- Email: [email protected]
Estimating energy expenditure in wildland fire fighters using a physical activity monitor
Applied Ergonomics 33 (2002) 405–413
Estimating energy expenditure in wildland fire fighters using a physical activity monitor Daniel P. Heil* Movement Science/Human Performance Lab, Department of Health and Human Development, Montana State University, Hoseaus 101, Bozeman, MT 59717-3540, USA Received 1 August 2001; received in revised form 4 March 2002; accepted 10 May 2002
Abstract This study piloted the use of an electronic activity monitor (MTI AM 7164–1.2) as a tool for estimating activity (EEACT, kcal day1) and total (EETOT, kcal day1) energy expenditure in wildland fire fighters during extended periods of wildland fire suppression. Ten Hot Shot fire fighters (9 men, 1 woman) volunteered to wear a MTI monitor during every work shift for 21 consecutive days. Summarizing whole-body motion data each 1 min, the raw activity data (counts min1) were transformed into units of kcal min1 using a custom computer program with standard conversion equations. EETOT averaged (Mean7SD) 47687478 kcal day1, while EEACT averaged 25857406 kcal day1, neither of which differed significantly (P ¼ 0:198 and 0.268, respectively) from literature values reported for Hot Shots using the doubly labeled water technique. These data suggest that the electronic activity monitor provided reasonable estimates of EE in wildland fire fighters. This study should be verified, however, with a more complete validation methodology to ensure these findings. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Activity monitor; Firefighting; Total energy expenditure
1. Introduction Each summer, the western United States is plagued with wildland fires that are responsible for extensive property damage, as well as personal injury or even loss of life to those responsible for fighting wildland fires. During the 2000 wildland fire season, a total of 122,827 fires burned 8,422,237 acres (including 861 structures) which required 30,000+firefighting personnel at an estimated cost of 1.3 billion dollars (National Fire News, 2001). The personal risk that wildland fire fighters face when battling wildland fires cannot be overemphasized. On July 6th, 1994, for example, 12 wildland fire fighters and two helitack crew members were killed on Storm King Mountain fighting the South Canyon Fire (near Glenwood Springs, CO) (Butler et al., 1998). Since the middle of the last century, the most effective method for fighting wildland fires in the United States has been the use of specially trained elite fire fighting crews called Smoke Jumpers and Hot Shots. The *Tel.: +1-406-994-6324; fax: +1-406-994-6314. E-mail address: [email protected] (D.P. Heil).
regional deployment of both crew types is controlled by the National Interagency Fire Center (NIFC; Boise, ID), a coalition of six federal agencies concerned with fighting wildland fires. The Smoke Jumpers are airborne fire fighters that parachute from planes to attack wildland fires in remote and inaccessible areas. The Hot Shot crews, in contrast, are often given the toughest fire fighting duties that include fire suppression and cleanup activities (e.g. building firelines, setting backfires, and mopping up) that emphasize the use of specialized hand tools (e.g. chainsaws, shovels, and specialized tools called Pulaskis and McLeods) and even fireline explosives. Thus, while Smoke Jumpers are often first on the scene of a wildland fire, the Hot Shot crews are trained and expected to endure consecutive days or weeks of 10–16 h work shifts of mostly manual labor under extremes of high heat, high altitude, and poor air quality (i.e. smoke from the fires). In addition, Hot Shots perform these duties while carrying an 18–20 kg pack (which includes daily necessities like food and water, as well as emergency supplies such as a fire shelter) and wearing head-to-foot fire retardant clothing.
0003-6870/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 3 - 6 8 7 0 ( 0 2 ) 0 0 0 4 2 - X
406
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
Clearly, the physiological demands of fighting wildland fires as a Hot Shot are extremely demanding, potentially dangerous, and even life threatening. In light of incidents like the South Canyon Fire on Storm King Mountain (Butler et al., 1998), it seems prudent to better understand the physiological demands of long-term wildland fire fighting to help minimize the inherent health and safety risks of this particular occupation. With the exception of a series of projects by Ruby et al. (1999, 2000, 2001), however, very little is known about the demands of wildland firefighting in Hot Shot crews while actually fighting fires. This gap in the literature is due in part to the difficulty of collecting valid physiological data in remote and environmentally extreme conditions while not interfering with the Hot Shots’ ability to perform their jobs. Australian researchers have by addressed this methodological issue by using standard indirect calorimetry procedures (via the Douglas bag technique) in the field to measure energy expenditure (EE) during simulated fireline activities (Brotherhood et al., 1997), and controlled experimental fires (Budd et al., 1997). While the use of indirect calorimetry in lab or field settings is well suited for assessing EE of specific activities over short duration, it is not suitable for measuring EE over long duration and especially not during actual wildland fire suppression. An alternative measurement strategy for both occupationally (Burstein et al., 1996; Forbes-Ewan et al., 1989; Hoyt et al., 1991) and athletically (Pulfrey and Jones, 1996; Westerterp et al., 1986) demanding activities has been to determine total EE (instead of activity specific EE) resulting from long-term cumulative physiological stress. Ruby et al. (2001), for example, determined both total and activity EE (EETOT and EEACT, respectively, both in kcal day1) using the doubly labeled water (DLW) methodology during actual wildland fire suppression in three Hot Shot crews. Values for EETOT averaged 4878 and 3541 kcal day1 for Hot Shot men (n ¼ 8) and women (n ¼ 9), respectively, while all subjects together averaged 4160 kcal day1. Ruby et al. also reported average EEACT values of 2628 and 1754 kcal day1 for the same men and women, respectively, with a sample average of 2115 kcal day1. These results, in turn, have been used by Ruby et al. (1999, 2000) to evaluate issues of gender differences, body composition changes, adequacy of nutritional intake, and overall energy balance during active fire suppression. Methodologically, however, the DLW technique is exceptionally difficult under the environmental conditions that characterize wildland fire fighting, as well as being very expensive to perform on a large number of subjects. The DLW technique involves tracking the washout kinetics of body water that has been dosed via ingestion of isotope labeled hydrogen and oxygen (i.e. 2 H2O and H18 2 O). The rate at which these isotopes are
cleared from the body, as determined through analysis of urine samples collected over at least several consecutive days, is primarily a function of metabolic rate. Thus, not only does this technique rely on the collection, storage, and analysis of multiple urine samples, but the cost can be extremely high for physically active subjects ($1000–$1500 per subject) for a one-time analysis due to the cost of the isotope samples and subsequent analysis procedures. A complete description of the theory and recommended methodology underlying the DLW technique can be found in a report by the International Atomic Energy Agency (IAEA, 1990). In recent years, researchers studying the health consequences of free-living physical activity EE have used various commercially available electronic activity monitors. The monitors are typically very small and lightweight (similar in size and weight to a large wristwatch) and are typically worn around the waist like a pager such that daily activity patterns are minimally influenced by the measurement procedure (hence the term free-living). All electronic activity monitors are similar in that bodily movements, sensed by an accelerometer-type mechanism, are summarized by a battery-driven microprocessor. Once download to a computer, the stored raw movement data, which are in units specific to each activity device, are then converted to physiologically meaningful units (e.g. kcal min1) by way of a validated prediction equation. Several EE prediction equations specific to electronic activity monitors have been reported in the literature (Freedson et al., 1998; Hendelman et al., 2000; Swartz et al., 2000), each of which were derived from data collected for a variety of locomotion activities (e.g. walking at various speeds) indoor household tasks (e.g. window washing, sweeping, mopping, dusting, vacuuming, cooking), outdoor household tasks (lawn mowing, gardening), as well as physical conditioning activities (e.g. doubles tennis, softball, golf). While electronic activity monitors are fairly expensive ($400–$1000 each depending upon the brand), once purchased the monitors can be used repeatedly to provide minute-by-minute recordings of whole-body motion for consecutive days or weeks at a time. In addition, the detailed recordings can also be used to characterize domains of physical activity not possible with the DLW technique (i.e. frequency, intensity, and duration of individual activity bouts), as well as broader summary measures like EETOT and EEACT. To date, however, electronic activity monitors have not been rigorously tested for use in extreme work environments such as those experienced by wildland fire fighters. It was the purpose of the present study, therefore, to pilot the use of a commercially available electronic activity monitor for estimating both EETOT and EEACT in Hot Shot fire fighters during actual wildland fire suppression. Specifically, the present study was to
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
evaluate the methodology for collecting activity monitor data, as well as the algorithm for processing the activity monitor data. It was of interest to determine whether reasonable estimates of EETOT and EEACT could be obtained after several weeks of continuous data collection. These estimates were then compared to values derived using the DLW technique and reported in the literature by Ruby et al. (2001).
