Cumulative Spine Loading - Fundamentals and Assessment Tools for Occupational Ergonomics - page 317

static load magnitude dependent on lifting speed and load magnitude (Danz and Ayoub, 1992). None of
the published cumulative loading studies have used a dynamic modeling approach. Another data
reduction method to reduce the quantity of data precludes dynamic analyses and will be discussed
shortly. Several studies appear to have used quasi-dynamic approaches with representative peak measures
taken of hand loads (Daynard et al., 2001; Godin et al., 2003; Norman et al., 1998) with all other studies
using static models (Jager et al., 2000; Kumar, 1990) and recall of static loads (Seidler et al., 2001, 2003).
The use of a static assumption in cumulative analyses does not induce the same magnitude of errors seen
in peak analyses. Cumulative exposure average errors across varying tasks, masses and lifting speeds were
approximately 10% compared to a dynamic model (Callaghan et al., 2005). The quasi-dynamic model
faired much better with average errors well below 5% compared to a dynamic model. These
lower errors in cumulative loading compared to peak loading can be attributed to the changing accelera-
tion of hand loads when being lifted and lowered. When being lifted the hand load will exceed
the actual static load of the object. But when being lowered the effective hand load will be less than
the static load of the object thereby letting over- and underprediction errors cancel out when dynamic
components are not included. This study used a model with sampled hand loads so that at any
instant in time the hand load was known. This is impractical in an industrial application and the
studies that have employed quasi-dynamic models have utilized a representative hand load for the
entire lifting period or fraction of the task when the load is in contact with the body. This is effectively
a static or nonchanging load and may prove to be no better than simply using the mass of the objects
lifted or manipulated.
13.8.3 Extrapolation of Short-Time Periods to Represent a Shift Exposure
Determining the cumulative loading of an entire shift involves documenting the loading in all of the tasks
a worker performs. To document the task cumulative exposure involves calculating the force-time or
moment-time histories that are exerted on the lumbar joint at a sufficient sampling rate to capture an
accurate representation of exposure. Rather than having to measure the entire shift or exposure
period, would it be sufficient to document the components of a job, the number of times each was per-
formed and then extrapolate to predict a shift exposure? This approach has not been directly validated
but has been widely employed by most cumulative studies with the exception of Jager et al. (2000), who
undertook the commendable task of documenting with great detail the loading exposure of eight indi-
viduals for entire shifts. Secondary evidence that this approach may be acceptable is found in research
that has examined the variability of how workers perform their jobs. Granata et al. (1999) demonstrated
that peak compressive loads in repetitive tasks have a fair amount of variability between trials.
By taking three repeats of a task for a worker, the standard errors of the trunk kinematics can be
reduced (Allread et al., 2000). This led to some of our own work examining the repeatability of trial
to trial cumulative load (Keown et al., 2002). If a single cycle is used to extrapolate to shift exposure,
the potential to substantially under- or overestimate actual cumulative loading is high. Keown et al.
(2002) found that by averaging 3 to 6 cycles a more stable estimate of cycle loading was achieved,
which would give a better representation of shift exposure when extrapolation techniques are used.
Additionally, the time of day that the samples were taken at or the day that they were taken on
had little influence on the average cycle cumulative loading (Keown et al., 2002).
13.8.4 Reducing the Information to Represent Task Exposure
If it is possible to reduce the documentation time down to several task repeats and then extrapolate to a
shift, would it be equally feasible to reduce the information content required to determine cycle cumu-
lative exposure? Several different approaches have been employed to reduce the data required to represent
task-based exposures. Many of the different approaches to represent cycle cumulative loading were
examined to determine the error that the various assumptions introduced (Callaghan et al., 2001).