Activities of Daily Living in Healthy Adults and Children

Reliability of Registrations with Multiple Body Worn Sensors

Ryanne Lemmens

1,2

, Henk Seelen

1,2

, Yvonne Janssen-Potten

1,2

, Annick Timmermans

1,3

Marlous Schnackers

, Annet Eerden

, Richard Geers

and Rob Smeets

1,2

Department of Rehabilitation Medicine, Research School CAPHRI, Maastricht University, Maastricht, Netherlands

Adelante, Centre of Expertise in Rehabilitation and Audiology, Hoensbroek, Netherlands

BIOMED Biomedical Research Institute, Hasselt University, Hasselt, Belgium

1 OBJECTIVES

Patients with stroke or cerebral palsy often

encounter arm-hand problems during daily life.

Assessment is important to determine the progress of

arm-hand performance in patients during

rehabilitation, and to ascertain the effectiveness of

therapies. Many instruments are available to assess

capacity or perceived performance, but instruments

assessing actual performance are scarce (Lemmens

et al., 2012). Inertial sensors may be used to assess

actual performance. However, signal reliability

during execution of activities of daily living (ADL)

should be determined first.

Aim of this study was to examine the reliability

of the data of upper extremity skill performance in a

standardized setting, in both healthy adults and

healthy children, registered using a combination of

multiple body worn sensors.

2 METHODS

In this non-randomised cross-sectional study, both

healthy adults (aged > 50 years) and healthy children

were included. Because motor control processes may

mature with age, resulting in differences in skill

execution, the children were divided into two age

groups, i.e. 6-11 years and 12-18 years. Four 9-DOF

sensor devices, each containing a tri-axial

accelerometer, gyroscope and magnetometer were

attached to the dominant hand, wrist, upper arm and

chest of participants.

Data were registered during the execution of 5

repetitions of 2 tasks, i.e. ‘drinking from a cup’ and

‘eating with knife and fork’. Tasks were first

performed without extensive instructions (e.g. with

the instruction: “drink from the cup.”), and

subsequently with extensive instructions on how to

perform the task (e.g. “reach to the cup, grasp it,

bring it to your mouth, take a sip, put the cup back

on the table and go back to your starting position.”).

Signals were filtered with a 4

order zero-time

lag low-pass Butterworth filter (cut off frequency of

1.28 Hz). Repetitions of each specific task were

identified and intra-class correlation coefficients

(ICC) for each sensor and signal type were

determined as a measure of reliability, both within

and between subjects. For every person, a mean ICC

was measured. Since data was not normally

distributed, medians were calculated. The ICCs were

classified based on the kappa statistic classification

of Landis and Koch, i.e. ICC between 0.8-1.0=very

good; 0.6-0.8=good; 0.4-0.6=moderate; 0.2-0.4=fair;

<0.2=slight (Landis and Koch, 1977).

3 RESULTS

Figure 1: Within-subject reliability expressed as intra-class

correlation coefficients, for the skills eating and drinking.

Dark grey bars represent the performance of the task

without instruction, whereas light grey bars represent the

performance of the task with instruction.

Lemmens R., Seelen H., Janssen-Potten Y., Timmermans A., Schnackers M., Eerden A., Geers R. and Smeets R..

Activities of Daily Living in Healthy Adults and Children - Reliability of Registrations with Multiple Body Worn Sensors.

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Thirty adults were included (14 women, 16 men,

mean age 58.0 ± 5.1 years), 16 children aged

between 6-11 years (9 girls, 7 boys, mean age 8.5 ±

1.7 year) and 16 children aged between 12 and 18

years (8 girls, 8 boys, mean age 14.6 ± 1.5 years).

Figure 1 displays within-subject reliability of the

skills drinking and eating for both the adults and the

children. With regard to the within subject

reliability, the median ICC’s were good for the skill

eating and very good for the skill drinking.

Reliability was better for the skill performed with

instruction compared to the skill performed without

instruction, especially for the skill eating.

Furthermore it can be seen that the children aged

between 12-18 years showed a slightly higher

reliability compared to the children aged between 6-

12 years.

Figure 2 displays between-subject reliability of

the skills drinking and eating for both the adults and

the children.

Figure 2: Between-subject reliability expressed as intra-

class correlation coefficients, for the skills eating and

drinking. Dark grey bars represent the performance of the

task without instruction, whereas light grey bars represent

the performance of the task with instruction.

Between-subject reliability was good to very

good for both skills performed by adults. The skills

drinking performed by the children had a very good

reliability whereas the skill eating had a fair to

moderate reliability in the youngest children, and a

good reliability in the older children. Especially for

the skill eating, a big difference was seen regarding

the reliability of the performance without instruction

compared to the performance with instruction. For

the skill drinking, reliability was comparable

between younger children and the older children

whereas for the skill eating, performance of the older

children had a higher reliability.

4 DISCUSSION

Overall, the skill drinking as well as the skill eating

had a good to very good within-subject reliability in

both adults and children. Performance with

instruction had a higher reliability compared to

performance without instruction. By giving

instructions about how to perform the task, the

variability in execution of the task was reduced,

thereby increasing the reliability.

The performance of the skill drinking had a

higher reliability compared to the skill eating. This

can be explained by the complexity of the skills, i.e.

drinking is a rather simple skill, which cannot be

performed in many ways, whereas the skill eating

consists of more sub movements and can, in

addition, be performed in many different ways.

Between-subject reliability of the skill eating was

relatively low, especially for the performance

without instruction in the youngest children. This

can be explained by the fact that many children did

not use the knife to cut the food, whereas other

children had difficulties manipulating the knife. For

the performance with instruction, they were told

how to use the knife.

In conclusion, we have shown that a

combination of multiple body worn sensors is able

to reliably register activities of daily living in

healthy adults as well as in healthy children. Future

research will focus on the investigation of signal

reliability during activities of daily living performed

by patients and in a daily life setting.

REFERENCES

Landis, J. R. and Koch, G. G. (1977) The measurement of

observer agreement for categorical data. Biometrics,

33, 159-74.

Lemmens, R., Timmermans, A., Janssen-Potten, Y.,

Smeets, R. and Seelen, H. (2012) Valid and reliable

instruments for arm-hand assessment at ICF activity

level in persons with hemiplegia: a systematic review.

BMC Neurology, 12, 21.