2.7 errors, P ¼ 0.80, respectively). Compared to sitting while typing, RPE was slightly higher during pedaling while typing at 5 min (6.9 0.8 vs. 6.1 0.3 scale units, P < 0.01) and 10 min (6.9 0.7 vs. 6.1 0.3 scale units, P < 0.01).
Pedaling
0
3. Results 3.1. Physiological responses and typing performance
Sitting
.
Separate one-way repeated measures analysis of variance _ , V_ E , (ANOVA) procedures were used to assess differences in VO 2 and heart rate values during the 10 min experimental trials (sitting while typing vs. pedaling while typing). Additionally, paired student’s t-tests were used to compare differences in metabolic cost, METS, and RPE between sitting while typing and pedaling while typing trials. Separate paired student’s t-test were used to compare differences in the time required to complete the standardized typing task and number of typing errors between sitting while typing and pedaling while typing trials. Power-pedaling rate relationships determined for each resistance level with the power measurement system are reported for descriptive purposes. To assess the accuracy of the exercise cycle console to measure power, mean differences of actual power data and the power displayed on the exercise cycle console were determined. Further, the test-retest reliability of the exercise cycle was evaluated using a Pearson productemoment correlation. Data are presented as mean standard deviation and alpha was set to 0.05.
VO2 (L∙min )
2.7. Data analysis
80
1243
90
*
80 70 60 50 0
2
4
6
8
10
Time (min) Fig. 3. Time course of alterations for selected mechanical, metabolic, and cardiorespiratory variables during sitting while typing and pedaling while typing. Values are reported as Mean SD. *P < 0.01 vs. sitting while typing (main effect of exercise condition).
suggesting that firmware in the console bases the power calculation on measured pedaling rate and indicated resistance level. Further, power values determined by the power measurement system before and after the experimental pedaling trial were highly reliable as demonstrated by a Pearson productemoment correlation coefficient of 0.998 (P < 0.01). This suggests that the exercise cycle resistance settings, which control the position of magnets relative to the flywheel, were highly repeatable. Note that, to improve the clarity of Fig. 5 relationships were illustrated for only those resistance levels utilized by the participants in this investigation (resistance levels 2, 3, 4, and 5).
1244
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
A
A
150
*
300
125
250
Actual Power (W)
-1
Metabolic Cost (kcal∙h )
350
200 150 100 50
100 Level 2 Level 3 Level 4 Level 5
75
50
0
Sitting 4.0
B
Pedaling
25
*
0
3.5
30
40
2.5
METS
50
60
70
80
90
Pedaling Rate (rpm)
3.0
B
250
2.0 1.5
0.5 0.0
Sitting
Pedaling
Fig. 4. Comparison of metabolic cost (A) and metabolic equivalents (METS) (B) between sitting while typing and pedaling while typing. Dashed line represents baseline sitting values (no typing). Values are reported as Mean SD. *P < 0.01 vs. sitting while typing.
4. Discussion An important driving force in the obesity epidemic is the decrease in physical activity during working hours (Church et al., 2011). Consequently, facilitating physical activity in the workplace is essential in combating obesity and improving worker health. Employers have a vested interest in employee heath because they ultimately bear the costs of most insurance programs (Kate Bundorf, 2002). In this study, we evaluated the use of a cycling workstation to facilitate physical activity and its effect on typing performance. With our modified cycling workstation participants chose to pedal at a power output of 38 W (actual power) while performing a standardized typing task, an intensity they rated as “very very light” exertion. These results support previous reports documenting the use of a cycling workstation (Straker et al., 2009) in which participants pedaled at 30 W while performing typing task and rated their intensity between “fairly light” and “moderate”. Further, metabolic rates during pedaling at 38 W in the current study (240 kcal h1) compare favorably to those estimated during treadmill walking at 1.5 mph (w162 kcal h1) (John et al., 2011; Levine and Miller, 2007). Importantly, pedaling at this intensity was not uncomfortable for participants and did not significantly compromise the time required to type The Gettysburg Address or number of typing errors. Taken together, these results indicate that a cycling workstation may provide an additional viable option for increasing physical activity in desk-bound workers without compromising typing performance.
Display Power (W)
200
1.0
Level 2 Level 3 Level 4 Level 5
150
100
50
0 0
50
100
150
200
250
Actual Power (W) Fig. 5. Power-pedaling rate relationships for the exercise cycle. Power required to “pedal” the exercise cycle for several resistance levels and pedaling rates as measured by the calibrated SRM power meter (A). Power-pedaling rate relationships were essentially linear which implies a constant resistant torque provided by the exercise cycle internal braking system. Comparison of actual power required (as measured by the calibrated SRM power meter) with power displayed on the exercise cycle control console (B). The exercise cycle console overestimated power in all conditions and the error ranged from 14 to 138%. For the settings and pedaling rates used by participants in this study while typing, power displayed by the console was approximately twice the actual power.
4.1. Potential outcomes Pedaling while typing on a cycling workstation elevated metabolic cost by 155 kcal h1 compared to typing alone. This additional energy expenditure could allow desk-bound office workers to reduce sedentary time which can significantly improve glucose and insulin tolerance (Dunstan et al., 2012) and reduce the risk of diabetes (Wilmot et al., 2012). Prolonged use of the cycling workstation at this intensity could also dramatically influence energy balance and replace workplace physical activity that has been lost over the last 50 years (Church et al., 2011). For example, pedaling 10 min every hour throughout an 8 h workday (80 min day1) could
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
1245
Table 1 Differences in actual power measured from the power measurement system and power displayed on the exercise cycle console. Note that, positive values indicate that the exercise cycle console power was greater than the actual power. Pedaling rate
40 50 60 70 80
Resistance level 1
2
3W (14%) 9W (31%) 14W (38%) 21W (47%) 25W (45%)
12W 20W 32W 39W 51W
3 (54%) (65%) (86%) (76%) (84%)
22W 31W 45W 59W 78W
(100%) (88%) (102%) (105%) (118%)
result in 200 kcal day1 or 1000 kcal wk1 of additional energy expenditure compared to typing alone. Pedaling intermittently for two to 4 h each day at this intensity would increase energy expenditure by 300e600 kcal day1 or 1500e3000 kcal wk1 compared to typing alone. One thousand kcal wk1 is significant because it meets the Center for Disease Control and Prevention (CDC) criteria for moderate physical activity (Health and Human Services, 2008). Assuming one pound of weight loss could result from each 3500 kcal of negative energy balance, these values of 200e600 kcal day1 could facilitate weight loss of approximately one pound every 6 to 17 workdays, or 15 to 45 pounds per year, provided energy intake and other physical activity remained constant. In addition to weight loss, physical activity per se is known to reduce the risk of cardiovascular disease (Morris et al., 1953; Weller and Corey, 1998). Therefore, use of a cycling workstation could dramatically reduce weight and improve health of desk-bound workers. On a similar note, facilitating physical activity to maintain and prevent weight gain is also important (Church et al., 2011) and staying active can improve cognitive function (Weuve et al., 2004). An important next step would be to determine the selfselected intensity and actual daily use of cycling workstations by overweight and obese individuals. 