Translation of risk factor estimates into on-farm interventions and their effect on Campylobacter broiler flock prevalence

Helle Mølgaard Sommer, Maarten Nauta, Hanne Rosenquist

Research output: Contribution to journalJournal articleResearchpeer-review

Abstract

Before deciding upon interventions to control Campylobacter in broiler flocks, it would be useful to estimate the potential effects of different interventions. Certain previously identified risk factors for colonization of broiler flocks with Campylobacter may seem to have large impact on the broiler flock prevalence. Nevertheless, interventions related to these risk factors may have only limited effect on the overall prevalence estimate, since in practice only a relatively small fraction of farms are actually amenable for an intervention related to a given risk factor.We present a novel method for the risk assessor that predicts effects of interventions at the farm, based on results from a risk factor study that included data from six European countries (Denmark, Netherlands, Norway, Poland, Spain and United Kingdom). In the present study, five previously identified risk factors, which had shown to have significant impact on Campylobacter flock prevalence, were translated into practical on-farm interventions. Given the implementation of these interventions the population prevalence was predicted by developing and using a statistical method anchored in the ideas behind standardized population estimations using logistic regression. To obtain population estimates per country, the predicted prevalence values were multiplied by the frequencies (no. of farms) in a reference population based on data from the risk factor study and a large questionnaire. The latter was included to improve the representativeness of the reference population. Population prevalence estimates were calculated before and after implementation of a given intervention in the six countries. Results showed that if biosecurity was not accounted for, some individual interventions resulted in a limited reduction of the population prevalence. The reduction differed between countries depending on the current farm management practices and the actual flock prevalence level. In general, the most effective interventions were "building new houses with strict biosecurity for all houses older than 15 years" and "apply drinkers with nipples without cups". In conclusion, the novel method translates results from risk factor studies into effects of on-farm interventions for the reduction of the prevalence of Campylobacter in broiler flocks. The method is very useful for providing the basis for risk management decisions. The usefulness would improve further when the results are integrated with costs of interventions in a cost effectiveness study. The approach was developed for Campylobacter in broiler flocks, but it can also be applied to other pathogens and other farm animals, given that the required data are available.
Original languageEnglish
JournalMicrobial Risk Analysis
Volume2-3
Pages (from-to)27-37
Number of pages11
ISSN2352-3522
DOIs
Publication statusPublished - 2016

Keywords

  • European study
  • Intervention
  • Logistic regression
  • On-farm Campylobacter control
  • Standardized population estimate
  • Translation model

Cite this

@article{486ad971e94b493886e8e7ab11d87dea,
title = "Translation of risk factor estimates into on-farm interventions and their effect on Campylobacter broiler flock prevalence",
abstract = "Before deciding upon interventions to control Campylobacter in broiler flocks, it would be useful to estimate the potential effects of different interventions. Certain previously identified risk factors for colonization of broiler flocks with Campylobacter may seem to have large impact on the broiler flock prevalence. Nevertheless, interventions related to these risk factors may have only limited effect on the overall prevalence estimate, since in practice only a relatively small fraction of farms are actually amenable for an intervention related to a given risk factor.We present a novel method for the risk assessor that predicts effects of interventions at the farm, based on results from a risk factor study that included data from six European countries (Denmark, Netherlands, Norway, Poland, Spain and United Kingdom). In the present study, five previously identified risk factors, which had shown to have significant impact on Campylobacter flock prevalence, were translated into practical on-farm interventions. Given the implementation of these interventions the population prevalence was predicted by developing and using a statistical method anchored in the ideas behind standardized population estimations using logistic regression. To obtain population estimates per country, the predicted prevalence values were multiplied by the frequencies (no. of farms) in a reference population based on data from the risk factor study and a large questionnaire. The latter was included to improve the representativeness of the reference population. Population prevalence estimates were calculated before and after implementation of a given intervention in the six countries. Results showed that if biosecurity was not accounted for, some individual interventions resulted in a limited reduction of the population prevalence. The reduction differed between countries depending on the current farm management practices and the actual flock prevalence level. In general, the most effective interventions were {"}building new houses with strict biosecurity for all houses older than 15 years{"} and {"}apply drinkers with nipples without cups{"}. In conclusion, the novel method translates results from risk factor studies into effects of on-farm interventions for the reduction of the prevalence of Campylobacter in broiler flocks. The method is very useful for providing the basis for risk management decisions. The usefulness would improve further when the results are integrated with costs of interventions in a cost effectiveness study. The approach was developed for Campylobacter in broiler flocks, but it can also be applied to other pathogens and other farm animals, given that the required data are available.",
keywords = "European study, Intervention, Logistic regression, On-farm Campylobacter control, Standardized population estimate, Translation model",
author = "Sommer, {Helle M{\o}lgaard} and Maarten Nauta and Hanne Rosenquist",
year = "2016",
doi = "10.1016/j.mran.2016.06.001",
language = "English",
volume = "2-3",
pages = "27--37",
journal = "Microbial Risk Analysis",
issn = "2352-3522",
publisher = "Elsevier",

}

Translation of risk factor estimates into on-farm interventions and their effect on Campylobacter broiler flock prevalence. / Sommer, Helle Mølgaard; Nauta, Maarten; Rosenquist, Hanne.