2. Methods 2.1. Subjects Ten Hot Shot wildland fire fighters from the Helena, MT, based crew volunteered to participate in the present study, which was approved by the Montana State University (Bozeman, MT) Human Subjects Review Committee. Due to a limited number of activity monitors (only five were available for this project), data was collected in two separate measurement periods (n ¼ 5 each period) during the summer of 2000 while fighting fires in the state of Montana. The first measurement period lasted from July 15 to August 5, while the second period was from August 20 to September 11. Informed consent was obtained from all subjects prior to their participation in the study. 2.2. Procedures and instrumentation Demographic data (i.e. body height, body mass, age) for each subject was obtained via questionnaire. To participate in the study, each subject was required to wear an MTI physical activity monitor (MTI AM 71641.2; Manufacturing Technology, Inc. Fort Walton Beach, FL)1 during all working hours for at least 21 consecutive days. Fire fighters were instructed to keep the monitors in the chest pocket of their work shirts which were worn only while on duty. The MTI monitors were mounted to 8 10 cm2 plastic cards that oriented the monitors vertically and minimized extraneous movement of the monitors within the work shirt pocket. When the measurement period ended, the MTI monitors were mailed back to the investigator for data downloading and analysis. The MTI monitors were programmed to summarize and record data about wholebody motion each 1 min after starting which would saturate the monitor’s memory after 21 consecutive days. The monitors were initialized and mailed by the investigator to the Hot Shots’ headquarters in Helena, MT, where they were then transferred to the Hot Shots themselves (with instructions) by their supervisor. The 1
This is the same activity monitor previously marketed and referred to in the research literature as the CSA AM 7164-1.2 by Computer Science Application, Inc., Shalimar, FL.
407
MTI monitors were all calibrated by the manufacturer prior to their use in this study. The MTI activity monitor is small (1.5 4.0 5.0 cm3) and light weight (42.5 g), has long-term data storage capacity (21 consecutive days with 1 min record intervals), will download stored data via a serial interface to a computer, and utilizes a single vertically oriented accelerometer (i.e. one dimensional, or uniaxial) to detect and sum activity counts while recording (e.g. activity counts per unit time). The accelerometer, which is designed to measure and record time varying accelerations (0.5–2.0 g0 s) with a frequency response of 0.25–2.5 Hz, is specifically designed to detect normal human motion and to reject motion from other sources (e.g. high-frequency vibrations). The acceleration signal is filtered by an analog bandpass filter and then digitized by an 8 bit A-to-D converter at 10 Hz. Individual A-toD values are then summed over a user-specified interval of time (epoch), where 1 min epochs are the most common interval used in human physical activity research. At 1 min epochs, the MTI AM 7164-1.2 activity monitor can record up to 22 consecutive days of data. The actual data collected by the CSA activity monitor is a series of numbers representing the level or intensity of activity in each epoch. Thus, the design of the MTIs motion quantifying mechanism makes it possible to summarize both the duration (by time) and intensity (by counts min1 or kcal min1) of individual physical activity bouts. Algorithms for converting counts min1 to kcal min1 during free-living EE using the MTI monitor have been validated with the monitor located at the waist line just behind the iliac crest (typically clipped to a belt) (Freedson et al., 1998; Hendelman et al., 2000; Swartz et al., 2000). Location of the MTI monitor on the waist line has been criticized for an inability to detect work performed by the upper body (Hendelman et al., 2000; Swartz et al., 2000). Given the amount of upper body work performed by Hot Shots, and the 18 kg backpack carried by each fire fighter at all times, location of the MTI monitor along the waist line was not an option. The chest pocket location was chosen primarily because there was little chance of losing or damaging the monitors. It was also anticipated that the chest pocket location would be more sensitive to upper body work. A separate study was performed to check the validation of the use of the MTI monitor in the shirt pocket placement versus the hip location (see Section 2.4 below). 2.3. Data processing algorithm The MTI activity monitors were programmed to save activity data once per minute which resulted in data files (after 21 days) totaling 30,800 data points each. The raw MTI data files (ASCII format) were then read and
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
408
processed by a custom data processing program written in Visual Basic (V6.0). The data processing algorithm included multiple steps for the estimation of both total (EETOT, kcal day1) and activity (EEACT, kcal day1) EE. EETOT was defined as the total daily caloric need of each fire fighter, while EEACT was defined as the daily energy expended during physical activity above basal metabolism. First, after a file was read into the data processing program, each 1 min data point (units of activity counts min1) was transformed into METs (i.e. metabolic equivalent) using a standard conversion formula from the literature (Swartz et al., 2000):
0.9 and 1.1 METs, respectively (Ainsworth et al., 2000), while EEDIT was assumed to equal 10% of EETOT. Estimates for EESLEEP and EEREST were then estimated inserting the respective MET values into Eq. (2) and multiplying the result by the amount time spent sleeping (T SLEEP) or resting (T REST). EETOT for each fire fighter for each full work day was then computed as the sum of EESLEEP, EEREST, and the kcal min1 values computed from Eq. (2) that corresponded to minutes of physical activity. X EETOT ¼ EESLEEP þ EEREST þ EEDIT þ EEi ; ð4Þ
METs ¼ 2:606 þ 0:0006863
P where EEi are individual kcal min1 values from Eq. (2) for each minute ‘i’ of the activity time period ranging from 1 to T ACT. Daily values P for EEACT, in turn, were derived from daily values EEi as follows:
ðMTI output; counts min1 Þ;
ð1Þ
where one MET is assumed to be equal to 3.5 ml kg1 min1 of oxygen consumed for aerobic metabolism (American College of Sports Medicine, 2000). Next, each transformed data point was converted to units of EE, kcal min1, as follows (American College of Sports Medicine, 2000): EEðkcal min 1Þ ¼ 3:5 ðMETsÞ ðTotal Mass; kgÞ=200;
ð2Þ
where METs is the value calculated from Eq. (1) and total mass is the sum of each fire fighter’s body mass and an assumed average backpack weight of 18 kg. From Eq. (2) results a minute-by-minute estimation of the rate of expenditure for each fire fighter. These EE estimates, however, are only valid for ‘‘light’’ to ‘‘vigorous’’ intensity physical activities (Swartz et al., 2000) and not resting or basal metabolism. Prior to deriving estimates of EETOT and EEACT, the time spent being physically active (T ACT, min day1), time awake but not active (T REST, min day1), and the time spent sleeping (T SLEEP, min day1) had to be determined. Values for T ACT were calculated as the sum of minutes that each fire fighter spent at a MTI output value >250 counts min1 (an activity count equivalent to above resting levels). Further, it was also assumed that T SLEEP was equal to 8 h per night (sleep logs were not kept), which left the calculation of the time spent in resting metabolism (T REST) as TREST ðminÞ ¼ 1440 TSLEEP TACT ;
ð3Þ
where 1440 is the total number minutes in a 24 hour period, T SLEEP is the number of minutes assumed sleeping (i.e. 480 min), and T ACT is the number of minutes spent being physically active (i.e. MTI output >250 counts min1). Estimated values of EETOT were assumed to be the sum of estimates for basal metabolism (EESLEEP, kcal day1; during sleep), resting metabolism (EEREST, kcal day1; awake but not physically active), and the thermic effect of diary intake (EEDIT, kcal day1). MET values for EESLEEP and EEREST were assumed equal to
EEACT ðkcal day1 Þ X ¼ ½EEi ðTACT Þi ðEESLEEP ; kcal min1 Þi ; ð5Þ where (EESLEEP)i is the 1 min estimate of EESLEEP, and EEACT is computed as the sum of 1 min EE values above that required during sleep. Since the actual value of EEDIT is not known, Eq. (4) cannot be solved and must be rewritten in a computationally equivalent manner such that it can be solved. Therefore, by assuming that EEDIT=0.10 EETOT (i.e. 10% of total EE), substituting 0.10 EETOT into Eq. (4) for EEDIT and then combining like terms, and finally solving for EETOT gives EETOT ¼ ½EESLEEP þ EEREST þ EEACT =0:90;
ð6Þ
where all of the unknowns (EESLEEP, EEREST, EEACT) can be computed as described above. For the purpose of data analyses, values of EETOT and EEACT were averaged over all full work days within the respective measurement periods for each fire fighter. 2.4. Validation of activity monitor placement Separate data collection and analyses were performed to determine the sensitivity of the MTI monitor in the chest pocket to both upper and lower body work activities common to wildland fire fighters. Three nonfire fighters wore MTIs in the chest pocket and the right hip during outdoor activities that emphasized lower body (two speeds of self-paced trail walking) and upper body work activities (wood chopping and chainsawing) while wearing a 20 kg backpack. The MTI monitors were programmed to record motion data over 1 -min intervals while each activity was performed for 5 min each at self-paced steady rates. For each activity, the last 3 min of data (for each monitor) were averaged for comparison.
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
2.5. Statistical analysis The statistical analysis for this project was primarily descriptive in nature with reference to the EETOT and EEACT values reported by Ruby et al. (2001) for comparison. Specifically, the present study values of EETOT and EEACT were compared directly to Ruby’s data using an independent T-test. In addition, MTI data from the activity monitor placement sub-study were evaluated using a 3-factor repeated measures ANOVA and Tukey’s post hoc test. This analysis focused on determining whether the activity monitor output varied as a function of monitor location (hip versus shirt pocket) and interactions between monitor location and type of activity (upper versus lower body activities). All tests were performed at an alpha level of 0.05.
