Hall David C, Le Quynh B
Department of Ecosystem and Public Health, Faculty of Veterinary Medicine, University of Calgary, Calgary, Alberta, T2N 4N1, Canada.
Trans R Soc Trop Med Hyg. 2017 Jun 1;111(6):270-277. doi: 10.1093/trstmh/trx043.
More than 70 million Vietnamese rely on small-scale farming for some form of household income. Water on many of those farms is contaminated with waste, including animal manure, partly due to non-sustainable waste management. This increases the risk of water-related zoonotic disease transmission. The purpose of this research was to examine the impact of various demographic and management factors on the likelihood of finding Escherichia coli in drinking water sourced from wells and rainwater on farms in Vietnam.
A Bayesian Belief Network (BBN) was designed to describe association between various deterministic and probabilistic variables gathered from 600 small-scale integrated (SSI) farmers in Vietnam. The variables relate to E. coli content of their drinking water sourced on-farm from wells and rainwater, and stored in on-farm large vessels, including concrete water tanks. The BBN was developed using the Netica software tool; the model was calibrated and goodness of fit examined using concordance of predictability.
Sensitivity analysis of the model revealed that choice variables, including engagement in mitigation of water contamination and livestock management activities, were particularly likely to influence endpoint values, reflecting the highly variable and impactful nature of preferences, attitudes and beliefs relating to mitigation strategies. Quantitative variables including numbers of livestock (particularly chickens) and income also had a high impact. The highest concordance (62%) was achieved with the BBN reported in this paper.
This BBN model of SSI farming in Vietnam is helpful in understanding the complexity of small-scale agriculture and how various factors work in concert to influence contamination of on-farm drinking water as indicated by the presence of E. coli. The model will also be useful for identifying and estimating the impact of policy options such as improved delivery of clean water management training for rural areas, particularly where such analysis is combined with other analytical and policy tools. With appropriate knowledge translation, the model results will be particularly useful in helping SSI farmers understand their options for engaging in public health mitigation strategies addressing clean water that do not significantly disrupt their agriculture-based livelihoods.
超过7000万越南人依靠小规模农业获取某种形式的家庭收入。这些农场中的许多水源都受到包括动物粪便在内的废物污染，部分原因是废物管理不可持续。这增加了与水相关的人畜共患病传播风险。本研究的目的是调查各种人口统计学和管理因素对越南农场井水和雨水饮用水中发现大肠杆菌可能性的影响。
设计了一个贝叶斯信念网络（BBN）来描述从越南600名小规模综合（SSI）农民收集的各种确定性和概率性变量之间的关联。这些变量与他们从农场水井和雨水中获取并储存在农场大型容器（包括混凝土水箱）中的饮用水中的大肠杆菌含量有关。使用Netica软件工具开发BBN；使用可预测性的一致性对模型进行校准并检验拟合优度。
模型的敏感性分析表明，选择变量，包括参与水污染缓解和牲畜管理活动，特别有可能影响终点值，反映了与缓解策略相关的偏好、态度和信念的高度变异性和影响力。包括牲畜数量（特别是鸡）和收入在内的定量变量也有很大影响。本文报告的BBN达到了最高一致性（62%）。
越南SSI农业的这个BBN模型有助于理解小规模农业的复杂性，以及各种因素如何协同作用以影响农场饮用水污染（以大肠杆菌的存在为指标）。该模型还将有助于识别和估计政策选项的影响，例如改善农村地区清洁水管理培训的提供，特别是在这种分析与其他分析和政策工具相结合的情况下。通过适当的知识转化，模型结果将特别有助于SSI农民了解他们参与解决清洁水问题的公共卫生缓解策略的选项，而不会严重扰乱他们以农业为基础的生计。
