Prediction Of Muscle Power In Elderly Using Functional Screening Data by "Vivek Vijay" From IIT, Jodhpur .The presentation was done at #doppa17 DevOps++ Global Summit 2017. All the copyrights are reserved with the author.
Prediction Of Muscle Power In Elderly Using Functional Screening Data
1. Prediction of muscle power in elderly
using functional screening data
Indian Institute of Technology Jodhpur 1
By
Dr Vivek Vijay & Brajesh Shukla
2. Ageing
What happens when people age?
A. Functional decline
B. Loss of muscle strength and power
C. Falling
2
3. Old age
fall
statistics
Ageing inWorld - 8.5%
Fall rate in elderly adults – 30-60% [1]
Injury/Death rate – 10-20 % [1]
Reasons of fall – age, living alone, foot
problem, visual impairment, psychological
status, etc. [1]
3
[1] Rubenstein, L.Z., Falls in older people: epidemiology, risk factors and strategies for
prevention. Age and Ageing, 2006. 35(suppl 2): p. ii37-ii41
4. Old age
fall
statistics
INDIA
Ageing in India - 7.7 % > age 60 [3]
Prediction up to 2050- 19%
Prevalence of fall - 47-53% [4]
[3] Krishnaswamy, B. and U. Gnanasambandam, Falls in older people, India, in
National/ Regional review. 2007, Department of Geriatric Medicine, Madras
Medical College and government General Hospital,Chennai City,Tamil Nadu
State, India. p. 19.
4
9. 5Times sit
to stand test
9
http://www.drdenizdogan.com/2012/03/5-defa-oturup-kalkma-testi.html
10. Data
collection
• 101 participants (25 females, 76 males)
• Average age 70.099 ± 5.3575 years
• No major medical conditions in last 24
months
• Ethical approval from S.N. Medical
College of Jodhpur
• Collected different functional screening
tests such as OLS, STS,TUG, Gait
Velocity, BQT, Grip strength and
different questionnaire
10
11. Explanatory
data analysis
Descriptive statistics (numerical summaries):
mean, range, variance, skewness, kurtosis
etc.
Non-parametric tests were used to check
the normality of data.
Graphical methods:
I. frequency distribution histograms
II. scatter plots
III.Normal probability plots
11
12. Bagging
( Bootstrap
aggregating)
It is an ensemble method: a method of combining
multiple predictors. Used most commonly to provide a
measure of accuracy of a parameter estimate,
especially when the data set is small
12
An example
X=(3.12, 0, 1.57,
19.67, 0.22, 2.20)
Mean=4.46
X1=(1.57,0.22,19.67,
0,0,2.2,3.12)
Mean=4.13
X2=(0, 2.20, 2.20,
2.20, 19.67, 1.57)
Mean=4.64
X3=(0.22, 3.12,1.57,
3.12, 2.20, 0.22)
Mean=1.74
13. Cross
Validation
Used to estimate test set prediction error
rates associated with a given learning
method to evaluate its performance.
The data set is divided into k subsets, and the
method is repeated k times.
Each time, one of the k subsets is used as the
test set and the other k-1 subsets are put
together to form a training set.
Then the average error across all k trials is
computed.
13
14. Previous
works
(Takai etAll)
Sit-to-standTest to Evaluate Knee Extensor Muscle Size and
Strength in the Elderly: A Novel Approach– Takai et. Al
(2009) [5]
The authors used an equation to estimate muscle power in
a 5 time sit to stand
The variables used in the power equation were leg length,
body weight, and the time to complete the 5 STS
Power=
𝐿−0.4 × 𝐵𝑜𝑑𝑦 𝑚𝑎𝑠𝑠 ×5𝑔
𝑡𝑖𝑚𝑒 𝑆𝑇𝑆
14
15. Previous
works
(Smith et.All)
Simple equations to predict concentric lower-body
muscle power in older adults using the 30-second
chair-rise test: a pilot study – Smith et. Al (2010) [6]
Two variable equation - body weight and number
of stands completed in 30 sec
Y= B0+B1X1+B2X2+B3X3+B4X4+Error
X1 – no of stands, X2 – weight, X3 – femur length,
X4 – gender
X3 and X4 do not contribute significantly
15
16. Prediction
of muscle
power
Linear regression to predict muscle power in
elderly using functional screening data
Linear regression to develop a predictive model
for muscle power
Five features – BMI, Gait velocity, STS,TUG, Grip
strength
Y= B0+B1X1+B2X2+B3X3+B4X4+ B5X5+Error
X1 – BMI, X2 – GaitVelocity, X3 – STS, X4 –TUG,
X5- Grip strength
16
17. Prediction
of muscle
power
Estimation of parameters
Y= 1.8657+2.3440X1+17.67X2-3.8458X3+0.8633X4+
2.040X5+Error
X1 – BMI, X2 – GaitVelocity, X3 – STS, X4 –TUG,
X5- Grip strength
R
2
= 0.7509 F statistics=44.01 P value=0.0000
17
18. Residual
Analysis
The linear regression model
Y= B0+B1X1+B2X2+B3X3+B4X4+Error
The residuals are defined as the n difference.
The random departures ( residuals) are
assumed
i. To have zero mean
ii. To have a constant variance
iii. Independent and follow a
normal distribution.
18
19. Prediction
of muscle
power
Regression to predict muscle power in elderly using
functional screening data
Improving result by adding non-linearity of STS.
Y= 82.1964+2.4144X1+5.0720X2-10.7678X3+0.1670X3
2
+
0.3995X4+1.9240X5+Error
R
2
=0.7900, F statistics= 45.1402, P value=0.000
19
-3
-2
-1
0
1
2
3
-60 -40 -20 0 20 40 60
Q-Q Plot of Power Residual
20. Conclusion
We use linear regression to predict muscle power
in elderly using functional screening data.
By predicting muscle power we can easily detect
sarcopenia ( Loss of skeletal muscle mass) in
elderly.
Sarcopenia is major reason behind functional
decline and fraility in Elderly.
20
22. References
[1] Rubenstein, L.Z., Falls in older people: epidemiology, risk factors and strategies for
prevention. Age and Ageing, 2006. 35(suppl 2): p. ii37-ii41
[2] Heinrich, S., et al., Cost of falls in old age: a systematic review. Osteoporosis
International, 2010. 21(6): p. 891-902.
[3] Krishnaswamy, B. and U. Gnanasambandam, Falls in older people, India, in
National/ Regional review. 2007, Department of Geriatric Medicine, Madras Medical
College and government General Hospital, Chennai City,Tamil Nadu State, India. p.
19.
[4] Dsouza, S.A., et al. Falls in Indian older adults: A barrier to active ageing. Asian
Journal of Gerontology and Geriatrices, 2014. 9, 33-40.
[5]Takai,Yohei, et al. "Sit-to-stand test to evaluate knee extensor muscle size and
strength in the elderly: a novel approach." Journal of physiological anthropology 28.3
(2009): 123-128.
[6] Smith, Wesley N., et al. "Simple equations to predict concentric lower-body
muscle power in older adults using the 30-second chair-rise test: a pilot study." Clin
Interv Aging 5.5 (2010): 173-180.
22