3. Results 3.1. Summary of field data Demographic data for the 10 Hot Shot fire fighters are provided in Table 1. During the first measurement period, MTI data was collected during a total of 12 full work days, whereas only seven full work days were recorded during the second measurement period (the monitors arrived at the fire fighters camp 1 week after the monitors had been programmed to begin collecting data). MTI data (after transformation to units of kcal min1) for a single subject during the first measurement period is shown in Fig. 1. Note how full work days, which were defined as those days in which recorded physical activity during most of the day, are very apparent when compared to other work days. The total time spent in physical activity during full work
409
days (Mean7SD) was 6.470.9 h which required an average EEACT of 24227375 kcal day1, while EETOT averaged 47167435 kcal day1. When evaluated for the men only (9 of the 10 subjects), neither EEACT (24657371 kcal day1) nor EETOT (47647432 kcal day1) changed appreciably. Since 9 of the 10 Hot Shots evaluated were men, and the single women was similar in body size to many of the men (Table 1), the mean EETOT and EEACT values reported by Ruby et al. (2001) for men were compared to the present results. Neither EETOT nor EEACT differed significantly from Ruby’s values of 4878 kcal day1 (P ¼ 0:268) and 2628 kcal day1 (P ¼ 0:198), respectively. 3.2. Effect of activity monitor placement The results from the separate data collection that compared the placement of the activity monitor in the shirt pocket versus the hip location are shown in Table 2. The ANOVA results indicated no significant differences between the hip versus shirt pocket location of the activity monitors (P ¼ 0:75). In addition, the interaction between the activity monitor location and the type of activity (upper versus lower body) performed approached significance (P ¼ 0:09). Compared to the hip monitor, output for the chest monitor tended to be lower for trail walking (2.3% to 5.4%) and higher for upper body work activities (+13.2% to +13.6%), though none of these differences were significant.
4. Discussion Using a commercially available electronic physical activity monitor and a somewhat novel data processing strategy, the present study attempted to estimate total
Table 1 Demographic and summary statistics for estimating activity (EEACT) and total (EETOT) energy expenditure in Montana’s Hot Shot wildland fire fighters (n ¼ 10) Measurement period
Work days evaluateda
Gender
Body mass (kg)
Height (m)
Age (yr)
Minutes of activity/day
Average EEACT (kcal day1)
Average EETOT (kcal day1)
1 1 1 1 1 2 2 2 2 2
12 12 12 12 12 7 7 7 7 7
Women Man Man Man Man Man Man Man Man Man
79.54 84.09 90.91 84.09 79.54 84.09 95.46 72.73 81.82 70.46
1.70 1.83 1.78 1.83 1.70 1.83 1.96 1.80 1.75 1.70
29 25 20 21 44 33 20 26 31 30
379.5 446.2 487.7 482.2 426.8 366.1 347.7 380.8 370.0 344.7
2039 2715 3062 2922 2398 2482 2274 2302 1989 2042
4280 4983 5418 5125 4570 4921 5033 4397 4305 4125
Mean7SD
82.2777.48
1.7970.08
27.977.3
403.2753.6
24227375
47167435
a
‘‘Work days evaluated’’ only included full days of work on wildland fire suppression and ‘‘mop up’’ and did not include days with a significant amount of travel (and thus minimal energy expenditure) or days off (when the monitors were not worn).
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
410
Rate of Energy Expenditure (kcals/min)
12 W
W
W
10
8
6
4
2 0
1
2
3
4
5
6
7
8
(A)
9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (days)
Rate of Energy Expenditure (kcals/min)
12
W
W
W
W
W
W
W
W
10
8
6
4
2 12
13
14
15
(B)
16
17
18
19
20
21
22
Time (days)
Rate of Energy Expenditure (kcals/min)
12
(C)
W
W
10
8
6
4
2 14
15
16
Time (days)
Fig. 1. Example activity monitor output for a single Hot Shot fire fighter over 21 consecutive days (A), the last 10 days (B), as well as two consecutive days (C). Note that highly active work shifts (labeled with ‘‘W’’) are very apparent based upon the duration and intensity of activity for any given day. Numbering on x-axis is based upon the individual days of measurement for each subject.
(EETOT) and activity (EEACT) EE in a group of Hot Shot wildland fire fighters during actual wildland fire suppression. Coincidentally, the fire suppression activities evaluated by the activity monitors in this study
occurred during what is now considered one of the worst wildland fire seasons (summer of 2000) for the western US in over 50 years. Mean values for EETOT (4716 kcal day1) and EEACT (2422 kcal day1) for the
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
Helena Hot Shot crew members were statistically similar to the same variables reported by Ruby et al. (2001) (EETOT=4878 and EEACT=2628 kcal day1) using the DLW technique. Clearly, the proposed methodology by this investigation appears promising as a technique for estimating common indicators of cumulative physiological stress (i.e. EEACT and EETOT) in a unique and demanding occupational setting. It should be stressed, however, that this study was a pilot investigation of the proposed methodology and should be verified by a true field validation study where concomitant measures of EE from the same group of fire fighters are derived via electronic activity monitoring and a criterion measure of EE (e.g. DLW, indirect calorimetry, or direct observation). A methodological limitation of the present study was the comparison of EE values in one group of fire fighters using the MTI monitors with EE values from another group of fire fighters using the DLW technique as reported in the literature (Ruby et al., 2001). Given this limitation, the present study chose to determine average EEACT and EETOT values over multiple days of measurement (see Table 1) so that the influence of extremely active or inactive days was minimized. In addition, this methodology also helped to minimize the obvious issue with the present group of fire fighters not necessarily doing exactly the same types or volumes of activities as the fire fighters described by Ruby et al. (1999, 2000, 2001). However, the types of physical activities characterizing wildland fire suppression as described by Ruby et al. (2001) (extensive hiking with a 15–20 kg pack, chainsaw work, brush removal, and fire line construction with a Pulaski, as well as other manual labor specific to actual fire suppression and cleanup activities) are exactly the same as those activities characterizing the Helena Hot Shots. Therefore, by averaging the activity profiles for each fire fighter over multiple days, the present study design provided generalized EEACT and EETOT values that were not representative of any one type of activity common to wildland fire fighting. Indeed, the EE values reported by Ruby and coworkers were similarly derived with the DLW technique as averages over 3–5 successive days of measurement. The EETOT values for Hot Shot fire fighters reported by Ruby et al. (2001) and the present study are similar to those reported for other occupationally (Burstein et al., 1996; Forbes-Ewan et al., 1989; Hoyt et al., 1991) and athletically (Pulfrey and Jones, 1996; Westerterp et al., 1986) demanding activities. For example, an average EETOT of 4634 kcal day1 was estimated for six mountaineers climbing Mt. Shisha Pangma (8046 m elevation) (Pulfrey and Jones, 1996), while an average 8054 kcal day1 were required to ride in the Tour de France bicycle race (Westerterp et al., 1986). Similarly, an Army platoon training for jungle warfare averaged
411
4750 kcal day1 (Forbes-Ewan et al., 1989), while Marines expended 4919 kcal day1 on average during 11 days of cold weather field exercises (Hoyt et al., 1991). All of the above-mentioned studies have two commonalities: (1) Measurement of EETOT in physically active subjects in an extreme environment and/or under physically demanding circumstances; (2) measurement of EETOT by the DLW technique. Use of the DLW technique is certainly the preferred method by researchers when lab-based research is not possible and finances are not limiting. Use of electronic activity monitors, however, as outlined by the present study, may serve as a practical alternative measurement strategy in many of the above study environments. In addition to measures of EETOT and EEACT, it is also possible to derive measures of physical activity behavior from activity monitor data. Fig. 1C, for example, illustrates the continually changing pattern of EE for a single Hot Shot over two successive days. From this figure, several physical activity behavior patterns are apparent. First, the rate of EE rarely exceeded 8 kcal min1 and tended to oscillate between 4 and 6 kcal min1 most frequently. Clearly, experienced wildland fire fighters know that their capacity to perform work must last over 10–16 h each day and will pace themselves accordingly. This may suggest that submaximal work capacity (i.e. as indicated by lactate threshold) may be a better determinant of work capacity in Hot Shots than maximal aerobic capacity (i.e. maximal oxygen uptake). Secondly, physical activity in Fig. 1C tends to oscillate in 15–60 min bouts with distinct breaks between bouts. Again, Hot Shots are well educated in the need to stay hydrated and eat frequently in order to withstand the daily rigors of working the fire line. Lastly, the information collected by the MTI activity monitors is far more detailed than anything is possible using the DLW technique. It is possible, for example, to evaluate each fire fighter’s data day-by-day, hour-to-hour, or even minute-by-minute for upwards of 21 consecutive days. Clearly, this may actually be too much data for most practical needs, but with the use of customized computer programs the volume of data is really not an issue. While not the focus of the present study, the use of electronic monitoring devices to track physical activity behavior may allow researchers in the future to better understand issues related to hydration, energy intake, and energy balance during prolonged bouts of high EE in remote field settings. To the author’s knowledge, the present study is the first to attempt estimating EE with electronic activity monitors located on the chest (i.e. work shirt pocket). With the exception of rare attempts to wear activity monitors on the wrist (Swartz et al., 2000), virtually all other electronic activity monitor validation studies have been performed with the monitor located along the
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
412
Table 2 Comparison of MTI activity monitor output from the hip versus the chest pocket during four activities common to Hot Shot wildland fire fighters a
Physical activity
MTI monitor output at the hip (counts min1)
MTI monitor output in shirt chest pocket (counts min1)
Percent difference in output between hip and chest locations (%)
Self-paced ‘‘slow’’ walking Self-paced ‘‘moderate’’ walking Chopping wood Chainsawing wood
3630 5168 2614 533
3444 5051 3010 617
5.4 2.3 13.2 13.6
a
Percent difference=[1(MTI output at hip/MTI output at chest)] 100.