4.2. Typing performance Despite employers’ vested interest in employee health, some may be reluctant to facilitate physical activity if they believe it will compromise productivity. Previous authors have reported mixed results for computer operation (typing time, speed, word count, and/or accuracy) with cycling workstations (Straker et al., 2009), treadmill workstations (Funk et al., 2012; Koepp et al., 2013; Straker et al., 2009; Thompson and Levine, 2011), and other novel workstation modalities (Beers et al., 2008) when compared to normal sitting. Our finding of similar typing time agrees with some previous reports for physical activity during computer operation (Funk et al., 2012; Koepp et al., 2013) but contrast with others which indicated small but significant decrements in typing speed and accuracy (Funk et al., 2012; John et al., 2009; Straker et al., 2009; Thompson and Levine, 2011). These varied results likely reflect the degree to which various users could compensate for additional task complexity. That is, while walking the entire body is moving and thus complex control of the arms and hands is required to keep them stationary with respect to the keyboard. Cycling in an upright position on a narrow saddle may induce upper body movement via hip movement (Neptune and Hull, 1995) thereby also increasing task complexity. With our cycling workstation that included a recumbent seat system, upper body movement induced by pedaling was likely reduced compared to walking or upright cycling, and the arms were further stabilized by the supportive keyboard tray. Other novel workstation modalities for increasing physical activity while reducing upper body movement (e.g., seated stepping in place (McAlpine et al., 2007)), have not reported effects on computer operation and thus their feasibility in a workplace
4
5
6
7
27W (104%) 41W (105%) 59W (118%) 81W (135%) 101W (135%)
34W (117%) 48W (107%) 69W (119%) 94W (133%) 119W (138%)
36W (100%) 54W (102%) 80W (121%) 105W (125%) 135W (135%)
42W (108%) 63W (109%) 86W (110%) 119W (127%) 145W (121%)
setting are unknown. Finally, it is important to note that these results highlighting the efficacy of a cycling workstation are limited to transcribe typing which is only one task that might be performed during computer operation. 4.3. Ergonomic solutions Establishing usable ergonomics with a recumbent cycling workstation posed several challenges. First, when pedaling a recumbent exercise cycle the knees will be higher, relative to the shoulders, than they would when seated in a typical office chair. This requires that the tabletop be raised high enough to allow pedaling without the knees hitting the table. Because the tabletop and keyboard must be relatively higher than normal, the user might be forced to use muscular effort to hold up the arms and shoulders. This could result in fatigue and/or discomfort of those muscles. This issue was reduced, but not eliminated, by using shorter-than-normal cycle cranks to reduce the required desktop height. Further, the pedaling action induces a slight swaying motion of the torso which could increase task complexity for typing, resulting in decreased office performance. Our solution to these issues was to use a custom made keyboard tray with extended surface to support and stabilize the arms. This level of customization was essential as most computer desks have keyboard designs that do not extend far enough to adequately support the arms and minimize requirement for muscular elevation. It is also important to point out that the seat of the cycling workstation used in this investigation was reasonably similar to that of a standard office chair. Thus, workers could remain seated while not pedaling and resume pedaling at will. Moreover, none of the participants reported any noticeable seat-related discomfort which was a concern with previous cycling workstation designs (Straker et al., 2009). However, the extent to which long term use of the cycling workstation elicits any discomfort and/or fatigue remains to be investigated. 4.4. Power measurements We used a power meter-based system to determine powerpedaling rate relationships for the exercise cycle resistance settings used in this study. SRM power meters are known to be accurate and reliable (Abbiss et al., 2009; Gardner et al., 2004; Martin et al., 1998) and thus we are confident in our measures of actual power. Results demonstrated that power-pedaling rate relationships were highly linear for each resistance setting indicating that the current brake system provides nearly constant torque for any specific resistance (position of the magnets relative to the flywheel). Further, the agreement of our pre- and post-test power values indicates that the resistance settings of the exercise cycle were reliable. While the exercise cycle was reliable, the large differences in actual power and the power displayed on the exercise cycle console were unexpected. Specifically, the console power was generally over twice as great as the actual power except for the very lowest resistant settings. This could be important to users because
1246
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
the console also presents a value for energy expenditure which is based on the summation of power over time. This level of inaccuracy could be very misleading to a user who might carefully track energy expenditure while pedaling the exercise cycle. For example an individual might believe that he/she has burned a given number of kcals when in fact the total is less than half of that. Thus, users should be cautious in their expectations of accuracy for this specific commercial exercise cycle. While we only have actual power data on the exercise cycle used in this investigation we intend to determine accuracy of a variety of exercise cycles found in commercial exercise facilities. 4.5. Summary These data demonstrate the efficacy of a cycling workstation to facilitate physical activity without compromising typing performance. Use of a cycling workstation for 80 min to 4 h per day could result in significant weight loss over time and have additional health benefits associated with reduced sedentary time (Barnes et al., 2012). With its office-chair-like seat, a cycling workstation may provide a useful alternative to treadmill desks for those who would prefer non-weight-bearing physical activity. The specific exercise cycle we used exhibited good reliability as assessed by a power meter-based measurement system but the power displayed by the exercise cycle console and, thus all the calculations based on power, were inaccurate and tended to be around twice the actual values. Together, these results may have implications for prevention and/or treatment of overweight and obesity and associated health risks in desk-bound office works as well individuals who use commercial exercise cycles for recreational purposes. Funding source No outside funding was received for this study. Conflict of interest Steven Elmer declares no conflict of interest. James Martin has disclosed a potential conflict of interest. He is the inventor of intellectual property related to a cycling workstation for which the University of Utah has filed a provisional patent, and he holds equity in a company that has licensed the rights to this intellectual property from the University of Utah. Acknowledgments We thank the participants for their enthusiastic efforts during the experimental trials and Alex Seegmiller for assistance with the data collection and power measurement trials. References Abbiss, C.R., Quod, M.J., Levin, G., Martin, D.T., Laursen, P.B., 2009. Accuracy of the Velotron ergometer and SRM power meter. Int. J. Sports Med. 30, 107e112. Barnes, J., Behrens, T.K., Benden, M.E., Biddle, S., Bond, D., Brassard, P., Brown, H., Carr, L., Chaput, J.-P., Christian, H., 2012. Letter to the Editor: standardized use of the terms “sedentary” and “sedentary behaviours”. Appl. Physiol. Nutr. Metab. 37, 540e542. Beers, E.A., Roemmich, J.N., Epstein, L.H., Horvath, P.J., 2008. Increasing passive energy expenditure during clerical work. Eur. J. Appl. Physiol. 103, 353e360. Borg, G., 1970. Perceived exertion as an indicator of somatic stress. Scand. J. Rehabil. Med., 92e98. Church, T.S., Thomas, D.M., Tudor-Locke, C., Katzmarzyk, P.T., Earnest, C.P., Rodarte, R.Q., Martin, C.K., Blair, S.N., Bouchard, C., 2011. Trends over 5 decades in U.S. occupation-related physical activity and their associations with obesity. PloS One 6, e19657.
