Accuracy and Precision of Sampling for Worm Egg Count Estimates

Abstract

Background: Reliance on faecal worm egg counts (WEC) as the basis for tactical anthelmintic treatment is a critical part of integrated worm control programs for sheep. The importance of WEC highlights the need for sampling that provides a good estimation of the true group mean. The typical approach used in Australia is the collection of ten faecal samples from a mob of sheep which could typically contain 200-1000 sheep. The collection of ten samples under these situations accounts for 1-5% of the mob. This study was conducted to test the hypothesis that ten samples may be insufficient to provide an accurate and precise estimate of the true group mean WEC. Methods: Faecal samples were collected on four occasions from all sheep in a single mob of 500 adult Merino ewes in the Northern Tablelands of NSW, Australia. Sampling occurred in 2003 on the following dates: 14th August, 2nd September, 17th September and 2nd November. WEC was determined (1 egg = 60 epg) for each sheep and data entered into a spreadsheet. Ewes were treated with napthalophos and levamisole at recommended rates on 17th September and lambed over a 5-week period starting 24th September. For each date, the requisite number of animals to achieve 4 (20 sheep), 8 (40) and 12% (60) sampling was randomly chosen in-silico and this process repeated a further 4 times to provide 5 random selections from each sampling date. The accuracy and precision of the WEC estimate for each selection group (i.e. 4, 8 or 12%) was calculated. Results: Group mean WEC was 563, 1612, 2605 and 805 epg respectively for the four sampling dates with 'Haemonchus contortus' contributing 43, 54, 70 and 74%. In relation to the true group mean, accuracy of WEC estimates increased from 69 to 73 and to 81% and precision improved from 19 to 17 and to 12% respectively for the three sampling levels. Conclusions: Sampling of 10 animals from commercial mobs may not provide an accurate and precise estimate of the true group WEC mean. Increased level of sampling is more accurate and precise but may not always be practical. The choice of sampling 40 animals may be a practical compromise as it provides for a sampling level of 4-20% in most instances.