Moving Forward Using Automated Measures for Lameness Detection - - PowerPoint PPT Presentation

Moving Forward Using Automated Measures for Lameness Detection

Núria Chapinal, PhD Animal Welfare Program, UBC April 14, 2010

 Introduction  Visual/subjective methods of detection  Automated methods of detection

 Examples

 Do they work?

 Experimental results

 Conclusions and practical applications

Outline

Introduction

 Lameness is a major welfare and productivity

problem in dairy cattle

 Traditional assessment method: visual

• bservation

 Herds are getting larger  Producers have difficulties detecting lame cows

Simple (fast), accurate and repeatable

Introduction

 Automated methods of detection available

 Automated gait assessment  Automated monitoring of other behaviors

 Automated gait assessment

 Video motion analysis (Flower et al. 2005)  Ground reaction force (Rajkondawar et al.

2006)

Introduction

 Lame cows:

 Lie down for longer (e.g. Chapinal et al.,

2009)

 Change weight distribution among legs

when standing (e.g. Rushen et al. 2007; Pastell and Kujala 2007)

 Have reduced mobility (e.g. visit a milking

robot less frequently, Borderas et al. 2008)

 Subjective  Vague description of lameness degrees  Inter and intra observer reliability  Not properly validated  Training  Time consuming

Visual methods for gait assessment

Subjective methods for gait assessment

Swinging in/out Back arch Joint flexion Tracking up Head bob Asymmetric steps Reluctance to bear weight

(Flower & Weary 2006 J. Dairy Sci. 89:139-146)

1 = not lame 5 = severely lame More than 90% of cases correctly classified as having a sole ulcer or not.

Gait score can predict sole ulcers

1 1.5 2 2.5 3 3.5 4

• 8
• 4

4

Overall gait

Overall gait † ** * *

1 1.5 2 2.5 3 3.5 4

• 8
• 4

4

Overall gait

Overall gait † ** * *

Week relative to diagnosis Sole ulcer Hemorrhage No lesions

(Chapinal et al. 2009 J. Dairy Sci. 92: 4365-4374)

Swinging in/out Back arch Joint flexion Tracking up Head bob Asymmetric steps Reluctance to bear weight

(Flower & Weary 2006 J. Dairy Sci. 89:139-146) (Chapinal et al. 2009 J. Dairy Sci. 92: 4365-4374)

Subjective methods for gait assessment

 Objective = Repeatable  Reduced labor  Continuous monitoring (changes within cows)

= Increased accuracy

 Some haven’t been properly validated yet  Becoming affordable

Automated methods for lameness detection

 Visits to a milking robot  Activity

 Lying behavior (time, bouts)  Steps

 Walking acceleration patterns  Weight distribution while standing  Ground reaction force while walking

Automated methods for lameness detection

 IceTag accelerometer (IceRobotics)  AfiMilk Pedometer Plus Tag (SAE Afikim)  Hobo G pendant acceleration logger

(Onset Computer Corporation)

 H-tag motion sensor (SCR)

Activity

Activity

Activity measures

 Lying bouts/day  Lying bout duration  Lying time/day  Steps/day  Acceleration patterns

Acceleration patterns

• 3.5
• 2.5
• 1.5
• 0.5

0.5 1.5 2.5 3.5 1 2 3 4 5

Seconds Acceleration (g)

• 3.5
• 2.5
• 1.5
• 0.5

0.5 1.5 2.5 3.5 1 2 3 4 5

Seconds Acceleration (g)

A

De Passillé et al. J. Dairy Sci. in press

Weight distribution: weighing platform

Weight distribution and shifting among legs

FRONT LEFT FRONT RIGHT BACK LEFT BACK RIGHT Total WEIGHT

100 200 300 400 500 600 700

10:52:37 10:53:26 10:54:14 10:55:03 10:55:52

Time Kg

Ground reaction forces: Stepmetrix (BouMatic)

Rajkondawar et al. 2006 J. Dairy Sci. 89:4267-4275 Bicalho et al. 2007 J. Dairy Sci. 90:3294-3300

Lameness scored based on 5 limb movement variables (measures of stride and weight bearing)

Do they work?