waist line (Freedson et al., 1998; Hendelman et al., 2000; Swartz et al., 2001). The rationale for the waist location is based upon the assumption that most EE associated with physical activity is best estimated by tracking the motions of the body’s center of mass. The waist location was not an option for the Hot Shot fire fighters, however, because of the waist pack that is worn while working. If the activity monitor was located within the pack there was a chance that the monitor would move around with the contents of the pack and orientation of the monitor in a vertical direction is critical to proper motion detection. Similarly, location of the monitor on the waist strap of the packs was not considered an option because of the likelihood of physical damage or actually losing the monitor completely. Given these constraints, the chest pocket seemed to be a reasonable alternative especially given that Hot Shot fire fighters are well known for their habit of wearing their fire retardant work shirts only while on duty. The main problem with the chest pocket location, of course, was the lack of a validation study to support this methodological characteristic. To address this issue, a sub-study within the present study was performed on three subjects varying considerably in body size (57– 92 kg) and gender (two men and one women) where each subject performed various outdoor activities with an MTI activity monitor around their waist and another within their shirt chest pocket. Each subject wore a 20 kg backpack while performing two self-selected speeds of overground walking, wood chopping with an axe, as well as chainsawing. Given the amount of upper body activities required for wildland fire fighting (i.e. brush removal, fire line construction, etc.), the sub-study was designed to determine if a systematic bias could be expected with the activity monitor located within the chest pocket versus on the waist. The results, which are summarized in Table 2, suggested that an activity monitor located within the chest pocket would slightly underestimate physical activity EE when compared to a waist-mounted monitor during locomotion activities like hiking overground. However, for the activities that emphasized upper body work (wood chopping and chainsawing), the chest pocket monitor tended to overestimate physical activity EE. Based upon these
data it is certain that EE profiles derived for each fire fighter, such as that shown in Fig. 1, represents a combination of under- and overestimates for physical activity EE. The use of an activity monitor designed with a triaxial accelerometer (sensing bodily motion in three planes of motion instead of only one) could provide more accurate information about EE when evaluating specific activity bouts in Hot Shot fire fighters. A recent study by Hendelman et al. (2000), however, found that both uniaxial- and triaxial-based activity monitors have the same motion sensing problems when worn around the waist. Despite the fact that a triaxial accelerometer should provide more detailed information about bodily motion, it is possible that it is not the right kind of information given that EE for any specific activity is dependent upon factors unknown to the activity monitor (e.g. upper versus lower body work, fatigue, load carriage, etc.). In summary, the present study has demonstrated that use of electronic activity monitors to estimate total and activity EE during prolonged bouts of physical activity under environmentally extreme conditions may serve as a practical alternative to more invasive and costly techniques. A field validation study that incorporates concomitant measures of EE from both activity monitors and a criterion method (DLW technique, indirect calorimetry, or direct observation) should be performed prior to full acceptance of the proposed methodology and data processing algorithm. Future use of electronic activity monitors under similar circumstances may focus on physical activity behaviors of fire fighters and how this relates to hydration status and maintenance of energy balance.
Acknowledgements This project was supported in part by a loan of activity monitors from Manufacturing Technology, Inc. (Fort Walton Beach, FL). The author gratefully acknowledges the assistance of Larry Edwards, Supervisor of the Helena Hot Shots, for coordinating the assignment and retrieval of activity monitors to Hot
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
Shots on the fire lines, as well as the willingness of the Helena Hot Shots themselves to participate in this study.
References Ainsworth, B.E., Haskell, W.L., Whitt, M.C., Irwin, M.L., Swartz, A.M., Strath, S.J., O’Brien, W.L., Bassett, D.R., Schmitz, K.H., Emplaincourt, P.O., Jacobs, D.R., Leon, A.S., 2000. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exercise 32 (9), S498–S516. American College of Sports Medicine, 2000. ACSM’s Guidelines for Exercise Testing and Prescription, 6th Edition. Lippincott Williams & Wilkens, Baltimore, MD. Brotherhood, J.R., Budd, G.M., Hendrie, A.L., Jeffery, S.E., Beasley, F.A., Costin, B.P., Zhien, W., Baker, M.M., Cheney, N.P., Dawson, M.P., 1997. Project Aquarius 3. Effects of work rate on the productivity, energy expenditure, and physiological responses of men building fireline with a rakehoe in dry eucalypt forest. Int. J. Wildland Fire 7 (2), 87–98. Budd, G.M., Brotherhood, J.R., Hendrie, A.L., Jeffery, S.E., Beasley, F.A., Costin, B.P., Zhien, W., Baker, M.M., Cheney, N.P., Dawson, M.P., 1997. Project Aquarius 5. Activity distribution, energy expenditure, and productivity of men free-running wildland fires with hand tools. Int. J. Wildland Fire 7 (2), 105–118. Burstein, R., Coward, A.W., Askew, W.E., Carmel, K., Irving, C., Shpilberg, O., Moran, D., Pikarsky, A., Ginot, G., Sawye, M., Golan, R., Epstein, Y., 1996. Energy expenditure variations in soldiers performing military activities under cold and hot climate conditions. Mil. Med. 161 (12), 750–754. Butler, B.W., Bartlette, R.A., Bradshaw, L.S., Cohen, J.D., Andrews, P.L., Putnam, T., Mangan, R.J., 1998. Fire Behavior Associated with the South Canyon Fire on Storm King Mountain, Colorado (Research Paper RMRS-RP-9). USDA Forest Service, Rocky Mountain Research Station: Fort Collins, CO. Forbes-Ewan, C.H., Morrissey, B.L.L., Gregg, G.C., Waters, D.R., 1989. Use of doubly labeled water technique in soldiers training for jungle warfare. J. Appl. Physiol. 67 (1), 14–18.
413
Freedson, P.S., Melanson, E., Sirard, J., 1998. Calibration of the computer science and applications, Inc accelerometer. Med. Sci. Sports Exercise 30 (5), 777–781. Hendelman, D., Miller, K., Baggett, C., Debold, E., Freedson, P., 2000. Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Med. Sci. Sports Exercise 32 (9), S442–S449. Hoyt, R.W., Jones, T.E., Stein, T.P., McAninch, G.W., Lieberman, H.R., Askew, E.W., Cymerman, A., 1991. Doubly labeled water measurement of human energy expenditure during strenuous exercise. J. Appl. Physiol. 71 (1), 16–22. IAEA, 1990. In: Prentice, A. M. (Ed). The Doubly Labelled Water Method for Measuring Energy Expenditure: a Consensus Report by the IDECG Working Group. I/D/E/C/G International Dietary Energy Consultancy Group, NAHRES-4, IAEA, Vienna, Italy. National Fire News, 2001. Fire Season 2000 Statistics. National Interagency Fire Center, Boise, ID. Pulfrey, S.M., Jones, P.J.H., 1996. Energy expenditure and requirement while climbing above 6000 m. J. Appl. Physiol. 81 (3), 1306– 1311. Ruby, B.C., Zderic, T.W., Burks, C.A., Tysk, S., Sharkey, B.J., 1999. Total energy expenditure (doubly labeled water) and bone resorption during wildland fire suppression. Med. Sci. Sports Exercise 31 (5), S1863. Ruby, B.C., Schoeller, D.A., Sharkey, B.J., 2000. Gender differences in water turnover, energy balance and body composition during arduous work. Med. Sci. Sports Exercise 32 (5), S17. Ruby, B.C., Schoeller, D.A., Sharkey, B.J., 2001. Evaluation of total energy expenditure (doubly labeled water) across different measurement periods during arduous work. Med. Sci. Sports Exercise 33 (5), S1545. Swartz, A.M., Strath, S.J., Bassett, D.R., O’Brien, W.L., King, G.A., Ainsworth, B.E., 2000. Estimation of energy expenditure using MTI accelerometers at hip and wrist sites. Med. Sci. Sports Exercise 32 (9), S450–S456. Westerterp, K.R., Saris, W.H.M., van Es, M., ten Hoor, F., 1986. Use of the doubly labeled water technique in humans during heavy sustained exercise. J. Appl. Physiol. 61 (6), 2162–2167.
