**The 15-point Borg scale 6**

**3. RESULTS AND DISCUSSION**

**3.18 A commentary on cardiorespiratory endurance standards**

**Figure 85: Steady-state oxygen consumption during walking and stepping**
(N=20; ten males and ten females), with and without the personal protective
clothing and equipment (Taylor et al., 2012b). Points are values for different
individuals in absolute units (A and B), and then normalised to both body
mass, and the body plus the protective clothing and equipment masses (total
mass: specific oxygen consumption: C and D). Finally, data are presented as
residuals, with specific oxygen consumption values individually subtracted
from the group mean specific oxygen consumption (E and F).

In Figures 85C and 85D, the impact of the two normalisation strategies is immediately evident. Firstly, normalising for total mass elicits a lower specific oxygen consumption.

This is an entirely predictable consequence of changing any denominator. However, if one was to base a cardiorespiratory endurance standard upon data so obtained, then one must determine which normalisation strategy is more appropriate, for the resultant standard is inextricably linked with that choice.

Secondly, as others have demonstrated for maximal aerobic power (Taylor et al., 1981;

Schmidt-Nielsen, 1984; Åstrand and Rodahl, 1986; Nevill et al., 1992), there is a negative
relationship between specific oxygen consumption and body mass during each of these
controlled, steady-state exercise modes (black symbols of Figures 85C and 85D). Although
not strong (walking: r^{2}=0.06; bench stepping: r^{2}=0.13), this trend is evident, and it results
in an over-correction of data for individuals of greater body mass, and a higher specific
oxygen consumption for lighter subjects. That is, when one compares Figure 85A with 85C,
and 85B with 85D, this normalisation method has changed the relationship between

metabolic demand and mass from a positive (85A and 85B) to a negative slope (85C and 85D). However, when employment standards are being developed, one should endeavour to remove bias. To achieve this, normalisation should convert these slopes into flat, mass-independent relationships. Thus, when standards are developed for employment categories in which loads are carried on the body, as is the case for fire fighting, the possibility exists that the minimal standard may suffer from a mass bias if data are normalised to the body mass only, and also if the standard is developed using individuals drawn from a range of body sizes that inadequately represent that of the sub-population from which recruits may be drawn.

Thirdly, for each exercise mode, normalising to the total mass (body plus clothing and equipment) resulted in a flattening of these specific oxygen consumption to mass

relationships (Figures 85C and 85D). This was also predictable, for it is well established that the metabolic impact of a constant load carriage is greater on smaller people (Taylor et al., 1980), changing in direct proportion to the change in the specific load. Thus, a 5%

increase in the total load is accompanied by a 5% elevation in metabolic demand. For lighter individuals, the combined mass of the protective clothing and equipment represented a greater relative mass change, and therefore a greater metabolic burden, and the impact of the mass-dependence of load carriage is evident. However, when the absolute oxygen cost was normalised to the total mass, this impact was partially removed, and the corresponding relationships with mass levelled off (red symbols of Figures 85C and 85D). In this case, it appears as though this normalisation procedure minimised the mass bias.

If one now compares these different relationships within Figures 85C and 85D, it becomes apparent that the regression lines for the two normalisation methods converge on a

theoretical body mass of 145-150 kg for each exercise mode. At this point, the protective clothing and equipment mass (~20 kg) would be less than 3% of the body mass, and since this is within the resolution of the measurement equipment, it would not be detectable. One can extend these analyses to the current simulations, in which firefighters worked whilst wearing protective clothing and equipment. In this case, ambulatory simulations were chosen to represent walking on a flat surface (Figure 86A: hydrant simulation), moving up

an incline (Figure 86B: carrying the ventilation fan up stairs) and a vertical climb (Figure 86C: ladder climb). This choice allowed for an evaluation of these normalisation procedures across the broadest possible range of mass-dependent locomotion. Whilst these data are inherently noisy, as noted above, mass-dependence is present once more, and a similar converging trend exists, with the regression lines again coming together at a theoretical body mass of 140-150 kg.

**Figure 86: Specific oxygen consumption data for firefighters performing a**
hydrant simulation (A; Simulation 5: N=16), carrying a ventilation fan (35
kg) up stairs (B; Simulation 14: N=16) and a vertical ladder climb (C;

Simulation 13: N=14). Each simulation was performed whilst wearing full personal protective clothing and equipment (~20 kg).

The next analysis of these data provides a statistical justification for choosing one normalisation procedure over another. In Figures 85E and 85F, data are presented as

residuals, with the specific oxygen consumption for each individual subtracted from the group mean for these steady-state walking and stepping activities (respectively). The mean specific oxygen consumption is important, since it would be used for setting an employment standard, if the activities in question were criterion tasks. Both Figures reveal the same trend. That is, the residuals are smaller when these metabolic data were normalised to the total mass, and these differences were statistically significant (P<0.05). Therefore, a

statistically superior employment standard should result from the normalisation to total mass when load carriage on the body forms an integral characteristic of the working conditions.

To this point, we have only been considering linear (arithmetic) normalisation procedures.

Yet we know that such an approach is frequently inappropriate (Section 3.1.3), since a one-to-one relationship between oxygen consumption and body mass does not exist (Kleiber, 1932; Tanner, 1949; Taylor et al., 1981; Schmidt-Nielsen, 1984; Åstrand and Rodahl, 1986; Nevill et al., 1992). Thus, curvilinear normalisation appears to be more valid for circumstances in which the oxygen cost of an activity is mass dependent. A more complete discussion on this point is contained within Appendix Two.

Finally, we have seen from Table 51 that none of the simulations investigated within the current project could be classed as unloaded, cardiorespiratory endurance activities. Indeed, every activity involved firefighters wearing protective clothing and equipment in some form. Thus, when load carriage is an important occupational constraint, then one must evaluate physiological function under loaded situations (Vanderburgh and Flanagan, 2000;

Bilzon et al., 2001a; Vanderburgh, 2008; Vanderburgh et al., 2011). The shuttle-run test fails to meet this criterion, and is likely to provide an unreliable prediction of load-carriage performance (Bilzon et al., 2001a).

One may summarise this commentary as follows:

• Standards derived for load-carriage occupations that do not normalise data to the total mass of the participants, and their protective clothing and equipment masses, are artificially inflated.

• Where possible, mass bias needs to be removed from employment standards, and this cannot occur unless the total mass is appropriately considered.

• Since linear normalisation is fallacious, then nonlinear approaches need to be thoroughly investigated from an occupational perspective (such data have been incorporated into existing data summary Tables [Table 15 onwards]).

• Unloaded endurance tests are unreliable screening methods for occupations in which load carriage is an integral part of the working requirement.

If one accepts these points, then one must also arrive at two conclusions. Firstly, the
minimal cardiorespiratory endurance standard of 45 mL.kg^{-1}.min^{-1} may be artificially
inflated, due to an artefact arising during its derivation. Secondly, considering the
requirement for firefighters to perform predominantly loaded activities, the use of an
unloaded endurance test to predict the ability of recruits to meet this standard is now found
to be lacking in scientific support.