DOS 2017

DOS Kongressen 2017 ·

Using wearable sensors to determine knee range of

movement in knee arthroplasty patients. A pilot study.

Mie Christina Hansen, Rasmus Malik Thaarup Høegh, Jacob Fyhring Mortensen,

Helge Bjarup Dissing Sørensen, Anders Odgaard

Ortopædkirurgisk afdeling, Herlev-Gentofte Hospital, Gentofte; Department of

Electrical Engineering, Technical University of Demark (DTU); Ortopædkirurgisk

afdeling, Herlev-Gentofte Hospital, Gentofte; Department of Electrical Engi-

neering, Technical University of Demark (DTU); Ortopædkirurgisk afdeling, Her-

lev-Gentofte Hospital, Gentofte

Background:

Range of motion (ROM) under specific activities may be related

to patients’ experience of the quality of knee function. It may be superior com-

pared to measurements of passive ROM in an ambulatory.

Purpose / Aim of Study:

The objective of this study is to investigate and de-

velop a new method to determine dynamic knee function using wearable sen-

sors.

Materials and Methods:

35 subjects aged 60-75 were included in the study.

They met 1 of 3 criteria: Healthy (normal functioning knee), pre-operative or

3 months post-operative. We used 2 sensors with a magneto-, accelero- and

gyrometer that measure the angle between the femur and tibia. With 2 sensors

taped to the skin, patients performed 2 exercises; walk on treadmill 2 km/h and

individually chosen fast pace. For reproducibility the exercises were carried out

twice.

Findings / Results:

Significant difference in knee function between the 3

groups was found. Subjects with healthy knee had significant higher active ROM

(mean = 53.6°), angular velocity and acceleration than the 2 other groups. The

post-operative’s active ROM (mean = 43.4°) were significantly higher than

the pre-operative group (mean = 37.4°), but the angular velocity and accel-

eration were not significantly different from the pre-operative group. Retests

showed the method to be highly reproducible. The largest difference between

the groups was observed when instructed to walk at a fast pace.

Conclusions:

The sensors measures of dynamic ROM were reproducible and

consistent with the degree of pathology. This method is simple and allows gait

analysis to be telemetric in the patient’s daily life, and it thus fundamentally dif-

fers from costly laboratory tests. The method can be used to monitor patients

both pre- and postoperatively. Further studies will investigate whether the ap-

plication of machine learning on the data can predict Oxford Knee Score.

No conflicts of interest reported

39.