Dettori, J.R., Koepsell, T.D., Cummings, P., Corman, J.M., 2004. Erectile dysfunction after a long-distance cycling event: associations with bicycle characteristics. J. Urol. 172, 637e641. Dunstan, D.W., Kingwell, B.A., Larsen, R., Healy, G.N., Cerin, E., Hamilton, M.T., Shaw, J.E., Bertovic, D.A., Zimmet, P.Z., Salmon, J., 2012. Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care 35, 976e983. Felson, D.T., Anderson, J.J., Naimark, A., Walker, A.M., Meenan, R.F., 1988. Obesity and knee osteoarthritis. The Framingham study. Ann. Intern. Med. 109, 18e24. Finkelstein, E.A., Trogdon, J.G., Cohen, J.W., Dietz, W., 2009. Annual medical spending attributable to obesity: payer-and service-specific estimates. Health Aff. (Millwood) 28, 822e831. Ford, E.S., Kohl, H.W., Mokdad, A.H., Ajani, U.A., 2012. Sedentary behavior, physical activity, and the metabolic syndrome among US adults. Obes. Res. 13, 608e 614. Funk, R.E., Taylor, M.L., Creekmur, C.C., Ohlinger, C.M., Cox, R.H., Berg, W.P., 2012. Effect of walking speed on typing performance using an active workstation. Percept. Mot. Ski. 115, 309e318. Gardner, A.S., Stephens, S., Martin, D.T., Lawton, E., Lee, H., Jenkins, D., 2004. Accuracy of SRM and power tap power monitoring systems for bicycling. Med. Sci. Sports Exerc. 36, 1252e1258. Haskell, W.L., Lee, I., Pate, R.R., Powell, K.E., Blair, S.N., Franklin, B.A., Macera, C.A., Heath, G.W., Thompson, P.D., Bauman, A., 2007. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med. Sci. Sports Exerc. 39, 1423e1434. Health and Human Services, 2008. 2008 Physical Activity Guidelines for Americans: Be Active, Healthy, and Happy! US Government Printing Office. John, D., Bassett, D., Thompson, D., Fairbrother, J., Baldwin, D., 2009. Effect of using a treadmill workstation on performance of simulated office work tasks. J. Phys. Act. Health 6, 617e624. John, D., Thompson, D.L., Raynor, H., Bielak, K., Rider, B., Bassett, D.R., 2011. Treadmill workstations: a worksite physical activity intervention in overweight and obese office workers. J. Phys. Act. Health 8, 1034e1043. Kate Bundorf, M., 2002. Employee demand for health insurance and employer health plan choices. J. Health Econ. 21, 65e88. Koepp, G.A., Manohar, C.U., McCrady-Spitzer, S.K., Ben-Ner, A., Hamann, D.J., Runge, C.F., Levine, J.A., 2013. Treadmill desks: a 1-year prospective trial. Obesity 21, 705e711. Levine, J.A., Miller, J.M., 2007. The energy expenditure of using a “walk-and-work” desk for office workers with obesity. Br. J. Sports Med. 41, 558e561. Martin, J.C., Milliken, D.L., Cobb, J.E., McFadden, K.L., Coggan, A.R., 1998. Validation of a mathematical model for road cycling power. J. Appl. Biomech. 14, 276e 291. Mayo Clinic, 2011. Barriers to Fitness: Overcoming Common Challenges, Fitness. Mayo Foundation for Medical Education and Research, Rochester. Available from: http://www.mayoclinic.com/health/fitness/SM00085_D (accessed 08.12.13.). McAlpine, D.A., Manohar, C.U., McCrady, S.K., Hensrud, D., Levine, J.A., 2007. An office-place stepping device to promote workplace physical activity. Br. J. Sports Med. 41, 903e907. Morris, J., Heady, J., Raffle, P., Roberts, C., Parks, J., 1953. Coronary heart-disease and physical activity of work. Lancet 262, 1111e1120. Mummery, W.K., Schofield, G.M., Steele, R., Eakin, E.G., Brown, W.J., 2005. Occupational sitting time and overweight and obesity in Australian workers. Am. J. Prev. Med. 29, 91e97. Mutrie, N., 2001. The relationship between physical activity and clinically defined depression. In: Biddle, S.J.H., Fox, K.R., Boutcher, S.H. (Eds.), Physical Activity and Psychological Well-being. Routledge, London, pp. 46e62. Neptune, R.R., Hull, M.L., 1995. Accuracy assessment of methods for determining hip movement in seated cycling. J. Biomech. 28, 423e437. Schwarzer, U., Sommer, F., Klotz, T., Cremer, C., Engelmann, U., 2002. Cycling and penile oxygen pressure: the type of saddle matters. Eur. Urol. 41, 139e143. Straker, L., Levine, J., Campbell, A., 2009. The effects of walking and cycling computer workstations on keyboard and mouse performance. Hum. Factors 51, 831e844. Thompson, W.G., Levine, J.A., 2011. Productivity of transcriptionists using a treadmill desk. Work 40, 473e477. Weller, I., Corey, P., 1998. The impact of excluding non-leisure energy expenditure on the relation between physical activity and mortality in women. Epidemiology 9, 632e635. Weuve, J., Kang, J.H., Manson, J.E., Breteler, M.M., Ware, J.H., Grodstein, F., 2004. Physical activity, including walking, and cognitive function in older women. JAMA 292, 1454e1461. Wilmot, E., Edwardson, C., Achana, F., Davies, M., Gorely, T., Gray, L., Khunti, K., Yates, T., Biddle, S., 2012. Sedentary time in adults and the association with diabetes, cardiovascular disease and death: systematic review and meta-analysis. Diabetologia 55, 2895e2905. Wooles, A.L., Robinson, A.J., Keen, P.S., 2005. A static method for obtaining a calibration factor for SRM bicycle power cranks. Sports Eng. 8, 137e144. Zuntz, N., 1901. sber die Bedeutung der verschiedene Nahrstoffe als Erzeuber der Muskelkraft. Pflugers Arch. 83, 557e571.
Contents lists available at ScienceDirect
Applied Ergonomics journal homepage: www.elsevier.com/locate/apergo
A cycling workstation to facilitate physical activity in office settings Steven J. Elmer a, b, c, *, James C. Martin d a
Department of Exercise Science and STEM Education, University of Maine, 5740 Lengyel Hall, Orono, ME 04469, USA Department of Mechanical Engineering, University of Maine, Orono, ME, USA c Eastern Maine Medical Center, Bangor, ME, USA d Department of Exercise and Sport Science, University of Utah, Salt Lake City, UT, USA b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 8 December 2013 Accepted 6 March 2014
Facilitating physical activity during the workday may help desk-bound workers reduce risks associated with sedentary behavior. We 1) evaluated the efficacy of a cycling workstation to increase energy expenditure while performing a typing task and 2) fabricated a power measurement system to determine the accuracy and reliability of an exercise cycle. Ten individuals performed 10 min trials of sitting while typing (SITtype) and pedaling while typing (PEDtype). Expired gases were recorded and typing performance was assessed. Metabolic cost during PEDtype was w2.5 greater compared to SITtype (255 14 vs. 100 11 kcal h1, P < 0.01). Typing time and number of typing errors did not differ between PEDtype and SITtype (7.7 1.5 vs. 7.6 1.6 min, P ¼ 0.51, 3.3 4.6 vs. 3.8 2.7 errors, P ¼ 0.80). The exercise cycle overestimated power by 14e138% compared to actual power but actual power was reliable (r ¼ 0.998, P < 0.01). A cycling workstation can facilitate physical activity without compromising typing performance. The exercise cycle’s inaccuracy could be misleading to users. Ó 2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.
