Assessing Postures and Mechanical Loads during Patient Transfers
Sandra Hellmers
1 a
, Anna Brinkmann
1 b
, Conrad Fifelski-Von B
¨
ohlen
1
, Sandra Lau
2
,
Rebecca Diekmann
1
and Andreas Hein
1
1
Assistance Systems and Medical Device Technology, Carl von Ossietzky University Oldenburg, Oldenburg, Germany
2
Geriatric Medicine, Carl von Ossietzky University Oldenburg, Oldenburg, Germany
Keywords:
Motion Capture, IMU, EMG, Care, Nursing, Musculoskeletal, Technology, Wearable, Posture, Kinect.
Abstract:
Socio-Demographic developments in industrialized countries cause a discrepancy between potential recipients
and providers of care. Caregivers experience high musculoskeletal loads during their daily work, which leads
to back complaints and a high rate of absenteeism at work. Ergonomically correct working can significantly
reduce musculoskeletal load. In a study with 13 caregiver students, we analyzed body postures, muscle activ-
ities, and loads during the transfer of a patient from bed to wheelchair. Our measurement system consists of a
full-body motion capture system and a Multi Kinect System. Additionally, muscle activities were measured via
surface electromyography. According to recommendations for ergonomic working in the care sector, a system
was developed that recognizes potentially harmful postures based on the motion capture data. A result report
visualizes the skeleton model together with color-coded information about inclination and torsion angles. The
motion capture data was also related to EMG data and analyzed according to biomechanical assumptions.
1 INTRODUCTION
Socio-demographic developments of industrialized
countries are characterized by low birth rates and an
increasing life expectancy. Therefore, the number of
people reaching old age increases. The discrepancy
between the supply of caregivers and the demand for
caregivers will continuously grow. In an international
comparison, Germany already has the worst patient-
to-caregiver ratio in Europe, with measurable effects
on patient mortality rates and the stress experience of
caregivers (H
¨
ohmann et al., 2016; Aiken et al., 2012).
Manual patient handling is one of the physiological
risk factors and leads to high mechanical loads in the
lower back of caregivers (Kliner et al., 2017; Hwang
et al., 2019; Choi and Brings, 2016; J
¨
ager et al.,
2013). Particularly non-ergonomic movements and
postures lead to health problems. This results from
various strenuous activities such as deep bending or
twisting during manual patient transfer, e.g. transfer-
ring a patient from bed to wheelchair (Hwang et al.,
2019; Choi and Brings, 2016; J
¨
ager et al., 2013). But
also working in harmful postures leads to back com-
plaints. Caregivers spend a total of two hours per
a
https://orcid.org/0000-0002-1686-6752
b
https://orcid.org/0000-0001-5228-4947
shift in a bent posture or bend down 1500 times per
shift (Weißert-Horn et al., 2014). High musculoskele-
tal strains and related spinal complaints are one of
the main reasons for the high rate of absenteeism at
work as well as for leaving the profession in profes-
sional nursing (J
¨
ager et al., 2013; Weißert-Horn et al.,
2014). It is therefore important to relieve and sup-
port caregivers. Musculoskeletal stress can be signif-
icantly reduced by an ergonomically correct method
of caregiving (Hwang et al., 2019; Choi and Brings,
2016). Also, prevention programs including various
ergonomic measures can improve the well-being of
the back and reduce the physical strain on caregivers
(Michaelis and Hermann, 2010).
To prevent back complaints, it is therefore important
to train caregivers to work in a back-friendly way and
to avoid harmful postures and actions. Therefore, we
developed a system that can be used for the training
of caregivers and the detection of harmful postures.
2 STATE OF THE ART
Manual movement of persons in need of care leads
to high mechanical stress in the lower back area of
caregivers (Weißert-Horn et al., 2014). Unfavorably
long-lasting trunk postures as well as lifting, holding
Hellmers, S., Brinkmann, A., Böhlen, C., Lau, S., Diekmann, R. and Hein, A.
Assessing Postures and Mechanical Loads during Patient Transfers.
DOI: 10.5220/0010155300210029
In Proceedings of the 14th International Joint Conference on Biomedical Engineer ing Systems and Technologies (BIOSTEC 2021) - Volume 5: HEALTHINF, pages 21-29
ISBN: 978-989-758-490-9
Copyright
c
2021 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
21
Table 1: Classification of the upper body (UB) postures
into angle ranges (Freitag et al., 2007). Sagittal inclinations
above 60
are classified as critical. Lateral incliniations and
torsions above |20|
are also harmful.