Automated milking systems collect data on cow behaviour: Lame cows go to robotic milkers less often

20 40 60 80 100 120 Frequent visitors Infrequent visitors % cows Not lame Lame

Borderas et al. 2008

• Can. J. Anim. Sci. 88:1-8

Weight distribution

Lame cows do not distribute weight evenly between contralateral legs

100 200 300 400 500 600 700

10:52:37 10:53:26 10:54:14 10:55:03 10:55:52

Time Kg

BACK LEFT BACK RIGHT TOTAL

Lame cows shift weight more often between contralateral legs

100 200 300 400 500 600 700

10:52:37 10:53:26 10:54:14 10:55:03 10:55:52

Time Kg

BACK LEFT BACK RIGHT TOTAL

Weight distribution measures

 For each pair of legs (front and back)

 WEIGHT ASSYMETRY



Leg weight ratio = weight on lighter/weight on heavier leg



E.g. 50% on left leg, 50% on right leg LWR = 50/50 = 1



60% on left leg, 40% on right leg LWR = 40/60 = 0.67

 WEIGHT SHIFTING:



Variability (SD) over time of weight applied to each pair

• f legs



Number of kicks

Weight distribution

Pastell & Kujala 2007 J. Dairy Sci. 90:2283-2292

Not lame Mild lameness Moderate lameness Severe lameness Pastell & Kujala 2007 J. Dairy Sci. 90:2283-2292

Measures of weight distribution can detect lameness promptly

Pastell & Kujala 2007 J. Dairy Sci. 90:2283-2292

Combination of methods: Does accuracy increase?

Experimental set-up for gait scoring and measuring weight distribution

WEIGHING PLATFORM GAIT SCORE 9 m

Weight distribution and activity (Exp 1)

Chapinal et al. J. Dairy Sci. in press

Overall gait score correlated with:

• Weight shifting in the rear legs (SD):

r = 0.23 ; P = 0.01

• Symmetry of rear legs (leg weight ratio):

r = - 0.52; P = 0.002

• Frequency of steps:

r = - 0.43; P < 0.001

Weight distribution and activity (Exp 1)

Chapinal et al. J. Dairy Sci. in press

Cows with severe hoof infections were more asymmetric in the rear legs

• Mean leg weight ratio ± SE =

0.75 ± 0.05 vs. 0.80 ± 0.03; P = 0.006

• OR = 1.2 ; P = 0.03

for each 5% decrease in leg weight ratio

Day 1 Day 2 Day 3 Day 4 Lameness Detection (objective 1) Effect of analgesia (objective 2) Ketoprofen (3mg/kg BW) / Saline (im)

Weight distribution and activity (Exp 2)

* Lame cows: overall gait score > 3 (Flower & Weary 2006)

Lameness Detection: Weight distribution among legs

Variable Non-lame Lame OR 95%CI Rear legs weight variability (SD, kg) 24.1 ± 2.0 32.6 ± 2.2 * 1.4 1 1.1– 1.8 Front legs weight variability (SD, kg) 16.5 ± 1.5 22.6 ± 1.7 ** 1.6 1 1.1 – 2.3 Rear leg weight ratio 0.9 ± 0.02 0.8 ± 0.02 ** 0.7 2 0.5 – 0.9

1 OR adjusted to a 5-kg increase; 2 OR adjusted to a 5% increase

Chapinal et al. J. Dairy Sci. in press

Lameness Detection: Activity and walking speed

Variable Non-lame Lame OR 95%CI Lying time (min/day) 720.1 ± 23.2 787.6 ± 27.1 † 1.1 1 1.0 – 1.3 Lying bout duration (min) 73.9 ± 3.9 89.7 ± 4.6 * 1.5 1 1.1 – 2.1 Walking speed (m/s) 1.5 ± 0.4 1.3 ± 0.4 ** 0.7 2 0.5 – 0.9

1 OR adjusted to a 30-min increase; 2OR adjusted to a 0.1 m/s increase

Chapinal et al. J. Dairy Sci. in press

SD of the weight of the rear legs (AUC = 0.71) SD + lying bout duration (AUC = 0.76) SD + bout duration + speed (AUC= 0.83)

Combining measures of weight distribution, activity and walking speed improved lameness detection

Chapinal et al. J. Dairy Sci. in press

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

1 - Specificity Sensitivity

The SD of the weight applied to the rear legs significantly decreased after the ketoprofen injections

15 20 25 30 35 40 1 2 3 4

Day SD of the weight (kg)