Keywords: Workplace activity Energy expenditure Ergometer calibration
1. Introduction Thirty-five percent of adults in the United States are currently obese and at increased risk of developing several metabolic, cardiovascular, and psychological disorders (Haskell et al., 2007; Mutrie, 2001). Health care costs attributed to obesity are approximated to be $147 billion per year in the United States and those costs are predicted to increase to $200 billion per year over the next two decades (Finkelstein et al., 2009). Obesity results from a longterm excess of energy consumed vs. energy expended, a positive energy balance (Church et al., 2011). While overeating certainly contributes to a positive energy balance, the epidemic of obesity and related metabolic disorders is also driven by reductions in energy expended (Ford et al., 2012). With increased use of computers, office workers may remain seated and sedentary (Barnes et al., 2012) during which energy expenditure is minimal, making many offices obesogenic environments (Mummery et al., 2005). Sedentary behavior not only contributes to positive energy balance but is also an independent risk factor for diabetes, cardiovascular
* Corresponding author. Department of Exercise Science and STEM Education, University of Maine, 5740 Lengyel Hall, Orono, ME 04469, USA. Tel.: þ1 207 581 3314; fax: þ1 207 581 1206. E-mail address: [email protected] (S.J. Elmer). http://dx.doi.org/10.1016/j.apergo.2014.03.001 0003-6870/Ó 2014 Elsevier Ltd and The Ergonomics Society. All rights reserved.
disease, and all-cause mortality (Wilmot et al., 2012). Therefore, increasing physical activity and/or decreasing sedentary time are paramount to reducing obesity and associated disorders. However, traditional strategies to increase physical activity are compromised by insufficient time for exercise, perception that exercise is boring, concern about appearance during exercise, and fatigue after the workday (Mayo Clinic, 2011). These barriers could be reduced by facilitating physical activity during working hours. Treadmill workstations have been promoted as means to increase physical activity during working hours. John et al. (2011) reported that office workers, provided with a treadmill workstation for nine months, increased their walking time by 38e 75 min day1, compared to baseline, and averaged a walking speed of 1.5 mph. This duration and speed suggests an energy expenditure of approximately 100e200 kcal day1, above baseline (Levine and Miller, 2007). Compared to baseline, the participants in this study exhibited significant decreases in waist and hip circumference as well as improvements in blood lipids and cholesterol. Similarly, Koepp et al. (2013) recently reported that office workers who were provided with a treadmill workstation for one year increased physical activity and decreased weight, while maintaining productivity (Koepp et al., 2013). This finding of unchanged productivity is important because other authors have reported that use of a treadmill workstation may compromise computer use accuracy (Funk et al., 2012; John et al., 2009) and/or work productivity
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
(Thompson and Levine, 2011). A potential limitation of treadmill workstation use is that individuals who are overweight and obese are at increased risk for knee osteoarthritis (Felson et al., 1988) and thus may have limited ability to walk. As a non-weight bearing activity, cycling could serve as an alternative modality for facilitating increased physical activity in desk-bound office workers. Indeed, Straker et al. (2009) evaluated the effects of a cycling workstation (computer desk and an upright exercise cycle) on physiological responses and computer operation in office workers. Specifically, these authors reported that heart rate values increased by w25% when pedaling and typing compared to normal sitting and typing. Additionally, these authors reported that typing performance while cycling was slightly compromised but was better than typing performance during treadmill walking. To the best of our knowledge, no previous authors have reported the actual rate of energy expenditure that individuals might self-select for pedaling during an office task such as typing. It is also important to note that the previous cycling workstation design used by Straker et al. (2009) consisted of an upright exercise cycle with a relatively small seat. This design may present some ergonomic challenges, especially when using it for extended durations, as several participants in that study (9 out of 30) reported some form of hip or gluteal discomfort related to the exercise cycle seat. Further, such bicycle seats are known to reduce penile oxygen pressure, reflecting perineal compression (Schwarzer et al., 2002) and prolonged sitting on such seats can lead to perineal numbness and erectile dysfunction (Dettori et al., 2004). Thus, a cycling workstation with a recumbent seat position that is more similar to a standard office chair may facilitate prolonged activity without risks associated with upright cycling. Our primary purposes for conducting this investigation were to 1) determine the metabolic cost associated with self-selected pedaling intensity while performing a standardized typing task and 2) assess the influence of pedaling on typing performance (typing time and number of typing errors). These responses were compared to those associated with normal sitting while typing. We hypothesized that compared to sitting while typing, pedaling while typing would increase metabolic cost without compromising typing performance. Additionally, because the accuracy and reliability of commercial exercise cycles to quantify power are not documented, our secondary purposes were to 1) fabricate a power measurement system to determine the actual power associated with pedaling the exercise cycle across a range of resistance levels and pedaling rates before and after the experimental trials and 2) compare the actual power, as determined by the power measurement system, to the power displayed by the exercise cycle console. This study could serve as an important first step to utilizing a cycling workstation to reduce sedentary time and increase physical activity in office settings. 2. Methods 2.1. Participants Ten healthy males volunteered to participate in this study (age: 32 8 yr, body mass: 72 8 kg, height: 1.78 0.07 m, body mass index: 23 2 kg m2). Participants were college students and faculty who spent considerable time sitting at a desk each day (6.3 2.4 h day1, self-reported). Participants were also recreationally active in a variety of sports and familiar with cycling exercise (both upright and recumbent cycling positions). Experimental procedures used in this investigation were reviewed by the University of Utah Institutional Review Board. The protocol and procedures were explained verbally and all participants provided written informed consent prior to testing.