UB limited un-
posture acceptable acceptable acceptable
sagittal 0
-20
20
-60
>60
lateral 0
-|20|
- >|20|
torsion 0
-|20|
- >|20|
and pulling large parts of the patient’s weight without
the use of aids such as lifts or sliding mats lead to a
compression of the intervertebral discs of the carers
up to 9 kN during this activity. This is far above the
upper limit of 3.4 kN for lumbosacral-disc compres-
sive forces (J
¨
ager et al., 2013).
Freitag et al. reported that nurses working on a geri-
atric ward take an average of 1390 upper body incli-
nations above 20 degrees per work shift (Freitag et al.,
2007). Weißert-Horn et al. found similar results and
state that caregivers spend a total of two hours per
shift in a bent posture or bend down 1500 times per
shift (Weißert-Horn et al., 2014). Recurring loads can
cause pain and injuries in the back and upper extrem-
ities. It is therefore important to identify and avoid
harmful postures.
Movement and posture analysis systems based on
different types of sensors can be used to identify
harmful postures and to estimate the stresses and
strains that occur in the musculoskeletal system
of both the carer and the person being cared for.
Theilmeier et al. captured postures via an optoelec-
tronic motion capturing system and video cameras
(Theilmeier et al., 2010). Additionally, a care bed
was equipped with a framework attached to the
bedstead and connected to the bedspring frame
via force sensors at the four bed-corners. J
¨
ager et
al. analyzed nine different activities with a patient
transfer in or at the bed or chair with the same
system. They found that the load on the lumbar spine
can be reduced through biomechanically optimized
transfer instead of conventional methods (J
¨
ager et al.,
2013). Wei et al. described a system that uses data
from depth-image cameras to estimate the skeletal
poses and joint forces of wheelchair users during
transfers (Wei et al., 2018). A Microsoft Kinect V2
camera was used to record skeleton data during three
different care activities (Agrawal and Ertel, 2018).
Lin et al. used both inertial sensor technology and a
marker-based motion detection system on the patient
to evaluate the transfer performance by the nursing
staff (Lin et al., 2018). However, the possibilities of
ergonomic work design and assessment have hardly
been used systematically in nursing and health care
up to now (Ding et al., 2014). This points to the
necessary need for ergonomic work teaching systems
for caregivers.
One aspect of working in a back-friendly manner
is avoiding excessive upper body inclination angles
and working with a straight back. Thereby, the
movements of the upper body should be considered
in all three directions: Movements in the sagittal
direction mean the forward inclination of the trunk.
The lateral movement - away from the central axis -
is called a lateral inclination. The torsion of the upper
body between the thoracic spine and the lumbar
spine is defined as torsion. Freitag et al. made a
classification of the upper body postures into angle
ranges and their degree of risk (Freitag et al., 2007).
The classification is listed in Table 1. Here angles
in the range of 0
-20
belong to the neutral position.
But according to the DIN standard DIN EN 1005-4
(German Institute for Standardization, 2005) lateral
inclinations and torsions above ten degrees are also
classified as critical.
In the literature are many recommendations for
back-friendly working in patient care (Michaelis
and Hermann, 2010), especially from professional
associations (Baum et al., 2012; Kusma et al., 2015).
In the present work, we focused on the transfer
from the bed edge to a wheelchair. Additionally to
our ”Healthcare Prevention System” (Brinkmann
et al., 2020), we developed a posture recognition
system based on a motion capture suit to identify
potentially harmful postures considering the rec-
ommendations for ergonomic working in the care
sector.
3 MATERIALS AND METHODS
In the following, the study design, the biomechanics,
and the used sensors and methods are described.
3.1 Study Design
To analyze typical care activities and their ergonomic
execution, we conducted a study with 13 caregiver
students, aged between 18 and 55 years (10 women,
3 men). In this article, we analyzed the transfer from
the edge of a bed to a wheelchair. The transfer was
carried out by each of the 13 students. Additionally
to the sensors, a Kinaesthetics teacher observed and
evaluated the performance of the care activity under
consideration of ergonomic working methods and ki-
naesthetic aspects. A member of the research staff
HEALTHINF 2021 - 14th International Conference on Health Informatics
22
acted as a patient so that no real patients were in-
volved. To avoid overloadings, the patient cooperates
during the treatment.
