Loading...
Publication
Desenvolvimento de um modelo de amostragem, com base em risco, aplicado à pesquisa de beta-agonistas em explorações de bovinos, na região Lisboa e Vale do Tejo
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.99 MB | Adobe PDF |
Abstract(s)
Os resíduos químicos nos alimentos de origem animal, ao nível da produção primária,
incluem os medicamentos veterinários e os contaminantes ambientais.
Em Portugal, a vigilância desses resíduos é representada pelo Plano Nacional de Pesquisa
de Resíduos (PNPR), executado pela Direção Geral de Alimentação e Veterinária. Este
controlo encontra-se harmonizado a nível europeu, estando estabelecido que o plano de
amostragem depende do volume de produção animal do ano anterior.
Os princípios da análise de risco aplicados nos planos de vigilância permitem uma
abordagem mais custo-efetiva, ao direcionar as ações de controlo para as subpopulações
de maior risco.
A realização deste estudo teve como principal objetivo o desenvolvimento de um modelo de
amostragem, com base em risco, para monitorizar os resíduos químicos, ao nível da
produção primária, aplicado na seleção de explorações, na região Lisboa e Vale do Tejo.
Na análise retrospetiva do PNPR entre 2013 e 2017, verificou-se que a taxa de não
conformidade foi de 0,66%, destacando-se as substâncias beta-agonistas (2,3%) e
anticoccídeos (2,0%).
O modelo desenvolvido foi particularizado para o grupo de substâncias beta-agonistas em
bovinos, considerando os 25 resultados positivos das 280 colheitas realizadas (8,9%; IC
95%:6,1-12,8%) em 173 explorações, no âmbito do PNPR. As características dessas
explorações e os seus dados relativos ao abate foram avaliados, identificando-se como
principais indicadores de risco de positividade a beta-agonistas a “proporção média de
carcaças com classes de estado de gordura 1 e 2” (p=0,039; OR=74,86; IC 95%:1,91-
7663,0) e o “desvio médio da relação peso/idade” (p=0,035; OR=1,44; IC95%:1,03-2,05) ao
abate.
Uma árvore de decisão para seleção de explorações a incluir no modelo de amostragem,
obtida pelo algoritmo Fast Frugal Trees com as variáveis “número de abates nos últimos 6
meses”, “proporção média de carcaças com classes de estado de gordura 1 e 2” e
“proporção média de carcaças com classes de conformação S,E e U” ao abate, permitiu
elevar a taxa de detenção de 8,9% para 24%, conseguida pelo aumento do número de
casos positivos (n=68) ou pela redução do número de colheitas realizadas (n=103).
Com este trabalho conseguiu-se demonstrar que é possível uma vigilância mais custoefetiva
na pesquisa de resíduos químicos nos géneros alimentícios de origem animal
através de um modelo de amostragem com base no risco.
ABSTRACT - Development of a risk-based sampling model applied for the detection of beta-agonists in cattle farms in the region of Lisboa e Vale do Tejo - Chemical residues in food of animal origin at the level of primary production include veterinary medicine products and environmental contaminants. In Portugal, the surveillance of these residues is represented by the National Residue Control Plan (PNPR) and is performed by the General Directorate of Food and Veterinary. This control is harmonized at European level and it is established that the sampling plan depends on the animal production volume in the preceding year. In order to improve the cost-effectiveness of this approach, the principles of risk analysis are applied to surveillance plans to guide control actions to the higher risk subpopulations. The main objective of this study was to develop a risk-based sampling model for monitoring chemical residues at the primary production level, applied to herds selection in the region of Lisboa e Vale do Tejo. In the retrospective analysis of PNPR data from 2013 to 2017, it was found that the rate of non-compliance was 0.66% and the substances with the highest percentage of noncompliant results were beta-agonists (2.3%) and anticoccids (2.0%). The developed model was particularized for the group of beta-agonist substances in cattle, considering the 25 positive results in 280 analysed PNPR samples (8.9%, 95%CI: 6.1- 12.8%) in 173 cattle farms. The characteristics and the slaughter register data of these farms were analysed, and the main risk indicators on the use of beta-agonists in cattle farms were identified: “average proportion of carcasses with fat cover classes 1 and 2” (p = 0.039; OR=74.86; 95%CI: 1.91-7663.0) and the “average deviation in weight/age” (p = 0.035; OR=1.44; 95%CI: 1.03-2.05) at slaughter. One herd selection decision tree to be included for sampling model, obtained by the Fast and Frugal Tree algorithm with the variables “number of slaughterings in the last 6 months”, “average proportion of carcasses with fat cover classes 1 and 2” and “average proportion of carcasses with conformation classes S, E and U” at slaughter, allowed the increase of the detention rate from 8.9% to 24%, that can be achieved by increasing the number of positive cases (n=68) or reducing the number of samples (n=103). This work demonstrates that is possible to increase the cost-effectiveness of chemical residues surveillance in foodstuffs of animal origin using a risk-based sampling model.
ABSTRACT - Development of a risk-based sampling model applied for the detection of beta-agonists in cattle farms in the region of Lisboa e Vale do Tejo - Chemical residues in food of animal origin at the level of primary production include veterinary medicine products and environmental contaminants. In Portugal, the surveillance of these residues is represented by the National Residue Control Plan (PNPR) and is performed by the General Directorate of Food and Veterinary. This control is harmonized at European level and it is established that the sampling plan depends on the animal production volume in the preceding year. In order to improve the cost-effectiveness of this approach, the principles of risk analysis are applied to surveillance plans to guide control actions to the higher risk subpopulations. The main objective of this study was to develop a risk-based sampling model for monitoring chemical residues at the primary production level, applied to herds selection in the region of Lisboa e Vale do Tejo. In the retrospective analysis of PNPR data from 2013 to 2017, it was found that the rate of non-compliance was 0.66% and the substances with the highest percentage of noncompliant results were beta-agonists (2.3%) and anticoccids (2.0%). The developed model was particularized for the group of beta-agonist substances in cattle, considering the 25 positive results in 280 analysed PNPR samples (8.9%, 95%CI: 6.1- 12.8%) in 173 cattle farms. The characteristics and the slaughter register data of these farms were analysed, and the main risk indicators on the use of beta-agonists in cattle farms were identified: “average proportion of carcasses with fat cover classes 1 and 2” (p = 0.039; OR=74.86; 95%CI: 1.91-7663.0) and the “average deviation in weight/age” (p = 0.035; OR=1.44; 95%CI: 1.03-2.05) at slaughter. One herd selection decision tree to be included for sampling model, obtained by the Fast and Frugal Tree algorithm with the variables “number of slaughterings in the last 6 months”, “average proportion of carcasses with fat cover classes 1 and 2” and “average proportion of carcasses with conformation classes S, E and U” at slaughter, allowed the increase of the detention rate from 8.9% to 24%, that can be achieved by increasing the number of positive cases (n=68) or reducing the number of samples (n=103). This work demonstrates that is possible to increase the cost-effectiveness of chemical residues surveillance in foodstuffs of animal origin using a risk-based sampling model.
Description
Dissertação de Mestrado em Segurança Alimentar
Keywords
Resíduos vigilância análise de risco amostragem fatores de risco residues surveillance risk analysis sampling risk factors
Pedagogical Context
Citation
Leal DRBS 2020. Desenvolvimento de um modelo de amostragem, com base em risco, aplicado à pesquisa de beta-agonistas em explorações de bovinos, na região Lisboa e Vale do Tejo [dissertação de mestrado]. Lisboa: FMV-Universidade de Lisboa
Publisher
Universidade de Lisboa, Faculdade de Medicina Veterinária