1241
2.2. Experimental protocol Participants reported to the laboratory dressed in typical school or office attire and were instructed to refrain from eating for at least 3 h prior to testing. During this visit resting baseline data were first recorded. Subsequently, participants performed two experimental trials (10 min) while physiological responses were recorded (described below): 1) Sitting while typing 2) Pedaling while typing For the sitting while typing condition, participants were positioned on the cycling workstation (but did not pedal) and performed a standardized typing task (described below). For the pedaling condition, participants were positioned on the cycling workstation and pedaled while they performed the same standardized typing task. The two experimental trials were presented in a random order. At least 5 min of recovery was provided between each trial to allow physiological responses to return to baseline levels. 2.3. Typing task To quantify typing performance, we measured the time required to transcribe the Gettysburg Address, a famous speech delivered by United States President Abraham Lincoln in 1863, which participants practiced typing twice prior to the experimental protocol. For this task, a copy of the Gettysburg Address was placed adjacent to the computer screen and participants were instructed to transcribe the Gettysburg Address at a comfortable typing speed, a speed at which they typically performed school and/or office-related typing. Participants were also instructed to correct any typing errors that might have occurred as they typed so that the overall quality of the document was equivalent to that of school and/or office work. Because all participants regularly used a word processing autocorrect spelling/grammar feature this setting was enabled during the standardized typing task. The total time required to transcribe the Gettysburg Address and correct any potential errors was recorded. In addition, an investigator carefully reviewed the document and recorded the actual number of typing errors. If participants completed the typing task before the 10 min period was over then they were simply asked to start the typing task again but the second round was not timed or completed. 2.4. Cycling workstation A cycling workstation was constructed using a commercially available recumbent exercise cycle (Model R3i, LifeSpan Fitness, Salt Lake City, UT, USA), a set of adjustable height table legs (Model UpLift 700, The Human Solution, Austin, TX, USA), and a custom made keyboard tray with integrated arm support (Fig. 1). The seat pan of the exercise cycle was 44 cm wide, 26 cm deep, and positioned at an inclination of w1 (positive to horizontal). The back support was 40 cm wide, 52 cm tall, and positioned at an inclination of w15 (backward of vertical). The standard exercise cycle cranks (170 mm) were replaced with shorter cranks (114 mm, Unicycle.com) to facilitate usable ergonomics. Participants selected a seat position that allowed a slight bend at the knee when the leg was in its most extended position. The desk height was then adjusted to be as low as possible while still providing clearance for the knees during pedaling. The cutout in the keyboard tray was approximately 24 cm deep and 41 cm wide (Fig. 1). This depth allowed the tray to extend to the anterior-posterior midline of the torso and thus provide support for the elbows when the upper arm
1242
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
Fig. 1. Overview of the cycling workstation which consisted of a 1) commercially available recumbent exercise cycle, 2) set of adjustable height table legs, and 3) a custom made keyboard tray with integrated arm support that extended to the anterior-posterior midline of the torso. Participants pedaled at a resistance level and pedaling rate that they could comfortably maintain while performing a standardized typing task.
was in a near vertical orientation. Forearm/elbow support provided by the custom keyboard tray was an integral aspect of the ergonomic solution intended to stabilize the arms and reduce the requirement for muscular elevation (see discussion for further explanation). After these adjustments were made, participants were given time to become familiar with pedaling the cycling workstation. That is, participants were instructed to select a resistance level that they could comfortably maintain for prolonged periods while performing computer tasks. During the experimental pedaling condition, the preferred resistance level was set on the exercise cycle console and participants pedaled at that resistance level for 10 min while they were free to modulate their pedaling rate. 2.5. Power measurement system The commercial exercise cycle used in this study was equipped with a console that displayed, and recorded resistance level, pedaling rate, and power. The accuracy and reliability of power determined by the console of commercial exercise cycles in general are not well known. Therefore, prior to experimental data collection we determined the actual power required to drive the cranks at a range of pedaling rates and resistance settings. To facilitate power measurement, we constructed a system consisting of an electric motor, variable frequency drive, and a power meter (Fig. 2). Specifically, a 1.1 kW, 100 rpm, three-phase, electric gear motor (Baldor Electric Co. Model Syncromotor, Ft. Smith, AR, USA) was used to drive the power meter (Schoberer Rad Messtechnik, SRM, Jülich, Germany) via a chain-drive system. The power meter was connected in series to a driveshaft which attached to the exercise cycle crank axle. Therefore, the electric motor drove the power meter which ultimately “pedaled” the exercise cycle. Motor speed and thus pedaling rate were controlled by a variable frequency drive (TECO/Westinghouse, Model FM50-101-C, Round Rock, TX, USA). Thus, power required to pedal the exercise cycle was quantified using the power meter, a system that serves as an accurate method
Fig. 2. Schematic of the power measurement system used to determine the power required to drive the exercise cycle at a range of pedaling rates and resistance settings. An electric motor was connected to a power meter (light gray circle) via a chain drive system (dashed line). The power meter was connected in series to a driveshaft that attached to the exercise cycle. Motor speed and pedaling rate were controlled using a variable frequency drive (dark gray box). Thus, the electric motor drove the power meter which ultimately “pedaled” the exercise cycle.
to quantify cycling power (Abbiss et al., 2009; Gardner et al., 2004; Martin et al., 1998). Prior to this power evaluation, the power meter itself was calibrated using static calibration procedures (Wooles et al., 2005). With this power measurement system we evaluated the power required to pedal the exercise cycle at the lowest seven resistance levels at 40, 50, 60, 70, and 80 rpm. From these data, powerpedaling rate relationships were determined for each resistance level using linear regression. During the experimental pedaling condition, pedaling rate was measured using the exercise cycle console and actual power produced by the participant was calculated using the power-pedaling rate regression equations. We also determined the relationship between the actual power determined by our power measurement system (i.e., gold standard) and the power determined by the exercise cycle console, as this could be important for other researchers using commercial exercise cycles. Finally, this procedure was again performed after the completion of experimental pedaling trial in order to assess the test-retest reliability of the exercise cycle. 2.6. Metabolic, cardiorespiratory, and perceptual measures Gas exchange data were measured during each experimental condition using open circuit spirometry (True Max 2400, Parvo Medics, Sandy, UT, USA). The metabolic system was calibrated with a 3 L calibration syringe (Hans Rudolph, Kansas City, MO, USA) and medical gases of known concentrations (16.00% O2, 4.02% CO2, _ Þ, carbon dioxide probalanced N2). Oxygen consumption ðVO 2 _ _ duction ðVCO 2 Þ, minute ventilation ðVE Þ, respiratory exchange ratio (RER), and metabolic equivalents (METS) data were recorded every _ 15 s and averaged over min 5e10 of each condition. Absolute VO 2 1 and RER data were used to calculate metabolic cost (kcal h ) using the regression equation reported by Zuntz (1901) which is based upon the thermal equivalent of O2 for nonprotein respiratory equivalent. Heart rate was measured using a Timex monitor (Timex Ironman, Timex Group USA, Middlebury, CT, USA) and averaged
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
over min 5e10 of each trial. Finally, overall rating of perceived exertion (RPE) was assessed during min 4e5 and 9e10 using a Borg 6e20 scale (Borg, 1970).
Power-pedaling rate relationships determined from the power measurement system were highly linear for each resistance level evaluated (Fig. 5). Relationships between the actual power and power displayed by the exercise cycle console indicated that the exercise cycle console grossly overestimated power (Fig. 5 and Table 1). Specifically, mean power differences at the lowest resistance setting ranged from 3 to 25 W (14e47%) across the pedaling rates evaluated (40, 50, 60, 70, 80 rpm) and mean differences at the highest resistance setting ranged from 42 to 145 W (108e127%) across pedaling rates (Table 1). Power displayed by the exercise cycle console for each resistance setting and pedaling rate was identical before and after the experimental pedaling trial,
Power (W)
40 20
2
4
6
8
10
1.2 1.0 -1
*
0.8 0.6 0.4 0.2 0.0 0
2
4
6
8
10
30 25
VE (L∙min )
.