The test procedures were approved by the local ethics
committee (ethical vote: Carl von Ossietzky Univer-
sity Oldenburg and conducted in accordance with the
Declaration of Helsinki.
3.2 Applied Sensor Systems
Several sensor systems were used in the study and are
described in the following.
3.2.1 Motion Capture System
The caregiver students were equipped with a full-
body motion capture system of Motion Shadow
(Shadow, 2020). The system includes 17 motion
nodes. Figure 1 shows the human joints (blue
dots) and the positions of each motion node (green
boxes) of the motion capture system. A motion
node is a measurement unit, which includes a 3D-
accelerometer, 3D-gyroscope, and 3D-magnetometer.
The sampling rate of each sensor is 100 Hz.
3.2.2 Multi Kinect System
In addition, the nursing process was recorded with
a Multi Kinect System (Fifelski et al., 2018). This
system contains 4 Microsoft Kinect v2 depth cam-
eras. Each Kinect v2 is connected to an Intel NUC
Mini-PC, which acquires the data from the camera
and sends it to the main computer, where the data of
the four cameras is fused into a colored point cloud.
The depth cameras have to be registered to each other
Figure 1: Calculation of sagittal inclination angle: The total
angle α results from the segment angles SpineLow, Spine-
Mid and Chest.
Figure 2: Anatomy of human thigh. Electromyography
electrodes are on rectus femoris, vastus medialis and biceps
femoris. The dots on the muscles indicate the positions of
the electrodes. Image courtesy of Complete Anatomy (Es-
sential Anatomy 5) (3D4Medical, 2020).
to ensure that the four point clouds of cameras are
aligned. The Multi Kinect System enables a 3D view.
3.2.3 Electromyography
Surface Electromyography (EMG) was also used to
record the muscular activities associated with mus-
cle contraction in order to gain information on the
activation behavior of selected muscle groups during
the transfer from bed to wheelchair: vastus medialis
(VM), rectus femoris (RF), biceps femoris (BF) (see
Figure 2). Preparation, collection and processing pro-
tocols were consistent with SENIAM guidelines (Her-
mens et al., 1999). Signals of the bipolar surface elec-
trodes (14 mm diameter and 10 mm inter-electrode
distance, GE Medical/Hellige) were amplified with
2500 Hz by local amplifiers, then filtered (bandpass
10–700 Hz) and sent to Biovision Inputbox. The es-
timation of muscle activity in the caregiving process
was based on the potential level and set in relation
to the chronological sequence of the caregiving ac-
tivities. The processing of the raw EMG signals was
done via rectification, Root Mean Square (RMS) and
mean value.
3.3 Posture Analysis System
Figure 3 shows the system-workflow for processing
the IMU data of the motion capture system and au-
tomated analyzing and reporting risky postures dur-
ing care treatments. The IMU data is exported in
bvh-format and integrated into our system. The body
model defines an object-oriented data structure for
the motion data according to the recommendation of
the International Society of Biomechanics (Wu et al.,
2002). The model realizes access to the data of body
sections such as joints and realizes data related op-
erations. The identification component reads the raw
Assessing Postures and Mechanical Loads during Patient Transfers
23
Figure 3: System-Workflow for processing the IMU data of the motion capture system and analyzing and reporting physical
loading during care treatments.
data streams and converts them into the structure of
the defined data model. The identification component
returns the data object with the joint positions and an-
gles. The analysis component uses the data object
for motion analysis and executes the analysis strategy.
We implemented an algorithm for the analysis of risky
postures during care activities based on the physical
posture model. The physical posture model contains
various threshold values, which indicate harmful pos-
tures. We implemented rules for the identification of
risky postures regarding the classification of (Freitag
et al., 2007) mentioned in Table 1.
Joint positions are measured based on the neutral-
zero method. According to the joints’ neutral position
within the different body planes, a predefined perpen-
dicular is marked to measure the range of movement
(ROM) in degree. This procedure is most common
for peripheral joints of the lower and upper limbs. As-
sessing the ROM of the spine is more complex since
each vertebra is naturally located slightly different to
another. ROM increases with every involved segment.
Thus, the measurement of the spinal inclination
angle is challenging - especially in different postures.