In: Microbial Risk Analysis, Vol. 2-3, 2016, p. 27-37.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Translation of risk factor estimates into on-farm interventions and their effect on Campylobacter broiler flock prevalence

AU - Sommer, Helle Mølgaard

AU - Nauta, Maarten

AU - Rosenquist, Hanne

PY - 2016

Y1 - 2016

N2 - Before deciding upon interventions to control Campylobacter in broiler flocks, it would be useful to estimate the potential effects of different interventions. Certain previously identified risk factors for colonization of broiler flocks with Campylobacter may seem to have large impact on the broiler flock prevalence. Nevertheless, interventions related to these risk factors may have only limited effect on the overall prevalence estimate, since in practice only a relatively small fraction of farms are actually amenable for an intervention related to a given risk factor.We present a novel method for the risk assessor that predicts effects of interventions at the farm, based on results from a risk factor study that included data from six European countries (Denmark, Netherlands, Norway, Poland, Spain and United Kingdom). In the present study, five previously identified risk factors, which had shown to have significant impact on Campylobacter flock prevalence, were translated into practical on-farm interventions. Given the implementation of these interventions the population prevalence was predicted by developing and using a statistical method anchored in the ideas behind standardized population estimations using logistic regression. To obtain population estimates per country, the predicted prevalence values were multiplied by the frequencies (no. of farms) in a reference population based on data from the risk factor study and a large questionnaire. The latter was included to improve the representativeness of the reference population. Population prevalence estimates were calculated before and after implementation of a given intervention in the six countries. Results showed that if biosecurity was not accounted for, some individual interventions resulted in a limited reduction of the population prevalence. The reduction differed between countries depending on the current farm management practices and the actual flock prevalence level. In general, the most effective interventions were "building new houses with strict biosecurity for all houses older than 15 years" and "apply drinkers with nipples without cups". In conclusion, the novel method translates results from risk factor studies into effects of on-farm interventions for the reduction of the prevalence of Campylobacter in broiler flocks. The method is very useful for providing the basis for risk management decisions. The usefulness would improve further when the results are integrated with costs of interventions in a cost effectiveness study. The approach was developed for Campylobacter in broiler flocks, but it can also be applied to other pathogens and other farm animals, given that the required data are available.

AB - Before deciding upon interventions to control Campylobacter in broiler flocks, it would be useful to estimate the potential effects of different interventions. Certain previously identified risk factors for colonization of broiler flocks with Campylobacter may seem to have large impact on the broiler flock prevalence. Nevertheless, interventions related to these risk factors may have only limited effect on the overall prevalence estimate, since in practice only a relatively small fraction of farms are actually amenable for an intervention related to a given risk factor.We present a novel method for the risk assessor that predicts effects of interventions at the farm, based on results from a risk factor study that included data from six European countries (Denmark, Netherlands, Norway, Poland, Spain and United Kingdom). In the present study, five previously identified risk factors, which had shown to have significant impact on Campylobacter flock prevalence, were translated into practical on-farm interventions. Given the implementation of these interventions the population prevalence was predicted by developing and using a statistical method anchored in the ideas behind standardized population estimations using logistic regression. To obtain population estimates per country, the predicted prevalence values were multiplied by the frequencies (no. of farms) in a reference population based on data from the risk factor study and a large questionnaire. The latter was included to improve the representativeness of the reference population. Population prevalence estimates were calculated before and after implementation of a given intervention in the six countries. Results showed that if biosecurity was not accounted for, some individual interventions resulted in a limited reduction of the population prevalence. The reduction differed between countries depending on the current farm management practices and the actual flock prevalence level. In general, the most effective interventions were "building new houses with strict biosecurity for all houses older than 15 years" and "apply drinkers with nipples without cups". In conclusion, the novel method translates results from risk factor studies into effects of on-farm interventions for the reduction of the prevalence of Campylobacter in broiler flocks. The method is very useful for providing the basis for risk management decisions. The usefulness would improve further when the results are integrated with costs of interventions in a cost effectiveness study. The approach was developed for Campylobacter in broiler flocks, but it can also be applied to other pathogens and other farm animals, given that the required data are available.

KW - European study

KW - Intervention

KW - Logistic regression

KW - On-farm Campylobacter control

KW - Standardized population estimate

KW - Translation model

U2 - 10.1016/j.mran.2016.06.001

DO - 10.1016/j.mran.2016.06.001

M3 - Journal article

VL - 2-3

SP - 27

EP - 37

JO - Microbial Risk Analysis

JF - Microbial Risk Analysis

SN - 2352-3522

ER -