-1
3.2. Power measurements
60
0
*
20 15 10 5 0 0
2
4
6
8
10
100 -1
Heart Rate (b∙min )
All ten participants completed both the pedaling while typing and sitting while typing experimental trials. However, one participant accidently missed transcribing a phrase during the typing trial. Thus, values for typing time and number of typing errors were reported in only nine participants. Participants selected to pedal at a resistance level of 4 1 on the exercise cycle. Mean pedaling rate and power produced during the pedaling while typing trial were 51 14 rpm and 38 14 W (actual power based on power-pedaling rate relationships described below), respectively. Oxygen consumption, V_ E , and heart rate were substantially greater during pedaling while typing compared to sitting while typing (all P < 0.01, Fig. 3). Similarly, metabolic cost and METs were w2.5 times greater during pedaling while typing compared to sitting while typing (both P < 0.01, Fig. 4). Despite these differences, the time required to complete the standardized typing task and number of typing errors during pedaling while typing did not differ from that during sitting while typing (7.7 1.5 vs. 7.6 1.6 min, P ¼ 0.51; 3.3 4.6 vs. 3.8 2.7 errors, P ¼ 0.80, respectively). Compared to sitting while typing, RPE was slightly higher during pedaling while typing at 5 min (6.9 0.8 vs. 6.1 0.3 scale units, P < 0.01) and 10 min (6.9 0.7 vs. 6.1 0.3 scale units, P < 0.01).
Pedaling
0
3. Results 3.1. Physiological responses and typing performance
Sitting
.
Separate one-way repeated measures analysis of variance _ , V_ E , (ANOVA) procedures were used to assess differences in VO 2 and heart rate values during the 10 min experimental trials (sitting while typing vs. pedaling while typing). Additionally, paired student’s t-tests were used to compare differences in metabolic cost, METS, and RPE between sitting while typing and pedaling while typing trials. Separate paired student’s t-test were used to compare differences in the time required to complete the standardized typing task and number of typing errors between sitting while typing and pedaling while typing trials. Power-pedaling rate relationships determined for each resistance level with the power measurement system are reported for descriptive purposes. To assess the accuracy of the exercise cycle console to measure power, mean differences of actual power data and the power displayed on the exercise cycle console were determined. Further, the test-retest reliability of the exercise cycle was evaluated using a Pearson productemoment correlation. Data are presented as mean standard deviation and alpha was set to 0.05.
VO2 (L∙min )
2.7. Data analysis
80
1243
90
*
80 70 60 50 0
2
4
6
8
10
Time (min) Fig. 3. Time course of alterations for selected mechanical, metabolic, and cardiorespiratory variables during sitting while typing and pedaling while typing. Values are reported as Mean SD. *P < 0.01 vs. sitting while typing (main effect of exercise condition).
suggesting that firmware in the console bases the power calculation on measured pedaling rate and indicated resistance level. Further, power values determined by the power measurement system before and after the experimental pedaling trial were highly reliable as demonstrated by a Pearson productemoment correlation coefficient of 0.998 (P < 0.01). This suggests that the exercise cycle resistance settings, which control the position of magnets relative to the flywheel, were highly repeatable. Note that, to improve the clarity of Fig. 5 relationships were illustrated for only those resistance levels utilized by the participants in this investigation (resistance levels 2, 3, 4, and 5).
1244
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
A
A
150
*
300
125
250
Actual Power (W)
-1
Metabolic Cost (kcal∙h )
350
200 150 100 50
100 Level 2 Level 3 Level 4 Level 5
75
50
0
Sitting 4.0
B
Pedaling
25
*
0
3.5
30
40
2.5
METS
50
60
70
80
90
Pedaling Rate (rpm)
3.0
B
250
2.0 1.5
0.5 0.0
Sitting
Pedaling
Fig. 4. Comparison of metabolic cost (A) and metabolic equivalents (METS) (B) between sitting while typing and pedaling while typing. Dashed line represents baseline sitting values (no typing). Values are reported as Mean SD. *P < 0.01 vs. sitting while typing.
4. Discussion An important driving force in the obesity epidemic is the decrease in physical activity during working hours (Church et al., 2011). Consequently, facilitating physical activity in the workplace is essential in combating obesity and improving worker health. Employers have a vested interest in employee heath because they ultimately bear the costs of most insurance programs (Kate Bundorf, 2002). In this study, we evaluated the use of a cycling workstation to facilitate physical activity and its effect on typing performance. With our modified cycling workstation participants chose to pedal at a power output of 38 W (actual power) while performing a standardized typing task, an intensity they rated as “very very light” exertion. These results support previous reports documenting the use of a cycling workstation (Straker et al., 2009) in which participants pedaled at 30 W while performing typing task and rated their intensity between “fairly light” and “moderate”. Further, metabolic rates during pedaling at 38 W in the current study (240 kcal h1) compare favorably to those estimated during treadmill walking at 1.5 mph (w162 kcal h1) (John et al., 2011; Levine and Miller, 2007). Importantly, pedaling at this intensity was not uncomfortable for participants and did not significantly compromise the time required to type The Gettysburg Address or number of typing errors. Taken together, these results indicate that a cycling workstation may provide an additional viable option for increasing physical activity in desk-bound workers without compromising typing performance.
Display Power (W)
200
1.0
Level 2 Level 3 Level 4 Level 5
150
100
50
0 0
50
100
150
200
250
Actual Power (W) Fig. 5. Power-pedaling rate relationships for the exercise cycle. Power required to “pedal” the exercise cycle for several resistance levels and pedaling rates as measured by the calibrated SRM power meter (A). Power-pedaling rate relationships were essentially linear which implies a constant resistant torque provided by the exercise cycle internal braking system. Comparison of actual power required (as measured by the calibrated SRM power meter) with power displayed on the exercise cycle control console (B). The exercise cycle console overestimated power in all conditions and the error ranged from 14 to 138%. For the settings and pedaling rates used by participants in this study while typing, power displayed by the console was approximately twice the actual power.