In this approach, we decided to analyze the angle by
dividing the spine into three segments. The segments
in the skeleton model are called SpineLow, SpineMid,
and Chest. The upright neutral-zero position in the
sagittal plane is taken as the perpendicular which is in
this case identical to the longitudinal axis. To cal-
culate the sagittal inclination angle during transfer,
we shifted the perpendicular horizontally to L5, the
fifth lumbar spine vertebra. A straight line between
L5 and C7 (which covers all relevant segments) in-
dicates the angle corresponding to the perpendicular,
whereby C7 is the seventh cervical vertebra. The sys-
tem’s calculation of the sagittal inclination angle is
illustrated in Figure 1. To estimate the angle between
L5 and C7, the angles of segments SpineLow, Spine-
Mid and Chest were considered to calculate the total
inclination angle. Lateral inclination angles, as well
as torsions, are calculated accordingly.
In summary, the analyzer executes the analysis algo-
rithms and returns the result object. We implemented
a rule-based algorithm in the analyzer component,
which monitors the upper body inclination and tor-
sion and the compliance of the threshold values of the
physical posture model. If a threshold value is ex-
ceeded, the posture is classified as potentially harm-
ful. The report (result object) is a 3D virtual ob-
ject, which shows the visualization of skeleton model
(similar to the skeleton models in Figure 4) during
the execution of the care activity and indicates harm-
ful posture by specifying which of the defined rules
were exceeded.
3.4 Manual Patient Handling and
Biomechanics
The transfer from the edge of a bed to a wheelchair
can be divided into three main phases: Stand the pa-
tient up, turn the patient, and sit the patient down into
a wheelchair. In the following, the three main phases
and their subphases will be examined concerning
their ergonomic execution and biomechanics. Figure
4 shows the phases exemplarily for one student. In
the top of the Figure, 3D camera images of the Multi
Kinect System of each of the subphases are shown. In
the bottom of the Figure, the corresponding skeleton
models measured by the motion capture system are
visualized. At the beginning of the transfer, the
person requiring care sits at the edge of the bed. The
caregiver positions himself frontally to the patient
(stand). In the preparation phase, the caregiver gives
instructions to the patient and bends down to the
patient while squatting. The upper body remains
as straight as possible. The caregiver puts his arms
around the patient. The patient is also asked to put
his arms around the caregiver. In the lift phase,
the patient is lifted into a standing position by the
caregiver. At the end of the lift, the patient and the
caregiver stand in an upright position.
In the second main phase the caregiver rotates to-
gether with the patient in small cradle steps until the
patient is standing directly in front of the wheelchair.
The rotation in small steps prevents harmful upper-
HEALTHINF 2021 - 14th International Conference on Health Informatics
24
Figure 4: In the first phase of the transfer from the edge of the bed to a wheelchair, the caregiver lifts the patient into a standing
position. In the second phase the caregiver turns the patient in small steps until the patient is standing directly in front of the
wheelchair. In the third phase the caregiver lowers the patient into the wheelchair.
body torsions.
In the third phase of the transfer the caregiver
carefully lowers the patient into the wheelchair. The
transfer ends with an upright standing position.
In a biomechanical view, manual patient handling
leads to a compression, flexion and torsion of the
caregiver’s intervertebral discs depending on the
patient’s weight and the executed transfer mode
(J
¨
ager et al., 2013). Therefore, the intervertebral
discs are affected vertically by different compressive
forces. These forces are almost parallel to the flat
geometry of the sacroiliac joints (SIJ) (see Figure 2),
which are the direct connection between pelvis and
sacrum and primarily responsible for load transmis-
sion to the hip joints and finally to the legs and vice
versa (Vleeming and Stoeckart, 2007). Additional
muscle groups are essential in order to transfer loads
isolated and effectively from the lumbar spine to
the pelvis and to compensate for potential overload
effects (Vleeming and Stoeckart, 2007). Effective
load transfer is achieved, when muscle forces cause
compression of the SIJ, and thereby, preventing
shearing of the joints (Richardson et al., 2002; van
Wingerden et al., 2004). Both the muscles of the
lower limb and the back extensors influence the SIJ
movements and its stabilization mechanisms via
lever arms (Vleeming and Stoeckart, 2007). In the
present study, the muscles of lower limb are analyzed.
During hip flexion, torsional forces are transmitted to
the SIJ due to the connection of the rectus femoris
(RF) to the pelvis (Hammer et al., 2015) (see Figure
2). The rectus femoris (RF) is one of the extensors
of the knee and the flexors of the hip (Garrett and
Kirkendall, 2000). The vastus medialis (VM) is also
involved in knee extension (Mansfield and Neumann,
2008). It is assumed that increased activity of the RF
can lead to pain in the SIJ, and therefore may cause
lower back pain (Hammer et al., 2015). Furthermore,
due to the connection to the pelvis, the biceps femoris
(BF) performs knee flexion and hip extension and
thus affects the SIJ.
