Analysis of Dairy Cow Behavior during Milking Associated with Lameness

Round 1

Reviewer 1 Report

dairy-2599901-peer review-v1
Analysis of dairy cow behavior during milking associated with lameness.

REMARKS AND RECOMMENDATIONS
The authors compared the behavior of lame and non-lame cows in milking parlors in order to determine whether lameness could be related to behavioral changes in general and, in particular, to an increase in leg movements during the milking process.
In the following detailed list, you can find all remarks, comments and questions that should be considered and incorporated in the revised version for improving the manuscript:

ABSTRACT
Line 15-16: Please explain in the text, why do you use two scores for non-lame cows; the Sprecher score 2 indicates a clinically visible mild lameness, therefore it is not conclusive to classify this score as non-lame.

INTRODUCTION
Line 31-32: According to Becker et al. [1], in 2010 the prevalence of lameness in Swiss dairy cows reached about 14.8% on cow level and …  Please add here after prevalence of lameness (with LS ≥ 2 defined as clinical lameness).
Line 51-52: Therefore, to improve lameness detection in modern dairy housing systems, many different technical approaches have already been tested, most of them with the aim to detect lameness completely automatically [17,18]. Here you should add also the recently published approach of Werema CW, Yang DA, Laven LJ, Mueller KR, Laven RA. (2022): Evaluating alternatives to locomotion scoring for detecting lameness in pasture-based dairy cattle in New Zealand. Animals 12 (703); DOI: doi.org/10.3390/ani12060703.
Line 69-72: All the above-listed methods benefit from more or less sophisticated technology because devices have to be attached directly to the cow or need to be installed and integrated in the housing system. As a result, the systems tested may be less suitable for the regular use by farmers during the routine practices, because the maintenance is too complicated.                                                     Sorry, but here I see some contradiction: the methodology described in your study is also not simply available to the farmer, namely the PC-supported evaluation of the accelerometer data from the milking cluster for hindleg activity, the evaluation of the video data for front leg activity & head movements and their statistical calculation with a mixed model. The approach of Werema et al. 2022 seems to be a very practical way. Please rephrase this sentence.
Line 73-74: Because lame cows show increased stepping and weight shifting during prolonged standing compared with non-lame cows [14-16], … Here you should add also the "non-technical" approach of Werema et al. 2022.

MATERIAL & METHODS
Line 139, Table 1: Applied five-point locomotion scoring for dairy cattle after Sprecher et al. [6] …. The Sprecher score 2 (LS 2) describes a clinical lameness that can be assessed during walking of the cow; I recommend to group LS 2-cows also to the "lame" cow" group.
Line 193: … were recorded simultaneously with HLA measurements during one evening milking. This fact, measurements during one milking only in each cow, could be also a limiting aspect of your study; It would have been interesting to see the results after two or three consecutive measurements. Please discuss this aspect.
The authors could use the described statistical analysis in a second approach to calculate how the results change if they also include LS 2 as a lameness score because 249 cows, which accounts for almost one third of all animals examined, showed a locomotion score 2. These new results should then also be discussed.
Do you have data on which cows (ear tag number) entered the milking parlour at which time of the milking period? This information could also be used for statistical calculation, because there are studies that show that lame cows are preferably always the last cows to enter the milking parlour. If so, please insert and discuss this additional information.
RESULTS
Line 275: The difference between the two lameness categories… ;... two lameness categories...; non-lame is not a lameness category; please rephrase: better ... between the two groups ...

DISCUSSION
Line 299: They reported, for example, a lameness prevalence of 23% in Finland [2], 28.2% in the UK [40] and 28.3% in Ontario [3]. Please, indicate always which locomotion scoring method was used by the authors and if LS 2 was regarded as lameness or not; otherwise, these prevalences are very hard to compare.
Line 350: Please indicate how the stepping rate was assessed by Gygax et al. and Cerqueria et al. in their studies, also by an integrated accelerometer in the milking cluster or visually?
Line 414: Here you should also discuss the approach described by Werema et al. 2022.

Comments for author File: Comments.docx

Author Response

Please see the attachement.

Author Response File: Author Response.docx

Reviewer 2 Report

The study entitled “Analysis of dairy cow behavior during milking associated with lameness” is generally a well written study presenting some interesting results. However, this study was not designed as autonomous, and as such has in my opinion some serious flaws that lessen the value of the results. My recommendation is that the authors should strive to correct as many as possible of these issues. Firstly, the authors report in the abstract that they scored 647 cows in 20 farms. Thus, the reader assumes that this is the sample size of the study, which is actually largely misleading as they used data from 356 cows. In my opinion, the sample size that matters is the one analyzed, not what was the initial intention to have. I also do not understand why the authors did not additionally perform lesion scoring instead of only lameness scoring. Lameness scoring is unfortunately a very subjective procedure (which has been verified in many studies) and what is of interest for a farmer is not to detect cows that show not normal walking behavior but to identify cows that have lesions that need prompt treatment (which is not identical to showing lameness). In this context, I do not realize why you have only divided cows in lame vs. non-lame, and not tried to relate lameness score (3,4 and 5) to your measurements. This is crucial to perform in my opinion, because in case lameness 4 or 5 cows do not show increased restlessness, this will be indicative of serious flaw in your procedures. Additionally, the authors themselves are very critical regarding the limitations of the sensors regarding how many and which movements per second is being recorded (L361-400), however, they did not do anything to improve these limitations. For instance, an additional visual measurement regarding both type and intensity of movements would have been very helpful. I also do not understand why the authors have performed only one measurement per cow. Both lameness scoring and readings from the sensor would have been much more representative if performed for instance 3 times within 3 days and the mean would have been used. Repeatability of the measurements is also crucial in my opinion. Perhaps performing a small-scale repeatability study regarding the measurements of the sensors would strengthen your results (in case you find good repeatability). Another very important issue that has to be dealt (and can be performed) is in my opinion the lack of any sensitivity/specificity analysis. This is very important in studies that aim to develop automatic diagnostic tools.

Minor mistakes appear here and there (i.e. L25 and L369, could may). 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Very interesting work regarding dairy cattle production on a vey important issue: Lameness.

Some observations were made in the attached article.

In Methods: The description of hind leg activity, front leg and head movements is too exhaustive.

Comments for author File: Comments.pdf

Some minor English improvement

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The revised version of the manuscript is definitely improved. In my opinion the first figure the authors have provided in the response to my comments (relationship of HLA with lameness scoring, from 1 to 5) is of paramount importance, because it clearly shows (even if there are large deviations that distort the statistics), that HLA increases in a linear (or even perhaps exponential??) manner with increasing lameness score. Thus, the method is reliable (what in my opinion the authors unfortunately failed to fully convince me initially) and more importantly that the method can be very promising if someone invests time to optimize (which is actually why we all do research). Additionally, this figure will probably ease the same concerns another reviewer had regarding your discrimination of cows in lame and non-lame. Thus, I urge the authors to include it in their manuscript or in appendix.  

Author Response

Please see attachment.

Author Response File: Author Response.docx