Automated Assessment of Mobility in Bedridden Patients

Advisor: Dr. Chun-Ju Hou

Presenter: Si-Ping Chen

Date:2014/12/10

35th Annual International Conference of the IEEE EMBS Osaka, Japan, 3 - 7 July, 2013

Stephanie Bennett, Member, IEEE, Rafik Goubran, Fellow, IEEE, Kenneth Rockwood, and Frank Knoefel.

2

Outline

• Introduction• Methods• Results• Conclusions

3

Introduction

• Prevalence of population aging

4

Introduction

• The impact of population aging on society。National financial load increase。Declining economic growth。Political attention to the elderly-related policies。Business and consumer behavior change。Adjustment of the real estate industry。Shift the focus of education

5

Introduction

• Geriatric giants –The major categories of impairment that appear in elderly

people, especially as they begin to fail. These include• Immobility• Instability• Incontinence• Impaired intellect/memory

6

Introduction

• The Hierarchical Assessment of Balance and Mobility (HABAM)– Balance– Transfers– Mobility

• HABAM tools– Bedridden patients– Mildly functional impaired patients

7

Introduction

• Problems– Spend time on manually assessing HABAM The quality of

healthcare– Not currently used enough by hospital staff to be informative– Population aging Overloading health resources and staff

• Automatic and computerized measurement– Accelerometers and pressure sensors

• Sleep pattern, gait pattern, pressure ulcer, smart home environment

8

Purpose

• The goal of this work is often not only to automatic, but improve current health measurements particularly to impaired mobility and immobility.

9

Methods

• Equipment and Set-up– A laptop – A video camera – Three pressure sensitive mats manufactured by S4 sensors– Software

10

Methods

• Equipment and Set-up– Sensor captures this data at sampling rate of 10Hz.– The data generated by the mats is sent via Bluetooth to a laptop.

11

Methods

• Equipment and Set-up– Constructed of four panels:

。One large panel supporting the back.。Two small panels at the sacrum.。An intermediate sized panel supporting the legs.

Mat 80cm

25cm

12

Methods

• Equipment and Set-up

13

Methods

• Experimental Procedure– Data were collected from the mattress alone and five volunteers

performing entirely in-bed enactments of HABAM scores.– HABAM scores:

。Score 0: Needs positioning in bed.。Score 4: Positions self in bed.。Score 7: Lying-sitting independently.

14

Methods

• Data Analysis– Sums of sensors scores over time– Algorithms: HABAM score enactments

• Subsystem 1: eliminated the weight of the mattress from all volunteer performed enactment.

• Subsystem 2: determined if the volunteer was enacting a score of 0, or a score of 4.

• Subsystem 3: determined if the volunteer was enacting a score of 7.

15

Methods

• Subsystem 1: ‘zero-ed’ data

framesmattrawdatatzeroed sensorsensorsensor /)()()(

Average mattress pressure

16

Methods

• Subsystem 2– Only consider the middle mattress at sacrum region

1. Calculate baseline pressure ((t)) for each sensor

2. Calculate the percentage change over time from baseline for each sensor

3. Moving average filter for individual sensor (W=5)

Left Right

17

Methods

• Subsystem 24. Points in time at which a sensor recorded a percentage decrease of -

0.95 or less were recorded, least two sensors had simultaneously dropped below a percentage change of -0.95.

Left Right

-0.95 -0.95

Score 4( Positions self in bed )

Left Right

-0.95 X

Score 0( Needs positioning in bed )

18

Methods

• Subsystem 3 – Summed data from the top mat was divided by summed data from

the bottom mat to get a ratio of proportional distribution of the body over these two mats.

(t)/(t)

。≥ 1.0 Lying。< 1.0 Sitting

19

Results

• Subsystem 2 determined if the enactment was either of score 0: needs positioning in bed, or score 4: can position self in bed.– Relief of three sensors under the left hip during enactment

20

Results

• Location of relieved sensors in the mat.

21

Results

• Subsystem 3 determined if the enactment was of score 7:lying-sitting.

• This was done by calculating the sums, then ratios of the top mat and the bottom mat at every point in time.

22

Results

• Top over bottom mat ratio for scenario 3.

23

Results

• Overall results of the system can be observed in Table I.

24

Conclusions

• This paper aimed to automate a volunteer-based, partial HABAM assessment.

• Five volunteers performed three enactments each, on a standard hospital bed while pressure data was gathered from pressure mats underneath a hospital mattress.

25

Conclusions

• Subsystem 1 :‒ To identify and distinguish between a subject in a sitting or lying

position.

• Subsystem 2:– For expansion to include examination of pressure points and

associated patterns underneath a subject during HABAM enactments.

26

Conclusions

• Data revealed that the system had not assessed incorrectly.• This system, with relative engineering simplicity, was able to

better assess HABAM scores than an observing researcher.• HABAM

– Emphasize the importance of pervasive computing in the assessment and tracking of immobility.

27

Thanks for your attention