4 RESULTS
4.1 Analysis of Inclination Angles
After focusing on the different phases of the transfer,
the sagittal and lateral upper body inclination as well
as the upper body torsion were examined. Figure
5 shows the inclination and torsion angles during
the transfer. The inclination angles were calculated
according to Figure 1. The main phases are marked
in the graph.
The transfer itself took about 30 seconds. As ex-
pected, the highest inclination angles occur in the
first and last phase: stand up the patient and sit down
the patient. The first phase begins with a stand in a
neutral position with inclination angles below 20
.
The caregiver is only slightly inclined to the patient.
In the preparation phase, the caregiver squats. The
upper body also inclines in the sagittal direction.
In this preparation phase as well as during the lift,
angles of up to 40 degrees are reached, so that the
Assessing Postures and Mechanical Loads during Patient Transfers
25
Figure 5: Sagittal and lateral upper body inclination as well
as upper body torsion during the transfer from the edge of a
bed to a wheelchair.
limit to the neutral range (<20
) is exceeded. The
range between 20
-60
is limited acceptable (cf.
Table 1). Therefore, the time spent in this potentially
harmful position should be as short as possible. Ad-
ditional loads acting on the body or the back in such a
posture are particularly associated with a risk of back
injuries. During lift phase, part of the patient’s weight
acts on the body. The stand the patient up phase
ends in an upright position. The upright position is
maintained during the rotation so that the sagittal
inclination angle remains in the neutral range during
the entire phase. The turning phase has three peaks.
The number of peaks indicates the number of double
steps executed during the rotation.
In the last phase, the caregiver bents again to lower
the patient into the wheelchair. Here, angles signifi-
cantly above 20
are achieved, which indicates again
a potentially harmful posture. After setting the patient
down, the caregiver returns to an upright position. As
already described, the system generates a result report
that shows the corresponding color-coded angles as
well as the visualization of the skeleton model during
the care activity. If the threshold values are exceeded,
the corresponding angle is marked in yellow (sagittal
inclination angle between 20
-60
) or red (sagittal
inclination angle >60
) and the posture is classified
as potentially harmful.
The lateral inclination remains in an acceptable posi-
tion during the entire care activity. In the preparation
phase, the student takes a slightly right-bent posture
(positive inclination angles). During the lift and the
patient’s descent, the student shows a slight inclina-
tion to the left (negative inclination angles). During
the rotation, the angle is almost zero so that the stu-
dent has a straight upright position.
Figure 6: Mean muscle activity of vastus medialis (VM),
rectus femoris (RF), and biceps femoris (BF) during the
phases of the transfer.
A short torsion of more than 20
can be seen at the
beginning of the recording and thus before the actual
care action. During the care procedure itself, the stu-
dent remains in an acceptable position. Based on this
diagram, a distinction can also be made between the
preparation phase and the lift. During the preparation
phase, the student is squatting and is inclined in the
sagittal direction. During the lift, the direction of tor-
sion and lateral inclination changes. Special attention
should be paid for coupled movements (torsion + in-
clination, sagittal + lateral inclination) since they can
be particularly harmful (Panjabi and White, 1990).
4.2 Analysis of Muscle Activities
To evaluate the specific caregiving phases, the ac-
quired data of the applied sensor systems are related
to the surface EMG data and analyzed and interpreted
according to biomechanical assumptions. Figure 6
shows the mean muscle activity of rectus femoris
(RF), vastus medialis (VM), and biceps femoris (BF)
during the phases of the transfer. The mean muscle
activity of RF, VM and BF in the squatting position
are 322 mV, 175 mV and 77 mV. While lifting the pa-
tient the muscle activities are 425 mV, 212 mV and
141 mV. Turning the patient leads to mean muscle ac-
tivities of 121 mV, 173 mV and 88 mV and while sit-
ting the patient down the mean muscle activities are
204 mV, 142 mV and 192 mV.