4.1. Potential outcomes Pedaling while typing on a cycling workstation elevated metabolic cost by 155 kcal h1 compared to typing alone. This additional energy expenditure could allow desk-bound office workers to reduce sedentary time which can significantly improve glucose and insulin tolerance (Dunstan et al., 2012) and reduce the risk of diabetes (Wilmot et al., 2012). Prolonged use of the cycling workstation at this intensity could also dramatically influence energy balance and replace workplace physical activity that has been lost over the last 50 years (Church et al., 2011). For example, pedaling 10 min every hour throughout an 8 h workday (80 min day1) could
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
1245
Table 1 Differences in actual power measured from the power measurement system and power displayed on the exercise cycle console. Note that, positive values indicate that the exercise cycle console power was greater than the actual power. Pedaling rate
40 50 60 70 80
Resistance level 1
2
3W (14%) 9W (31%) 14W (38%) 21W (47%) 25W (45%)
12W 20W 32W 39W 51W
3 (54%) (65%) (86%) (76%) (84%)
22W 31W 45W 59W 78W
(100%) (88%) (102%) (105%) (118%)
result in 200 kcal day1 or 1000 kcal wk1 of additional energy expenditure compared to typing alone. Pedaling intermittently for two to 4 h each day at this intensity would increase energy expenditure by 300e600 kcal day1 or 1500e3000 kcal wk1 compared to typing alone. One thousand kcal wk1 is significant because it meets the Center for Disease Control and Prevention (CDC) criteria for moderate physical activity (Health and Human Services, 2008). Assuming one pound of weight loss could result from each 3500 kcal of negative energy balance, these values of 200e600 kcal day1 could facilitate weight loss of approximately one pound every 6 to 17 workdays, or 15 to 45 pounds per year, provided energy intake and other physical activity remained constant. In addition to weight loss, physical activity per se is known to reduce the risk of cardiovascular disease (Morris et al., 1953; Weller and Corey, 1998). Therefore, use of a cycling workstation could dramatically reduce weight and improve health of desk-bound workers. On a similar note, facilitating physical activity to maintain and prevent weight gain is also important (Church et al., 2011) and staying active can improve cognitive function (Weuve et al., 2004). An important next step would be to determine the selfselected intensity and actual daily use of cycling workstations by overweight and obese individuals. 4.2. Typing performance Despite employers’ vested interest in employee health, some may be reluctant to facilitate physical activity if they believe it will compromise productivity. Previous authors have reported mixed results for computer operation (typing time, speed, word count, and/or accuracy) with cycling workstations (Straker et al., 2009), treadmill workstations (Funk et al., 2012; Koepp et al., 2013; Straker et al., 2009; Thompson and Levine, 2011), and other novel workstation modalities (Beers et al., 2008) when compared to normal sitting. Our finding of similar typing time agrees with some previous reports for physical activity during computer operation (Funk et al., 2012; Koepp et al., 2013) but contrast with others which indicated small but significant decrements in typing speed and accuracy (Funk et al., 2012; John et al., 2009; Straker et al., 2009; Thompson and Levine, 2011). These varied results likely reflect the degree to which various users could compensate for additional task complexity. That is, while walking the entire body is moving and thus complex control of the arms and hands is required to keep them stationary with respect to the keyboard. Cycling in an upright position on a narrow saddle may induce upper body movement via hip movement (Neptune and Hull, 1995) thereby also increasing task complexity. With our cycling workstation that included a recumbent seat system, upper body movement induced by pedaling was likely reduced compared to walking or upright cycling, and the arms were further stabilized by the supportive keyboard tray. Other novel workstation modalities for increasing physical activity while reducing upper body movement (e.g., seated stepping in place (McAlpine et al., 2007)), have not reported effects on computer operation and thus their feasibility in a workplace
4
5
6
7
27W (104%) 41W (105%) 59W (118%) 81W (135%) 101W (135%)
34W (117%) 48W (107%) 69W (119%) 94W (133%) 119W (138%)
36W (100%) 54W (102%) 80W (121%) 105W (125%) 135W (135%)
42W (108%) 63W (109%) 86W (110%) 119W (127%) 145W (121%)
setting are unknown. Finally, it is important to note that these results highlighting the efficacy of a cycling workstation are limited to transcribe typing which is only one task that might be performed during computer operation. 4.3. Ergonomic solutions Establishing usable ergonomics with a recumbent cycling workstation posed several challenges. First, when pedaling a recumbent exercise cycle the knees will be higher, relative to the shoulders, than they would when seated in a typical office chair. This requires that the tabletop be raised high enough to allow pedaling without the knees hitting the table. Because the tabletop and keyboard must be relatively higher than normal, the user might be forced to use muscular effort to hold up the arms and shoulders. This could result in fatigue and/or discomfort of those muscles. This issue was reduced, but not eliminated, by using shorter-than-normal cycle cranks to reduce the required desktop height. Further, the pedaling action induces a slight swaying motion of the torso which could increase task complexity for typing, resulting in decreased office performance. Our solution to these issues was to use a custom made keyboard tray with extended surface to support and stabilize the arms. This level of customization was essential as most computer desks have keyboard designs that do not extend far enough to adequately support the arms and minimize requirement for muscular elevation. It is also important to point out that the seat of the cycling workstation used in this investigation was reasonably similar to that of a standard office chair. Thus, workers could remain seated while not pedaling and resume pedaling at will. Moreover, none of the participants reported any noticeable seat-related discomfort which was a concern with previous cycling workstation designs (Straker et al., 2009). However, the extent to which long term use of the cycling workstation elicits any discomfort and/or fatigue remains to be investigated. 4.4. Power measurements We used a power meter-based system to determine powerpedaling rate relationships for the exercise cycle resistance settings used in this study. SRM power meters are known to be accurate and reliable (Abbiss et al., 2009; Gardner et al., 2004; Martin et al., 1998) and thus we are confident in our measures of actual power. Results demonstrated that power-pedaling rate relationships were highly linear for each resistance setting indicating that the current brake system provides nearly constant torque for any specific resistance (position of the magnets relative to the flywheel). Further, the agreement of our pre- and post-test power values indicates that the resistance settings of the exercise cycle were reliable. While the exercise cycle was reliable, the large differences in actual power and the power displayed on the exercise cycle console were unexpected. Specifically, the console power was generally over twice as great as the actual power except for the very lowest resistant settings. This could be important to users because
1246
S.J. Elmer, J.C. Martin / Applied Ergonomics 45 (2014) 1240e1246
the console also presents a value for energy expenditure which is based on the summation of power over time. This level of inaccuracy could be very misleading to a user who might carefully track energy expenditure while pedaling the exercise cycle. For example an individual might believe that he/she has burned a given number of kcals when in fact the total is less than half of that. Thus, users should be cautious in their expectations of accuracy for this specific commercial exercise cycle. While we only have actual power data on the exercise cycle used in this investigation we intend to determine accuracy of a variety of exercise cycles found in commercial exercise facilities. 4.5. Summary These data demonstrate the efficacy of a cycling workstation to facilitate physical activity without compromising typing performance. Use of a cycling workstation for 80 min to 4 h per day could result in significant weight loss over time and have additional health benefits associated with reduced sedentary time (Barnes et al., 2012). With its office-chair-like seat, a cycling workstation may provide a useful alternative to treadmill desks for those who would prefer non-weight-bearing physical activity. The specific exercise cycle we used exhibited good reliability as assessed by a power meter-based measurement system but the power displayed by the exercise cycle console and, thus all the calculations based on power, were inaccurate and tended to be around twice the actual values. Together, these results may have implications for prevention and/or treatment of overweight and obesity and associated health risks in desk-bound office works as well individuals who use commercial exercise cycles for recreational purposes. Funding source No outside funding was received for this study. Conflict of interest Steven Elmer declares no conflict of interest. James Martin has disclosed a potential conflict of interest. He is the inventor of intellectual property related to a cycling workstation for which the University of Utah has filed a provisional patent, and he holds equity in a company that has licensed the rights to this intellectual property from the University of Utah. Acknowledgments We thank the participants for their enthusiastic efforts during the experimental trials and Alex Seegmiller for assistance with the data collection and power measurement trials. References Abbiss, C.R., Quod, M.J., Levin, G., Martin, D.T., Laursen, P.B., 2009. Accuracy of the Velotron ergometer and SRM power meter. Int. J. Sports Med. 30, 107e112. Barnes, J., Behrens, T.K., Benden, M.E., Biddle, S., Bond, D., Brassard, P., Brown, H., Carr, L., Chaput, J.-P., Christian, H., 2012. Letter to the Editor: standardized use of the terms “sedentary” and “sedentary behaviours”. Appl. Physiol. Nutr. Metab. 37, 540e542. Beers, E.A., Roemmich, J.N., Epstein, L.H., Horvath, P.J., 2008. Increasing passive energy expenditure during clerical work. Eur. J. Appl. Physiol. 103, 353e360. Borg, G., 1970. Perceived exertion as an indicator of somatic stress. Scand. J. Rehabil. Med., 92e98. Church, T.S., Thomas, D.M., Tudor-Locke, C., Katzmarzyk, P.T., Earnest, C.P., Rodarte, R.Q., Martin, C.K., Blair, S.N., Bouchard, C., 2011. Trends over 5 decades in U.S. occupation-related physical activity and their associations with obesity. PloS One 6, e19657.