Estimating energy expenditure in wildland fire fighters using a physical activity monitor Daniel P. Heil* Movement Science/Human Performance Lab, Department of Health and Human Development, Montana State University, Hoseaus 101, Bozeman, MT 59717-3540, USA Received 1 August 2001; received in revised form 4 March 2002; accepted 10 May 2002
Abstract This study piloted the use of an electronic activity monitor (MTI AM 7164–1.2) as a tool for estimating activity (EEACT, kcal day1) and total (EETOT, kcal day1) energy expenditure in wildland fire fighters during extended periods of wildland fire suppression. Ten Hot Shot fire fighters (9 men, 1 woman) volunteered to wear a MTI monitor during every work shift for 21 consecutive days. Summarizing whole-body motion data each 1 min, the raw activity data (counts min1) were transformed into units of kcal min1 using a custom computer program with standard conversion equations. EETOT averaged (Mean7SD) 47687478 kcal day1, while EEACT averaged 25857406 kcal day1, neither of which differed significantly (P ¼ 0:198 and 0.268, respectively) from literature values reported for Hot Shots using the doubly labeled water technique. These data suggest that the electronic activity monitor provided reasonable estimates of EE in wildland fire fighters. This study should be verified, however, with a more complete validation methodology to ensure these findings. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Activity monitor; Firefighting; Total energy expenditure
1. Introduction Each summer, the western United States is plagued with wildland fires that are responsible for extensive property damage, as well as personal injury or even loss of life to those responsible for fighting wildland fires. During the 2000 wildland fire season, a total of 122,827 fires burned 8,422,237 acres (including 861 structures) which required 30,000+firefighting personnel at an estimated cost of 1.3 billion dollars (National Fire News, 2001). The personal risk that wildland fire fighters face when battling wildland fires cannot be overemphasized. On July 6th, 1994, for example, 12 wildland fire fighters and two helitack crew members were killed on Storm King Mountain fighting the South Canyon Fire (near Glenwood Springs, CO) (Butler et al., 1998). Since the middle of the last century, the most effective method for fighting wildland fires in the United States has been the use of specially trained elite fire fighting crews called Smoke Jumpers and Hot Shots. The *Tel.: +1-406-994-6324; fax: +1-406-994-6314. E-mail address: [email protected] (D.P. Heil).
regional deployment of both crew types is controlled by the National Interagency Fire Center (NIFC; Boise, ID), a coalition of six federal agencies concerned with fighting wildland fires. The Smoke Jumpers are airborne fire fighters that parachute from planes to attack wildland fires in remote and inaccessible areas. The Hot Shot crews, in contrast, are often given the toughest fire fighting duties that include fire suppression and cleanup activities (e.g. building firelines, setting backfires, and mopping up) that emphasize the use of specialized hand tools (e.g. chainsaws, shovels, and specialized tools called Pulaskis and McLeods) and even fireline explosives. Thus, while Smoke Jumpers are often first on the scene of a wildland fire, the Hot Shot crews are trained and expected to endure consecutive days or weeks of 10–16 h work shifts of mostly manual labor under extremes of high heat, high altitude, and poor air quality (i.e. smoke from the fires). In addition, Hot Shots perform these duties while carrying an 18–20 kg pack (which includes daily necessities like food and water, as well as emergency supplies such as a fire shelter) and wearing head-to-foot fire retardant clothing.
0003-6870/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 3 - 6 8 7 0 ( 0 2 ) 0 0 0 4 2 - X
406
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
Clearly, the physiological demands of fighting wildland fires as a Hot Shot are extremely demanding, potentially dangerous, and even life threatening. In light of incidents like the South Canyon Fire on Storm King Mountain (Butler et al., 1998), it seems prudent to better understand the physiological demands of long-term wildland fire fighting to help minimize the inherent health and safety risks of this particular occupation. With the exception of a series of projects by Ruby et al. (1999, 2000, 2001), however, very little is known about the demands of wildland firefighting in Hot Shot crews while actually fighting fires. This gap in the literature is due in part to the difficulty of collecting valid physiological data in remote and environmentally extreme conditions while not interfering with the Hot Shots’ ability to perform their jobs. Australian researchers have by addressed this methodological issue by using standard indirect calorimetry procedures (via the Douglas bag technique) in the field to measure energy expenditure (EE) during simulated fireline activities (Brotherhood et al., 1997), and controlled experimental fires (Budd et al., 1997). While the use of indirect calorimetry in lab or field settings is well suited for assessing EE of specific activities over short duration, it is not suitable for measuring EE over long duration and especially not during actual wildland fire suppression. An alternative measurement strategy for both occupationally (Burstein et al., 1996; Forbes-Ewan et al., 1989; Hoyt et al., 1991) and athletically (Pulfrey and Jones, 1996; Westerterp et al., 1986) demanding activities has been to determine total EE (instead of activity specific EE) resulting from long-term cumulative physiological stress. Ruby et al. (2001), for example, determined both total and activity EE (EETOT and EEACT, respectively, both in kcal day1) using the doubly labeled water (DLW) methodology during actual wildland fire suppression in three Hot Shot crews. Values for EETOT averaged 4878 and 3541 kcal day1 for Hot Shot men (n ¼ 8) and women (n ¼ 9), respectively, while all subjects together averaged 4160 kcal day1. Ruby et al. also reported average EEACT values of 2628 and 1754 kcal day1 for the same men and women, respectively, with a sample average of 2115 kcal day1. These results, in turn, have been used by Ruby et al. (1999, 2000) to evaluate issues of gender differences, body composition changes, adequacy of nutritional intake, and overall energy balance during active fire suppression. Methodologically, however, the DLW technique is exceptionally difficult under the environmental conditions that characterize wildland fire fighting, as well as being very expensive to perform on a large number of subjects. The DLW technique involves tracking the washout kinetics of body water that has been dosed via ingestion of isotope labeled hydrogen and oxygen (i.e. 2 H2O and H18 2 O). The rate at which these isotopes are
cleared from the body, as determined through analysis of urine samples collected over at least several consecutive days, is primarily a function of metabolic rate. Thus, not only does this technique rely on the collection, storage, and analysis of multiple urine samples, but the cost can be extremely high for physically active subjects ($1000–$1500 per subject) for a one-time analysis due to the cost of the isotope samples and subsequent analysis procedures. A complete description of the theory and recommended methodology underlying the DLW technique can be found in a report by the International Atomic Energy Agency (IAEA, 1990). In recent years, researchers studying the health consequences of free-living physical activity EE have used various commercially available electronic activity monitors. The monitors are typically very small and lightweight (similar in size and weight to a large wristwatch) and are typically worn around the waist like a pager such that daily activity patterns are minimally influenced by the measurement procedure (hence the term free-living). All electronic activity monitors are similar in that bodily movements, sensed by an accelerometer-type mechanism, are summarized by a battery-driven microprocessor. Once download to a computer, the stored raw movement data, which are in units specific to each activity device, are then converted to physiologically meaningful units (e.g. kcal min1) by way of a validated prediction equation. Several EE prediction equations specific to electronic activity monitors have been reported in the literature (Freedson et al., 1998; Hendelman et al., 2000; Swartz et al., 2000), each of which were derived from data collected for a variety of locomotion activities (e.g. walking at various speeds) indoor household tasks (e.g. window washing, sweeping, mopping, dusting, vacuuming, cooking), outdoor household tasks (lawn mowing, gardening), as well as physical conditioning activities (e.g. doubles tennis, softball, golf). While electronic activity monitors are fairly expensive ($400–$1000 each depending upon the brand), once purchased the monitors can be used repeatedly to provide minute-by-minute recordings of whole-body motion for consecutive days or weeks at a time. In addition, the detailed recordings can also be used to characterize domains of physical activity not possible with the DLW technique (i.e. frequency, intensity, and duration of individual activity bouts), as well as broader summary measures like EETOT and EEACT. To date, however, electronic activity monitors have not been rigorously tested for use in extreme work environments such as those experienced by wildland fire fighters. It was the purpose of the present study, therefore, to pilot the use of a commercially available electronic activity monitor for estimating both EETOT and EEACT in Hot Shot fire fighters during actual wildland fire suppression. Specifically, the present study was to
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
evaluate the methodology for collecting activity monitor data, as well as the algorithm for processing the activity monitor data. It was of interest to determine whether reasonable estimates of EETOT and EEACT could be obtained after several weeks of continuous data collection. These estimates were then compared to values derived using the DLW technique and reported in the literature by Ruby et al. (2001).