5 DISCUSSION
Caregivers experience high musculoskeletal loads
during their daily work, which leads to back com-
plaints and a high rate of absenteeism at work. It
HEALTHINF 2021 - 14th International Conference on Health Informatics
26
has been shown, that ergonomically correct working
can lead to a significant reduction in musculoskele-
tal load (Brinkmann et al., 2020; Weißert-Horn et al.,
2014). However, the possibilities of ergonomic work
design and assessment have hardly been used system-
atically in nursing and health care up to now (Ding
et al., 2014). In consequence, we developed a posture
recognition system which is able to identify potential
harmful body postures according to the recommen-
dations for ergonomic working in the care sector and
integrated it into our Healthcare Prevention System
(Brinkmann et al., 2020). The posture recognition
system analyzes the data of a full-body motion cap-
ture system. A result report visualizes the skeleton
model together with color-coded information about
inclination and torsion angles during the care activ-
ity.
In the present work, we focused on the transfer from
the bed into a wheelchair and analyzed the posture
in terms of sagittal and lateral inclination angles as
well as torsions. We could show the concept of the
system and see its application mainly in the training
of nursing students. An advantage of this system is
the report and its visualization of the nursing handling
which allows a retrospective discussion of the activ-
ity with objective angle information. In addition, the
visualization of the skeleton model can be rotated, so
that the nursing activity can be evaluated from differ-
ent perspectives. The advantage of using a full-body
motion capture system over cameras is that the field
of vision cannot be obscured and the system does not
require additional space. The disadvantages are that
the patient is not recorded and therefore context in-
formation may be missing.
The Multi Kinect System (Fifelski et al., 2018;
Brinkmann et al., 2020) was used to record the scene
by depth camera data. This system also allows a 3D
view and therefore a viewing from different perspec-
tives. We are currently working on merging skeletal
models from the Kinect data to make posture analysis
with this data as well. So the camera system and the
motion capture system can be used interchangeably
depending on requirements.
As it is not possible to determine loads on the basis
of pure body postures, further information is needed
to assess the risk potential for a care activity. Con-
textual information about the nursing action itself and
the weight of the patient can be used to make rough
estimations about mechanical loads, which act on
the caregiver’s body. For more precise load estima-
tions we used surface electromyography to analyze
the muscle activities during the patient transfer. The
higher activity of the VM in comparison to the RF
while squatting and lifting was also observed in other
studies (Brinkmann et al., 2020; Dionisio et al., 2008;
Garrett and Kirkendall, 2000; Slater and Hart, 2017)
and can be considered as realistic. The mean activ-
ity of VM were around 46% higher in the squat po-
sition and around 50% higher while lifting the pa-
tient in comparison to the RF. This is assumed to
be due to the RF’s biarticular function as hip flexor
and knee extensor (Garrett and Kirkendall, 2000). In-
creased activity of the RF can lead to an increased
hip flexor/extensor torque and may cause pain in the
SIJ, and therefore lower back pain (Garrett and Kirk-
endall, 2000; Hammer et al., 2015). Furthermore, the
mean muscle activity of the BF increases around 45%
during the squat ascent compared to static squatting.
This was also observed in previous studies (Garrett
and Kirkendall, 2000; Dionisio et al., 2008; Slater and
Hart, 2017). Moreover, Brinkmann et al. found in a
further study conducting three different dynamic lift-
ing tasks that the mean activity of the quadriceps and
hamstring musculature increases with lifting higher
loads (Brinkmann et al., 2021). Also, an intra- as well
as an interindividual similarity of EMG muscle acti-
vation pattern regarding time and shape of the signals
for the selected muscles of the lower limb could be
observed.
ACKNOWLEDGEMENTS
Funded by the Lower Saxony Ministry of Science
and Culture under grant number 11-76251-12-10/19
ZN3491 within the Lower Saxony “Vorab“ of the
Volkswagen Foundation and supported by the Center
for Digital Innovations (ZDIN). The development of
the Multi Kinect System was funded by the German
Federal Ministry of Education and Research (Project
No. 16SV8532).
REFERENCES
3D4Medical (2020). Essential anatomy 5 (5.0.8).
3d4medical, dublin, ireland. accessed 10.05.2019.
Agrawal, A. and Ertel, W. (2018). Automatic nursing care
trainer based on machine learning. In KHD@ IJCAI,
pages 53–59.
Aiken, L. H., Sermeus, W., Van den Heede, K., Sloane,
D. M., Busse, R., McKee, M., Bruyneel, L., Raf-
ferty, A. M., Griffiths, P., Moreno-Casbas, M. T., et al.