Dettori, J.R., Koepsell, T.D., Cummings, P., Corman, J.M., 2004. Erectile dysfunction after a long-distance cycling event: associations with bicycle characteristics. J. Urol. 172, 637e641. Dunstan, D.W., Kingwell, B.A., Larsen, R., Healy, G.N., Cerin, E., Hamilton, M.T., Shaw, J.E., Bertovic, D.A., Zimmet, P.Z., Salmon, J., 2012. Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care 35, 976e983. Felson, D.T., Anderson, J.J., Naimark, A., Walker, A.M., Meenan, R.F., 1988. Obesity and knee osteoarthritis. The Framingham study. Ann. Intern. Med. 109, 18e24. Finkelstein, E.A., Trogdon, J.G., Cohen, J.W., Dietz, W., 2009. Annual medical spending attributable to obesity: payer-and service-specific estimates. Health Aff. (Millwood) 28, 822e831. Ford, E.S., Kohl, H.W., Mokdad, A.H., Ajani, U.A., 2012. Sedentary behavior, physical activity, and the metabolic syndrome among US adults. Obes. Res. 13, 608e 614. Funk, R.E., Taylor, M.L., Creekmur, C.C., Ohlinger, C.M., Cox, R.H., Berg, W.P., 2012. Effect of walking speed on typing performance using an active workstation. Percept. Mot. Ski. 115, 309e318. Gardner, A.S., Stephens, S., Martin, D.T., Lawton, E., Lee, H., Jenkins, D., 2004. Accuracy of SRM and power tap power monitoring systems for bicycling. Med. Sci. Sports Exerc. 36, 1252e1258. Haskell, W.L., Lee, I., Pate, R.R., Powell, K.E., Blair, S.N., Franklin, B.A., Macera, C.A., Heath, G.W., Thompson, P.D., Bauman, A., 2007. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med. Sci. Sports Exerc. 39, 1423e1434. Health and Human Services, 2008. 2008 Physical Activity Guidelines for Americans: Be Active, Healthy, and Happy! US Government Printing Office. John, D., Bassett, D., Thompson, D., Fairbrother, J., Baldwin, D., 2009. Effect of using a treadmill workstation on performance of simulated office work tasks. J. Phys. Act. Health 6, 617e624. John, D., Thompson, D.L., Raynor, H., Bielak, K., Rider, B., Bassett, D.R., 2011. Treadmill workstations: a worksite physical activity intervention in overweight and obese office workers. J. Phys. Act. Health 8, 1034e1043. Kate Bundorf, M., 2002. Employee demand for health insurance and employer health plan choices. J. Health Econ. 21, 65e88. Koepp, G.A., Manohar, C.U., McCrady-Spitzer, S.K., Ben-Ner, A., Hamann, D.J., Runge, C.F., Levine, J.A., 2013. Treadmill desks: a 1-year prospective trial. Obesity 21, 705e711. Levine, J.A., Miller, J.M., 2007. The energy expenditure of using a “walk-and-work” desk for office workers with obesity. Br. J. Sports Med. 41, 558e561. Martin, J.C., Milliken, D.L., Cobb, J.E., McFadden, K.L., Coggan, A.R., 1998. Validation of a mathematical model for road cycling power. J. Appl. Biomech. 14, 276e 291. Mayo Clinic, 2011. Barriers to Fitness: Overcoming Common Challenges, Fitness. Mayo Foundation for Medical Education and Research, Rochester. Available from: http://www.mayoclinic.com/health/fitness/SM00085_D (accessed 08.12.13.). McAlpine, D.A., Manohar, C.U., McCrady, S.K., Hensrud, D., Levine, J.A., 2007. An office-place stepping device to promote workplace physical activity. Br. J. Sports Med. 41, 903e907. Morris, J., Heady, J., Raffle, P., Roberts, C., Parks, J., 1953. Coronary heart-disease and physical activity of work. Lancet 262, 1111e1120. Mummery, W.K., Schofield, G.M., Steele, R., Eakin, E.G., Brown, W.J., 2005. Occupational sitting time and overweight and obesity in Australian workers. Am. J. Prev. Med. 29, 91e97. Mutrie, N., 2001. The relationship between physical activity and clinically defined depression. In: Biddle, S.J.H., Fox, K.R., Boutcher, S.H. (Eds.), Physical Activity and Psychological Well-being. Routledge, London, pp. 46e62. Neptune, R.R., Hull, M.L., 1995. Accuracy assessment of methods for determining hip movement in seated cycling. J. Biomech. 28, 423e437. Schwarzer, U., Sommer, F., Klotz, T., Cremer, C., Engelmann, U., 2002. Cycling and penile oxygen pressure: the type of saddle matters. Eur. Urol. 41, 139e143. Straker, L., Levine, J., Campbell, A., 2009. The effects of walking and cycling computer workstations on keyboard and mouse performance. Hum. Factors 51, 831e844. Thompson, W.G., Levine, J.A., 2011. Productivity of transcriptionists using a treadmill desk. Work 40, 473e477. Weller, I., Corey, P., 1998. The impact of excluding non-leisure energy expenditure on the relation between physical activity and mortality in women. Epidemiology 9, 632e635. Weuve, J., Kang, J.H., Manson, J.E., Breteler, M.M., Ware, J.H., Grodstein, F., 2004. Physical activity, including walking, and cognitive function in older women. JAMA 292, 1454e1461. Wilmot, E., Edwardson, C., Achana, F., Davies, M., Gorely, T., Gray, L., Khunti, K., Yates, T., Biddle, S., 2012. Sedentary time in adults and the association with diabetes, cardiovascular disease and death: systematic review and meta-analysis. Diabetologia 55, 2895e2905. Wooles, A.L., Robinson, A.J., Keen, P.S., 2005. A static method for obtaining a calibration factor for SRM bicycle power cranks. Sports Eng. 8, 137e144. Zuntz, N., 1901. sber die Bedeutung der verschiedene Nahrstoffe als Erzeuber der Muskelkraft. Pflugers Arch. 83, 557e571.