2. Methods 2.1. Subjects Ten Hot Shot wildland fire fighters from the Helena, MT, based crew volunteered to participate in the present study, which was approved by the Montana State University (Bozeman, MT) Human Subjects Review Committee. Due to a limited number of activity monitors (only five were available for this project), data was collected in two separate measurement periods (n ¼ 5 each period) during the summer of 2000 while fighting fires in the state of Montana. The first measurement period lasted from July 15 to August 5, while the second period was from August 20 to September 11. Informed consent was obtained from all subjects prior to their participation in the study. 2.2. Procedures and instrumentation Demographic data (i.e. body height, body mass, age) for each subject was obtained via questionnaire. To participate in the study, each subject was required to wear an MTI physical activity monitor (MTI AM 71641.2; Manufacturing Technology, Inc. Fort Walton Beach, FL)1 during all working hours for at least 21 consecutive days. Fire fighters were instructed to keep the monitors in the chest pocket of their work shirts which were worn only while on duty. The MTI monitors were mounted to 8 10 cm2 plastic cards that oriented the monitors vertically and minimized extraneous movement of the monitors within the work shirt pocket. When the measurement period ended, the MTI monitors were mailed back to the investigator for data downloading and analysis. The MTI monitors were programmed to summarize and record data about wholebody motion each 1 min after starting which would saturate the monitor’s memory after 21 consecutive days. The monitors were initialized and mailed by the investigator to the Hot Shots’ headquarters in Helena, MT, where they were then transferred to the Hot Shots themselves (with instructions) by their supervisor. The 1
This is the same activity monitor previously marketed and referred to in the research literature as the CSA AM 7164-1.2 by Computer Science Application, Inc., Shalimar, FL.
407
MTI monitors were all calibrated by the manufacturer prior to their use in this study. The MTI activity monitor is small (1.5 4.0 5.0 cm3) and light weight (42.5 g), has long-term data storage capacity (21 consecutive days with 1 min record intervals), will download stored data via a serial interface to a computer, and utilizes a single vertically oriented accelerometer (i.e. one dimensional, or uniaxial) to detect and sum activity counts while recording (e.g. activity counts per unit time). The accelerometer, which is designed to measure and record time varying accelerations (0.5–2.0 g0 s) with a frequency response of 0.25–2.5 Hz, is specifically designed to detect normal human motion and to reject motion from other sources (e.g. high-frequency vibrations). The acceleration signal is filtered by an analog bandpass filter and then digitized by an 8 bit A-to-D converter at 10 Hz. Individual A-toD values are then summed over a user-specified interval of time (epoch), where 1 min epochs are the most common interval used in human physical activity research. At 1 min epochs, the MTI AM 7164-1.2 activity monitor can record up to 22 consecutive days of data. The actual data collected by the CSA activity monitor is a series of numbers representing the level or intensity of activity in each epoch. Thus, the design of the MTIs motion quantifying mechanism makes it possible to summarize both the duration (by time) and intensity (by counts min1 or kcal min1) of individual physical activity bouts. Algorithms for converting counts min1 to kcal min1 during free-living EE using the MTI monitor have been validated with the monitor located at the waist line just behind the iliac crest (typically clipped to a belt) (Freedson et al., 1998; Hendelman et al., 2000; Swartz et al., 2000). Location of the MTI monitor on the waist line has been criticized for an inability to detect work performed by the upper body (Hendelman et al., 2000; Swartz et al., 2000). Given the amount of upper body work performed by Hot Shots, and the 18 kg backpack carried by each fire fighter at all times, location of the MTI monitor along the waist line was not an option. The chest pocket location was chosen primarily because there was little chance of losing or damaging the monitors. It was also anticipated that the chest pocket location would be more sensitive to upper body work. A separate study was performed to check the validation of the use of the MTI monitor in the shirt pocket placement versus the hip location (see Section 2.4 below). 2.3. Data processing algorithm The MTI activity monitors were programmed to save activity data once per minute which resulted in data files (after 21 days) totaling 30,800 data points each. The raw MTI data files (ASCII format) were then read and
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
408
processed by a custom data processing program written in Visual Basic (V6.0). The data processing algorithm included multiple steps for the estimation of both total (EETOT, kcal day1) and activity (EEACT, kcal day1) EE. EETOT was defined as the total daily caloric need of each fire fighter, while EEACT was defined as the daily energy expended during physical activity above basal metabolism. First, after a file was read into the data processing program, each 1 min data point (units of activity counts min1) was transformed into METs (i.e. metabolic equivalent) using a standard conversion formula from the literature (Swartz et al., 2000):
0.9 and 1.1 METs, respectively (Ainsworth et al., 2000), while EEDIT was assumed to equal 10% of EETOT. Estimates for EESLEEP and EEREST were then estimated inserting the respective MET values into Eq. (2) and multiplying the result by the amount time spent sleeping (T SLEEP) or resting (T REST). EETOT for each fire fighter for each full work day was then computed as the sum of EESLEEP, EEREST, and the kcal min1 values computed from Eq. (2) that corresponded to minutes of physical activity. X EETOT ¼ EESLEEP þ EEREST þ EEDIT þ EEi ; ð4Þ
METs ¼ 2:606 þ 0:0006863
P where EEi are individual kcal min1 values from Eq. (2) for each minute ‘i’ of the activity time period ranging from 1 to T ACT. Daily values P for EEACT, in turn, were derived from daily values EEi as follows:
ðMTI output; counts min1 Þ;
ð1Þ
where one MET is assumed to be equal to 3.5 ml kg1 min1 of oxygen consumed for aerobic metabolism (American College of Sports Medicine, 2000). Next, each transformed data point was converted to units of EE, kcal min1, as follows (American College of Sports Medicine, 2000): EEðkcal min 1Þ ¼ 3:5 ðMETsÞ ðTotal Mass; kgÞ=200;
ð2Þ
where METs is the value calculated from Eq. (1) and total mass is the sum of each fire fighter’s body mass and an assumed average backpack weight of 18 kg. From Eq. (2) results a minute-by-minute estimation of the rate of expenditure for each fire fighter. These EE estimates, however, are only valid for ‘‘light’’ to ‘‘vigorous’’ intensity physical activities (Swartz et al., 2000) and not resting or basal metabolism. Prior to deriving estimates of EETOT and EEACT, the time spent being physically active (T ACT, min day1), time awake but not active (T REST, min day1), and the time spent sleeping (T SLEEP, min day1) had to be determined. Values for T ACT were calculated as the sum of minutes that each fire fighter spent at a MTI output value >250 counts min1 (an activity count equivalent to above resting levels). Further, it was also assumed that T SLEEP was equal to 8 h per night (sleep logs were not kept), which left the calculation of the time spent in resting metabolism (T REST) as TREST ðminÞ ¼ 1440 TSLEEP TACT ;
ð3Þ
where 1440 is the total number minutes in a 24 hour period, T SLEEP is the number of minutes assumed sleeping (i.e. 480 min), and T ACT is the number of minutes spent being physically active (i.e. MTI output >250 counts min1). Estimated values of EETOT were assumed to be the sum of estimates for basal metabolism (EESLEEP, kcal day1; during sleep), resting metabolism (EEREST, kcal day1; awake but not physically active), and the thermic effect of diary intake (EEDIT, kcal day1). MET values for EESLEEP and EEREST were assumed equal to
EEACT ðkcal day1 Þ X ¼ ½EEi ðTACT Þi ðEESLEEP ; kcal min1 Þi ; ð5Þ where (EESLEEP)i is the 1 min estimate of EESLEEP, and EEACT is computed as the sum of 1 min EE values above that required during sleep. Since the actual value of EEDIT is not known, Eq. (4) cannot be solved and must be rewritten in a computationally equivalent manner such that it can be solved. Therefore, by assuming that EEDIT=0.10 EETOT (i.e. 10% of total EE), substituting 0.10 EETOT into Eq. (4) for EEDIT and then combining like terms, and finally solving for EETOT gives EETOT ¼ ½EESLEEP þ EEREST þ EEACT =0:90;
ð6Þ
where all of the unknowns (EESLEEP, EEREST, EEACT) can be computed as described above. For the purpose of data analyses, values of EETOT and EEACT were averaged over all full work days within the respective measurement periods for each fire fighter. 2.4. Validation of activity monitor placement Separate data collection and analyses were performed to determine the sensitivity of the MTI monitor in the chest pocket to both upper and lower body work activities common to wildland fire fighters. Three nonfire fighters wore MTIs in the chest pocket and the right hip during outdoor activities that emphasized lower body (two speeds of self-paced trail walking) and upper body work activities (wood chopping and chainsawing) while wearing a 20 kg backpack. The MTI monitors were programmed to record motion data over 1 -min intervals while each activity was performed for 5 min each at self-paced steady rates. For each activity, the last 3 min of data (for each monitor) were averaged for comparison.
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
2.5. Statistical analysis The statistical analysis for this project was primarily descriptive in nature with reference to the EETOT and EEACT values reported by Ruby et al. (2001) for comparison. Specifically, the present study values of EETOT and EEACT were compared directly to Ruby’s data using an independent T-test. In addition, MTI data from the activity monitor placement sub-study were evaluated using a 3-factor repeated measures ANOVA and Tukey’s post hoc test. This analysis focused on determining whether the activity monitor output varied as a function of monitor location (hip versus shirt pocket) and interactions between monitor location and type of activity (upper versus lower body activities). All tests were performed at an alpha level of 0.05.