(2012). Patient safety, satisfaction, and quality of hos-
pital care: cross sectional surveys of nurses and pa-
tients in 12 countries in europe and the united states.
Bmj, 344:e1717.
Baum, F., Beck, B., Fischer, B., Gl
¨
using, R., Graupner,
I., et al. (2012). Pr
¨
avention von r
¨
uckenbeschwerden;
Assessing Postures and Mechanical Loads during Patient Transfers
27
topas r–konzept der bgw f
¨
ur pflege und betreu-
ung. Hamburg, Germany: Berufsgenossenschaft f
¨
ur
Gesundheitsdienst und Wohlfahrtspflege.
Brinkmann, A., Fifelski, C., Lau, S., Kowalski, C., Meyer,
O., Diekmann, R., Isken, M., Fudickar, S., and Hein,
A. (2020). The aal/care laboratory – a healthcare pre-
vention system for caregivers. Nanomaterials and En-
ergy, 9(1):1–10.
Brinkmann, A., Fifelski-von B
¨
ohlen, C., Hellmers, S.,
Meyer, O., Diekmann, R., and Hein, A. (2021). Physi-
cal burden in manual patient handling – quantification
of lower limb emg muscle activation pattern of healthy
individuals lifting different loads ergonomically. In
HEALTHINF 2021.
Choi, S. D. and Brings, K. (2016). Work-related muscu-
loskeletal risks associated with nurses and nursing as-
sistants handling overweight and obese patients: A lit-
erature review. Work, 53(2):439–448.
Ding, J., Lim, Y.-J., Solano, M., Shadle, K., Park, C., Lin,
C., and Hu, J. (2014). Giving patients a lift-the robotic
nursing assistant (rona). In 2014 IEEE International
Conference on Technologies for Practical Robot Ap-
plications (TePRA), pages 1–5. IEEE.
Dionisio, V. C., Almeida, G. L., Duarte, M., and Hirata,
R. P. (2008). Kinematic, kinetic and emg patterns dur-
ing downward squatting. Journal of Electromyogra-
phy and Kinesiology, 18(1):134–143.
Fifelski, C., Brinkmann, A., Ortmann, S. M., Isken, M., and
Hein, A. (2018). Multi depth camera system for 3d
data recording for training and education of nurses. In
2018 International Conference on Computational Sci-
ence and Computational Intelligence (CSCI), pages
679–684. IEEE.
Freitag, S., Fincke, I., Dulon, M., Ellegast, R., and Nien-
haus, A. (2007). Messtechnische analyse von belas-
tenden k
¨
orperhaltungen bei pflegekr
¨
aften: eine geri-
atrische station im vergleich mit anderen kranken-
hausstationen. ErgoMed, 31:130–140.
Garrett, W. E. and Kirkendall, D. T. (2000). Exercise and
sport science. Lippincott Williams & Wilkins.
German Institute for Standardization (2005). Safety of ma-
chinery - human physical performance - part 4: Eval-
uation of working postures and movements in relation
to machinery; german version. DIN EN 1005-4:2005.
Hammer, N., M
¨
obius, R., Schleifenbaum, S., Hammer, K.-
H., Klima, S., Lange, J. S., Soisson, O., Winkler, D.,
and Milani, T. L. (2015). Pelvic belt effects on health
outcomes and functional parameters of patients with
sacroiliac joint pain. PloS one, 10(8).
Hermens, H. J., Freriks, B., Merletti, R., Stegeman, D.,
Blok, J., Rau, G., Disselhorst-Klug, C., and H
¨
agg, G.
(1999). European recommendations for surface elec-
tromyography. Roessingh research and development,
8(2):13–54.
H
¨
ohmann, U., Lautenschl
¨
ager, M., and Schwarz, L. (2016).
Belastungen im pflegeberuf: Bedingungsfaktoren, fol-
gen und desiderate. Pflege-Report, pages 73–89.
Hwang, J., Kuppam, V. A., Chodraju, S. S. R., Chen, J., and
Kim, J. H. (2019). Commercially available friction-
reducing patient-transfer devices reduce biomechani-
cal stresses on caregivers’ upper extremities and low
back. Human factors, 61(7):1125–1140.
J
¨
ager, M., Jordan, C., Theilmeier, A., Wortmann, N., Kuhn,
S., Nienhaus, A., and Luttmann, A. (2013). Lumbar-
load analysis of manual patient-handling activities for
biomechanical overload prevention among healthcare
workers. Annals of occupational hygiene, 57(4):528–
544.
