Australian Livestock Scanning Services Group

12. Methods of handling ewes for scanning and scanning accuracy

The way in which sheep are handled for scanning influences the likelihood of errors (Wolfe et al 1985). Direct extracts from this study, which was commissioned by the Australian Meat and Livestock Research and Development Corporation (AMLR&DC), are shown hereunder in italic text. The report is held at 10 different libraries throughout Australia.

The orientation of ewes for scanning should be such that it allows the best possible images for ease and accuracy of diagnosis. However, we must also consider the handling aspects of delivery and presentation as these will affect the rate of throughput and the range of equipment that might be used to assist the operation. If there are any compromises to be made in sacrificing accuracy or image clarity for increased speed, the consequences should be carefully assessed. (Wolfe et al 1985:page 47).

Given that there are a number of commercially available machines that are used for handling sheep, the purpose of this part of the study was not to compare machines but rather to compare the manner in which the ewe was presented at the point where the scan was to be conducted.

We designed an experiment to examine the image clarity and accuracy of diagnoses made when ewes were scanned in a variety of positions. These positions were chosen to cover the range of likely variations in presentation that might be possible (Wolfe et al 1985 page 47).

scanning1 image scanning1 image scanning1 image

Experimental procedure
     In this experiment we used a group of Merino and crossbred (Border Leicester x Merino) ewes. They were synchronized for mating during June/July 1982 using a standard progestagen treatment and half were treated with P.M.S.G. (Pregnecol®) to ensure a high proportion of multiple pregnancies.” Empty ewes were excluded from the experiment.
    Approximately 50 ewes were used in each of 6 position treatments. They were scanned by each of 2 operators on each of three occasions being randomly allocated between treatments on each occasion. The stages of pregnancy at these times were 45 – 54, 66 – 75, and 87 – 96 post conception. The positions presented to the operator for scan were presented at random.
     The positions examined were as follows:- 1 ewe lying on her back (horizontal) restrained in a cradle; 2 as above with the lumbar region raised by a soft support; 3 held in a cradle, inclined with head up; 4 inclined in cradle , inclined with rump up; 5 standing upright, held in a bail; 6 held on her side in a cradle. (Wolfe et al 1985 page 47).

The results of the study are detailed in full within the report. The data were analysed using appropriate statistical procedures and the results are presented in the report. The following section draws attention to the key findings of the project.

Results and discussion
The treatment effects [the various positions in which the ewes were presented for scanning ed.] can be seen by comparison of means within rows of the above table [2 tables of results are presented ed.] Many of the treatment differences in these analyses were not statistically significant due to low error rates overall. The treatment effects must therefore be judged for their consistency and absolute levels which are reflected in their rankings (shown later). In considering the single/multiple classification (Table 43), positions 1 and 2 had the most consistently low error rates overall with no values exceeding 4.5%. In contrast positions 5 and 6 were consistently highest, with 5 being the worst [with error rates of the order of 15 to 17 %] of the two. Error rates for positions 3 and 4 varied, being zero for both operators at the intermediate stage, but having several values exceeding 5% at the early and late stages. (Wolfe et al 1985 page 48).
Error rates in Table 44, when diagnoses of litter sizes of 3 or more were required, showed an almost identical pattern to the above. Accuracy was most affected by position at the early stage. Positions 1 and 2 again had the most consistently low error rates. (Wolfe et al 1985 page 49).

In this study, 2 independent operators have found that when sheep are scanned in the standing position there are error rates of the order of 15% to 17%.

Wolfe, E.C., Robards, G.E., Fowler, D.G., Wilkins, J.F., Chick, B.B., Weissel, D.A., (1985) Developing a Field Technique for Determining Litter Number during Pregnancy in Sheep. N.S.W. Department of Agriculture, Miscellaneous bulletin/Division of Animal Production 120pp