3. Results 3.1. Summary of field data Demographic data for the 10 Hot Shot fire fighters are provided in Table 1. During the first measurement period, MTI data was collected during a total of 12 full work days, whereas only seven full work days were recorded during the second measurement period (the monitors arrived at the fire fighters camp 1 week after the monitors had been programmed to begin collecting data). MTI data (after transformation to units of kcal min1) for a single subject during the first measurement period is shown in Fig. 1. Note how full work days, which were defined as those days in which recorded physical activity during most of the day, are very apparent when compared to other work days. The total time spent in physical activity during full work
409
days (Mean7SD) was 6.470.9 h which required an average EEACT of 24227375 kcal day1, while EETOT averaged 47167435 kcal day1. When evaluated for the men only (9 of the 10 subjects), neither EEACT (24657371 kcal day1) nor EETOT (47647432 kcal day1) changed appreciably. Since 9 of the 10 Hot Shots evaluated were men, and the single women was similar in body size to many of the men (Table 1), the mean EETOT and EEACT values reported by Ruby et al. (2001) for men were compared to the present results. Neither EETOT nor EEACT differed significantly from Ruby’s values of 4878 kcal day1 (P ¼ 0:268) and 2628 kcal day1 (P ¼ 0:198), respectively. 3.2. Effect of activity monitor placement The results from the separate data collection that compared the placement of the activity monitor in the shirt pocket versus the hip location are shown in Table 2. The ANOVA results indicated no significant differences between the hip versus shirt pocket location of the activity monitors (P ¼ 0:75). In addition, the interaction between the activity monitor location and the type of activity (upper versus lower body) performed approached significance (P ¼ 0:09). Compared to the hip monitor, output for the chest monitor tended to be lower for trail walking (2.3% to 5.4%) and higher for upper body work activities (+13.2% to +13.6%), though none of these differences were significant.
4. Discussion Using a commercially available electronic physical activity monitor and a somewhat novel data processing strategy, the present study attempted to estimate total
Table 1 Demographic and summary statistics for estimating activity (EEACT) and total (EETOT) energy expenditure in Montana’s Hot Shot wildland fire fighters (n ¼ 10) Measurement period
Work days evaluateda
Gender
Body mass (kg)
Height (m)
Age (yr)
Minutes of activity/day
Average EEACT (kcal day1)
Average EETOT (kcal day1)
1 1 1 1 1 2 2 2 2 2
12 12 12 12 12 7 7 7 7 7
Women Man Man Man Man Man Man Man Man Man
79.54 84.09 90.91 84.09 79.54 84.09 95.46 72.73 81.82 70.46
1.70 1.83 1.78 1.83 1.70 1.83 1.96 1.80 1.75 1.70
29 25 20 21 44 33 20 26 31 30
379.5 446.2 487.7 482.2 426.8 366.1 347.7 380.8 370.0 344.7
2039 2715 3062 2922 2398 2482 2274 2302 1989 2042
4280 4983 5418 5125 4570 4921 5033 4397 4305 4125
Mean7SD
82.2777.48
1.7970.08
27.977.3
403.2753.6
24227375
47167435
a
‘‘Work days evaluated’’ only included full days of work on wildland fire suppression and ‘‘mop up’’ and did not include days with a significant amount of travel (and thus minimal energy expenditure) or days off (when the monitors were not worn).
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
410
Rate of Energy Expenditure (kcals/min)
12 W
W
W
10
8
6
4
2 0
1
2
3
4
5
6
7
8
(A)
9 10 11 12 13 14 15 16 17 18 19 20 21 22 Time (days)
Rate of Energy Expenditure (kcals/min)
12
W
W
W
W
W
W
W
W
10
8
6
4
2 12
13
14
15
(B)
16
17
18
19
20
21
22
Time (days)
Rate of Energy Expenditure (kcals/min)
12
(C)
W
W
10
8
6
4
2 14
15
16
Time (days)
Fig. 1. Example activity monitor output for a single Hot Shot fire fighter over 21 consecutive days (A), the last 10 days (B), as well as two consecutive days (C). Note that highly active work shifts (labeled with ‘‘W’’) are very apparent based upon the duration and intensity of activity for any given day. Numbering on x-axis is based upon the individual days of measurement for each subject.
(EETOT) and activity (EEACT) EE in a group of Hot Shot wildland fire fighters during actual wildland fire suppression. Coincidentally, the fire suppression activities evaluated by the activity monitors in this study
occurred during what is now considered one of the worst wildland fire seasons (summer of 2000) for the western US in over 50 years. Mean values for EETOT (4716 kcal day1) and EEACT (2422 kcal day1) for the
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
Helena Hot Shot crew members were statistically similar to the same variables reported by Ruby et al. (2001) (EETOT=4878 and EEACT=2628 kcal day1) using the DLW technique. Clearly, the proposed methodology by this investigation appears promising as a technique for estimating common indicators of cumulative physiological stress (i.e. EEACT and EETOT) in a unique and demanding occupational setting. It should be stressed, however, that this study was a pilot investigation of the proposed methodology and should be verified by a true field validation study where concomitant measures of EE from the same group of fire fighters are derived via electronic activity monitoring and a criterion measure of EE (e.g. DLW, indirect calorimetry, or direct observation). A methodological limitation of the present study was the comparison of EE values in one group of fire fighters using the MTI monitors with EE values from another group of fire fighters using the DLW technique as reported in the literature (Ruby et al., 2001). Given this limitation, the present study chose to determine average EEACT and EETOT values over multiple days of measurement (see Table 1) so that the influence of extremely active or inactive days was minimized. In addition, this methodology also helped to minimize the obvious issue with the present group of fire fighters not necessarily doing exactly the same types or volumes of activities as the fire fighters described by Ruby et al. (1999, 2000, 2001). However, the types of physical activities characterizing wildland fire suppression as described by Ruby et al. (2001) (extensive hiking with a 15–20 kg pack, chainsaw work, brush removal, and fire line construction with a Pulaski, as well as other manual labor specific to actual fire suppression and cleanup activities) are exactly the same as those activities characterizing the Helena Hot Shots. Therefore, by averaging the activity profiles for each fire fighter over multiple days, the present study design provided generalized EEACT and EETOT values that were not representative of any one type of activity common to wildland fire fighting. Indeed, the EE values reported by Ruby and coworkers were similarly derived with the DLW technique as averages over 3–5 successive days of measurement. The EETOT values for Hot Shot fire fighters reported by Ruby et al. (2001) and the present study are similar to those reported for other occupationally (Burstein et al., 1996; Forbes-Ewan et al., 1989; Hoyt et al., 1991) and athletically (Pulfrey and Jones, 1996; Westerterp et al., 1986) demanding activities. For example, an average EETOT of 4634 kcal day1 was estimated for six mountaineers climbing Mt. Shisha Pangma (8046 m elevation) (Pulfrey and Jones, 1996), while an average 8054 kcal day1 were required to ride in the Tour de France bicycle race (Westerterp et al., 1986). Similarly, an Army platoon training for jungle warfare averaged
411
4750 kcal day1 (Forbes-Ewan et al., 1989), while Marines expended 4919 kcal day1 on average during 11 days of cold weather field exercises (Hoyt et al., 1991). All of the above-mentioned studies have two commonalities: (1) Measurement of EETOT in physically active subjects in an extreme environment and/or under physically demanding circumstances; (2) measurement of EETOT by the DLW technique. Use of the DLW technique is certainly the preferred method by researchers when lab-based research is not possible and finances are not limiting. Use of electronic activity monitors, however, as outlined by the present study, may serve as a practical alternative measurement strategy in many of the above study environments. In addition to measures of EETOT and EEACT, it is also possible to derive measures of physical activity behavior from activity monitor data. Fig. 1C, for example, illustrates the continually changing pattern of EE for a single Hot Shot over two successive days. From this figure, several physical activity behavior patterns are apparent. First, the rate of EE rarely exceeded 8 kcal min1 and tended to oscillate between 4 and 6 kcal min1 most frequently. Clearly, experienced wildland fire fighters know that their capacity to perform work must last over 10–16 h each day and will pace themselves accordingly. This may suggest that submaximal work capacity (i.e. as indicated by lactate threshold) may be a better determinant of work capacity in Hot Shots than maximal aerobic capacity (i.e. maximal oxygen uptake). Secondly, physical activity in Fig. 1C tends to oscillate in 15–60 min bouts with distinct breaks between bouts. Again, Hot Shots are well educated in the need to stay hydrated and eat frequently in order to withstand the daily rigors of working the fire line. Lastly, the information collected by the MTI activity monitors is far more detailed than anything is possible using the DLW technique. It is possible, for example, to evaluate each fire fighter’s data day-by-day, hour-to-hour, or even minute-by-minute for upwards of 21 consecutive days. Clearly, this may actually be too much data for most practical needs, but with the use of customized computer programs the volume of data is really not an issue. While not the focus of the present study, the use of electronic monitoring devices to track physical activity behavior may allow researchers in the future to better understand issues related to hydration, energy intake, and energy balance during prolonged bouts of high EE in remote field settings. To the author’s knowledge, the present study is the first to attempt estimating EE with electronic activity monitors located on the chest (i.e. work shirt pocket). With the exception of rare attempts to wear activity monitors on the wrist (Swartz et al., 2000), virtually all other electronic activity monitor validation studies have been performed with the monitor located along the
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
412
Table 2 Comparison of MTI activity monitor output from the hip versus the chest pocket during four activities common to Hot Shot wildland fire fighters a
Physical activity
MTI monitor output at the hip (counts min1)
MTI monitor output in shirt chest pocket (counts min1)
Percent difference in output between hip and chest locations (%)
Self-paced ‘‘slow’’ walking Self-paced ‘‘moderate’’ walking Chopping wood Chainsawing wood
3630 5168 2614 533
3444 5051 3010 617
5.4 2.3 13.2 13.6
a
Percent difference=[1(MTI output at hip/MTI output at chest)] 100.