Injections

Ketoprofen Saline Chapinal et al. J. Dairy Sci. in press

 Lame cows show:

 Asymmetry in weight distribution  Frequent weight transfer

 Lame cows usually have

 Longer lying bouts  Longer daily lying times  Decreased activity (steps)

although differences not always significant! Lameness Detection

Variability in activity measures

Lying time (h/d) Ito et al. 2009 J. Dairy Sci. 92:4412-4420

Farm ID Lying time (h/day)

20 40 60 80 100 120 140 160 180 200

2 4 6 8 10 12 14 16 18 20 22 Hour of day Steps/h

Variability in activity measures

Chapinal et al. J. Dairy Sci. in press 2 milkings / day

Steps/h Hour of day

Non-lame Lame

Variability in activity measures

Chapinal et al. J. Dairy Sci. in press

20 40 60 80 100 120 140 160 180 200

2 4 6 8 10 12 14 16 18 20 22 Hour of day Steps/h

3 milkings / day

Steps/h Hour of day

Non-lame Lame

Acceleration patterns

Chapinal et al. 2010. First North American Conference on Precision Dairy Management

Overall gait score Symmetry of acceleration (%)

Acceleration patterns

Chapinal et al. 2010. First North American Conference on Precision Dairy Management

 Automated methods of weight distribution and

activity show promise for on-farm lameness detection, particularly when combined

 These methods may provide a tool for future

evaluation of lameness therapies

Conclusions

 Continuous monitoring of activity

(heat detection, lameness,

• ther diseases)

 Milking robots (+ weighing platform?)

Practical applications

Borderas, T.F., A. R. Fournier, J. Rushen, and A.M. de Passillé. 2008. Effect of lameness on dairy cows' visits to automatic milking systems. Can. J. Ani. Sci. 88:1- 8. Bicalho, R. C., S. H. Cheong, G. Cramer, and C. L. Guard. 2007. Association between a visual and an automated locomotion score in lactating Holstein cows. J. Dairy Sci. 90:3294-3300. Chapinal, N., A. M. de Passille, and J. Rushen. 2009. Weight distribution and gait in dairy cattle are affected by milking and late pregnancy. J. Dairy Sci. 92:581-588. Chapinal, N., A. M. de Passillé, J. Rushen, and S. Wagner. Automated methods for the detection of lameness and analgesia in dairy cattle. J. Dairy Sci. (in press). Chapinal, N., A. M. de Passillé, J. Rushen, and S. Wagner. Effect of hoof trimming

• n gait, weight distribution and activity of dairy cattle. J. Dairy Sci. (in press).

Chapinal, N., M. Pastell, L. Hänninen, J. Rushen, A.M. de Passillé. 2010. Walking acceleration patters as a method for lameness detection. Proceedings of the First North American Conference on Precision Dairy Management, p.126-127.

References

De Passillé, A. M., M. B. Jensen, N. Chapinal, and J. Rushen. Technical note: Use

• f accelerometers to describe gait patterns in dairy calves. J. Dairy Sci. (in press).

Flower, F. C., D. J. Sanderson, and D. M. Weary. 2005. Hoof pathologies influence kinematic measures of dairy cow gait. J. Dairy Sci. 88:3166-3173. Flower, F. C. and D. M. Weary. 2006. Effect of hoof pathologies on subjective assessments of dairy cow gait. J. Dairy Sci. 89:139-146. Ito, K., D. M. Weary, and M. A. G. von Keyserlingk. 2009. Lying behavior: Assessing within- and between-herd variation in free-stall-housed dairy cows. J. Dairy Sci. 92: 4412-4420. Rushen, J., E. Pombourcq, and A. M. d. Passillé. 2007. Validation of two measures

• f lameness in dairy cows. Appl. Anim. Behav. Sci. 106:173-177.

Pastell, M. E. and M. Kujala. 2007. A probabilistic neural network model for lameness detection. J. Dairy Sci. 90:2283-2292. Rajkondawar, P. G., M. Liu, R. M. Dyer, N. K. Neerchal, U. Tasch, A. M. Lefcourt, B. Erez, and M. A. Varner. 2006. Comparison of models to identify lame cows based on gait and lesion scores, and limb movement variables. J. Dairy Sci. 89:4267-4275.

References