An observational study on the management of digital dermatitis through a repeated risk assessment on 19 Dutch dairy herds

Machine learning approaches to predict and detect early-onset of digital dermatitis in dairy cows using sensor data

The present study aimed to employ machine learning algorithms based on sensor behavior data for (1) early-onset detection of digital dermatitis (DD) and (2) DD prediction in dairy cows. Our machine learning model, which was based on the Tree-Based Pipeline Optimization Tool (TPOT) automatic machine learning method, for DD detection on day 0 of the appearance of the clinical signs has reached an accuracy of 79% on the test set, while the model for the prediction of DD 2 days prior to the appearance of the first clinical signs, which was a combination of K-means and TPOT, has reached an accuracy of 64%. The proposed machine learning models have the potential to help achieve a real-time automated tool for monitoring and diagnosing DD in lactating dairy cows based on sensor data in conventional dairy barn environments. Our results suggest that alterations in behavioral patterns can be used as inputs in an early warning system for herd management in order to detect variances in the health and wellbeing of individual cows.

A case-control study regarding factors associated with digital dermatitis in Norwegian dairy herds

Background

Digital dermatitis (DD) is a contagious bovine foot disease causing painful lesions, lameness, and reduced animal welfare. Previous studies indicate a complex aetiology of the disease. The aim of this study was to compare DD negative and DD positive herds to identify factors associated with DD in Norwegian dairy herds by analysing data obtained in a questionnaire and data recorded in the Norwegian Dairy Herd Recording System (NDHRS). The questionnaire was e-mailed to the owners of all 380 herds recorded with DD in 2019 and to 1530 randomly selected herds with no recordings of DD. Altogether 559 dairy herds were included in the final study population, of which 113 was classified as DD positive (≥ one cow recorded with DD in NDHRS) and 446 as DD negative. When stratified by housing system, the ratio between DD positive and negative herds was 108/281 in free stalls and 5/165 in tie stalls. Multilevel logistic regression was used to model the association between potential risk factors and variables related to the detection and diagnosis of DD, and the outcome in the free-stall population. Geographical area (county) was included in the model as a random effect.

Results

In the final study population 108/113 (96%) of the DD positive herds were housed in free stalls versus 5/113 (4%) in tie stalls. The free-stall herds’ DD status was associated with purchase of cattle during the last 5 years (baseline: 0 animals, OR = 2.30 for category 12–27 animals, OR = 4.34 for 28–52 animals, and OR = 5.39 for  ≥ 53 animals). The DD status was also associated with claw trimming frequency (Baseline: 1 < 2/year, OR = 0.41 for category  < 1/year, and OR = 4.09 for  ≥ 3/year), whether the claw trimming was done by a certified professional trimmer or not (baseline:  ≤ 90% of the cows, OR = 3.98 for category  ≥ 90% of the cows), cleaning of feet in the chute before trimming (baseline: no cleaning, OR = 1.98 for category cleaning), and alley flooring (baseline: slatted floor, OR = 2.36 for category solid floor).

Conclusions

Digital dermatitis was far more frequent in Norwegian dairy herds housed in free stalls versus those housed in tie stalls. In the free-stall herds purchase of cattle, increasing trimming frequency, use of certified professional trimmer, cleaning of the feet in the chute, and solid flooring in the alleys were associated with increased odds of recorded DD.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13028-022-00635-0.