waist line (Freedson et al., 1998; Hendelman et al., 2000; Swartz et al., 2001). The rationale for the waist location is based upon the assumption that most EE associated with physical activity is best estimated by tracking the motions of the body’s center of mass. The waist location was not an option for the Hot Shot fire fighters, however, because of the waist pack that is worn while working. If the activity monitor was located within the pack there was a chance that the monitor would move around with the contents of the pack and orientation of the monitor in a vertical direction is critical to proper motion detection. Similarly, location of the monitor on the waist strap of the packs was not considered an option because of the likelihood of physical damage or actually losing the monitor completely. Given these constraints, the chest pocket seemed to be a reasonable alternative especially given that Hot Shot fire fighters are well known for their habit of wearing their fire retardant work shirts only while on duty. The main problem with the chest pocket location, of course, was the lack of a validation study to support this methodological characteristic. To address this issue, a sub-study within the present study was performed on three subjects varying considerably in body size (57– 92 kg) and gender (two men and one women) where each subject performed various outdoor activities with an MTI activity monitor around their waist and another within their shirt chest pocket. Each subject wore a 20 kg backpack while performing two self-selected speeds of overground walking, wood chopping with an axe, as well as chainsawing. Given the amount of upper body activities required for wildland fire fighting (i.e. brush removal, fire line construction, etc.), the sub-study was designed to determine if a systematic bias could be expected with the activity monitor located within the chest pocket versus on the waist. The results, which are summarized in Table 2, suggested that an activity monitor located within the chest pocket would slightly underestimate physical activity EE when compared to a waist-mounted monitor during locomotion activities like hiking overground. However, for the activities that emphasized upper body work (wood chopping and chainsawing), the chest pocket monitor tended to overestimate physical activity EE. Based upon these
data it is certain that EE profiles derived for each fire fighter, such as that shown in Fig. 1, represents a combination of under- and overestimates for physical activity EE. The use of an activity monitor designed with a triaxial accelerometer (sensing bodily motion in three planes of motion instead of only one) could provide more accurate information about EE when evaluating specific activity bouts in Hot Shot fire fighters. A recent study by Hendelman et al. (2000), however, found that both uniaxial- and triaxial-based activity monitors have the same motion sensing problems when worn around the waist. Despite the fact that a triaxial accelerometer should provide more detailed information about bodily motion, it is possible that it is not the right kind of information given that EE for any specific activity is dependent upon factors unknown to the activity monitor (e.g. upper versus lower body work, fatigue, load carriage, etc.). In summary, the present study has demonstrated that use of electronic activity monitors to estimate total and activity EE during prolonged bouts of physical activity under environmentally extreme conditions may serve as a practical alternative to more invasive and costly techniques. A field validation study that incorporates concomitant measures of EE from both activity monitors and a criterion method (DLW technique, indirect calorimetry, or direct observation) should be performed prior to full acceptance of the proposed methodology and data processing algorithm. Future use of electronic activity monitors under similar circumstances may focus on physical activity behaviors of fire fighters and how this relates to hydration status and maintenance of energy balance.
Acknowledgements This project was supported in part by a loan of activity monitors from Manufacturing Technology, Inc. (Fort Walton Beach, FL). The author gratefully acknowledges the assistance of Larry Edwards, Supervisor of the Helena Hot Shots, for coordinating the assignment and retrieval of activity monitors to Hot
D.P. Heil / Applied Ergonomics 33 (2002) 405–413
Shots on the fire lines, as well as the willingness of the Helena Hot Shots themselves to participate in this study.
References Ainsworth, B.E., Haskell, W.L., Whitt, M.C., Irwin, M.L., Swartz, A.M., Strath, S.J., O’Brien, W.L., Bassett, D.R., Schmitz, K.H., Emplaincourt, P.O., Jacobs, D.R., Leon, A.S., 2000. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exercise 32 (9), S498–S516. American College of Sports Medicine, 2000. ACSM’s Guidelines for Exercise Testing and Prescription, 6th Edition. Lippincott Williams & Wilkens, Baltimore, MD. Brotherhood, J.R., Budd, G.M., Hendrie, A.L., Jeffery, S.E., Beasley, F.A., Costin, B.P., Zhien, W., Baker, M.M., Cheney, N.P., Dawson, M.P., 1997. Project Aquarius 3. Effects of work rate on the productivity, energy expenditure, and physiological responses of men building fireline with a rakehoe in dry eucalypt forest. Int. J. Wildland Fire 7 (2), 87–98. Budd, G.M., Brotherhood, J.R., Hendrie, A.L., Jeffery, S.E., Beasley, F.A., Costin, B.P., Zhien, W., Baker, M.M., Cheney, N.P., Dawson, M.P., 1997. Project Aquarius 5. Activity distribution, energy expenditure, and productivity of men free-running wildland fires with hand tools. Int. J. Wildland Fire 7 (2), 105–118. Burstein, R., Coward, A.W., Askew, W.E., Carmel, K., Irving, C., Shpilberg, O., Moran, D., Pikarsky, A., Ginot, G., Sawye, M., Golan, R., Epstein, Y., 1996. Energy expenditure variations in soldiers performing military activities under cold and hot climate conditions. Mil. Med. 161 (12), 750–754. Butler, B.W., Bartlette, R.A., Bradshaw, L.S., Cohen, J.D., Andrews, P.L., Putnam, T., Mangan, R.J., 1998. Fire Behavior Associated with the South Canyon Fire on Storm King Mountain, Colorado (Research Paper RMRS-RP-9). USDA Forest Service, Rocky Mountain Research Station: Fort Collins, CO. Forbes-Ewan, C.H., Morrissey, B.L.L., Gregg, G.C., Waters, D.R., 1989. Use of doubly labeled water technique in soldiers training for jungle warfare. J. Appl. Physiol. 67 (1), 14–18.
413
Freedson, P.S., Melanson, E., Sirard, J., 1998. Calibration of the computer science and applications, Inc accelerometer. Med. Sci. Sports Exercise 30 (5), 777–781. Hendelman, D., Miller, K., Baggett, C., Debold, E., Freedson, P., 2000. Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Med. Sci. Sports Exercise 32 (9), S442–S449. Hoyt, R.W., Jones, T.E., Stein, T.P., McAninch, G.W., Lieberman, H.R., Askew, E.W., Cymerman, A., 1991. Doubly labeled water measurement of human energy expenditure during strenuous exercise. J. Appl. Physiol. 71 (1), 16–22. IAEA, 1990. In: Prentice, A. M. (Ed). The Doubly Labelled Water Method for Measuring Energy Expenditure: a Consensus Report by the IDECG Working Group. I/D/E/C/G International Dietary Energy Consultancy Group, NAHRES-4, IAEA, Vienna, Italy. National Fire News, 2001. Fire Season 2000 Statistics. National Interagency Fire Center, Boise, ID. Pulfrey, S.M., Jones, P.J.H., 1996. Energy expenditure and requirement while climbing above 6000 m. J. Appl. Physiol. 81 (3), 1306– 1311. Ruby, B.C., Zderic, T.W., Burks, C.A., Tysk, S., Sharkey, B.J., 1999. Total energy expenditure (doubly labeled water) and bone resorption during wildland fire suppression. Med. Sci. Sports Exercise 31 (5), S1863. Ruby, B.C., Schoeller, D.A., Sharkey, B.J., 2000. Gender differences in water turnover, energy balance and body composition during arduous work. Med. Sci. Sports Exercise 32 (5), S17. Ruby, B.C., Schoeller, D.A., Sharkey, B.J., 2001. Evaluation of total energy expenditure (doubly labeled water) across different measurement periods during arduous work. Med. Sci. Sports Exercise 33 (5), S1545. Swartz, A.M., Strath, S.J., Bassett, D.R., O’Brien, W.L., King, G.A., Ainsworth, B.E., 2000. Estimation of energy expenditure using MTI accelerometers at hip and wrist sites. Med. Sci. Sports Exercise 32 (9), S450–S456. Westerterp, K.R., Saris, W.H.M., van Es, M., ten Hoor, F., 1986. Use of the doubly labeled water technique in humans during heavy sustained exercise. J. Appl. Physiol. 61 (6), 2162–2167.