Kliner, K., Rennert, D., and Richter, M. (2017). Gesund-
heit und Arbeit-Blickpunkt Gesundheitswesen: BKK
Gesundheitsatlas 2017. MWV Medizinisch Wis-
senschaftliche Verlagsgesellschaft.
Kusma, B., Glaesener, J.-J., Brandenburg, S., Pietsch, A.,
Fischer, K., Schmidt, J., Behl-Sch
¨
on, S., and Pohrt,
U. (2015). Der pflege das kreuz st
¨
arken - individu-
alpr
¨
avention r
¨
ucken bei der berufsgenossenschaft f
¨
ur
gesundheitsdienst und wohlfahrtspflege. Trauma und
Berufskrankheit, 17(4):244–249.
Lin, C., Kanai-Pak, M., Maeda, J., Kitajima, Y., Naka-
mura, M., Kuwahara, N., Ogata, T., and Ota, J. (2018).
Translational acceleration, rotational speed, and joint
angle of patients related to correct/incorrect methods
of transfer skills by nurses. Sensors, 18(9):2975.
Mansfield, P. and Neumann, D. (2008). Essentials of Kine-
siology for the Physical Therapist Assistant. Mosby.
Michaelis, M. and Hermann, S. (2010). Evaluation des
pflegekonzepts r
¨
uckengerechter patiententransfer in
der kranken-und altenpflege. Langzeit-Follow-up zur
Ermittlung der Nachhaltigkeit pr
¨
aventiver Effekte.
Schriftenreihe der Bundesanstalt f
¨
ur Arbeitsschutz
und Arbeitsmedizin (F2196). Dortmund.
Panjabi, M. M. and White, A. (1990). Clinical biomechan-
ics of the spine. Lippincott Philadelphia, PA.
Richardson, C. A., Snijders, C. J., Hides, J. A., Damen, L.,
Pas, M. S., and Storm, J. (2002). The relation be-
tween the transversus abdominis muscles, sacroiliac
joint mechanics, and low back pain. Spine, 27(4):399–
405.
Shadow (2020). Motion Workshop, Motion Capture Sys-
tem, www.motionshadow.com, accessed: 28.04.2020.
Slater, L. V. and Hart, J. M. (2017). Muscle activation pat-
terns during different squat techniques. Journal of
strength and conditioning research, 31(3):667–676.
Theilmeier, A., Jordan, C., Luttmann, A., and J
¨
aGer, M.
(2010). Measurement of action forces and posture to
determine the lumbar load of healthcare workers dur-
ing care activities with patient transfers. Annals of
occupational hygiene, 54(8):923–933.
van Wingerden, J.-P., Vleeming, A., Buyruk, H., and
Raissadat, K. (2004). Stabilization of the sacroiliac
joint in vivo: verification of muscular contribution to
force closure of the pelvis. European Spine Journal,
13(3):199–205.
Vleeming, A. and Stoeckart, R. (2007). The role of the
pelvic girdle in coupling the spine and the legs: a
clinical–anatomical perspective on pelvic stability. In
Movement, Stability & Lumbopelvic Pain, pages 113–
137. Elsevier.
Wei, L., Tsai, C.-Y., and Koontz, A. M. (2018). The rela-
tionship between joint ranges of motion and joint ki-
HEALTHINF 2021 - 14th International Conference on Health Informatics
28
netics during sitting pivot wheelchair transfers. Pro-
ceedings of the Conference of the Rehabilitation En-
gineering and Assistive Technology Society of North
America.
Weißert-Horn, M., Meyer, M. D., Jacobs, M., Stern,
H., Raske, H.-W., and Landau, K. (2014). Er-
gonomisch richtige arbeitsweise beim transfer von
schwerstpflegebed
¨
urftigen. Zeitschrift f
¨
ur Arbeitswis-
senschaft, 68(3):175–184.
Wu, G., Siegler, S., Allard, P., Kirtley, C., Leardini, A.,
Rosenbaum, D., Whittle, M., D D’Lima, D., Cristo-
folini, L., Witte, H., et al. (2002). Isb recommendation
on definitions of joint coordinate system of various
joints for the reporting of human joint motion—part
i: ankle, hip, and spine. Journal of biomechanics,
35(4):543–548.
Assessing Postures and Mechanical Loads during Patient Transfers
29