Assessment of a hands-on method for FAMACHA© system training

Abstract

The FAMACHA©system is a method for selective anthelmintic treatment comprising earlydetection of haemonchosis in sheep and goats. In order to evaluate the hands-on trainingmethodology and the learning level of the participants, we analyzed data from 30 trainingevents involving 47 training classes conducted in the State of Paraná, Brazil, from July/2009to May/2011, during which period a total of 1004 participants did 20,080 FAMACHA©clas-sifications. In the practical training sessions, each participant individually evaluated 20animals with known haematocrit values. Every participant per training event was given aunique number, whereupon each of the animals in a given event was FAMACHA©classifiedby all the trainees involved, in the same trainee number sequence. After each consecutiveanimal had been evaluated by every one of the participants, its haematocrit and corre-sponding FAMACHA©category were announced before the next animal was presented. Thenumber of persons in training, which ranged from 5 to 39 per session, did not significantlyaffect the average error of the groups of participants involved (p > 0.05). The average errorin the classification of the first animal on a scale with a perfect score of zero was 2.5, signif-icantly greater than the error of 0.56 of the twentieth one (p < 0.05), indicating an inverserelationship between the error and the cumulative number of animals already evaluatedby each trainee involved, with the reduction in mean error per animal in a given trainingevent found by linear regression to be 0.0713. When the same animal was assessed twice inthe same training event, the average error of the second evaluation (1.05) was significantlylower than the 1.70 of the first (p < 0.05). While the total of 686 sheep used in the trainingevents (73%) was considerably larger than the corresponding number of 254 goats (27%),the average statistical errors, respectively, 1.34 and 1.23, were not significantly different(p > 0.05). Similarly, the average errors in FAMACHA©classification were not significantlyinfluenced by the occupation or gender of the participants, nor whether there were ani-mals in all five FAMACHA©categories or only in categories 1, 2, 3 and 4 per training event(p > 0.05).