Agent-Based Modeling as a Tool for the Prevention of Early Childhood Diarrhea

EPA Grant Number: FP917286
Title: Agent-Based Modeling as a Tool for the Prevention of Early Childhood Diarrhea
Investigators: Mellor, Jonathan E
Institution: University of Virginia
EPA Project Officer: Michaud, Jayne
Project Period: August 1, 2011 through July 31, 2014
Project Amount: $126,000
RFA: STAR Graduate Fellowships (2011) RFA Text |  Recipients Lists
Research Category: Fellowship - Emerging Environmental Approaches and Challenges: Information Science , Academic Fellowships


Early childhood diarrhea (ECD) causes nearly 1.5 million child deaths worldwide each year. But the causes of and the prevention strategies for ECD are numerous and complex. Therefore, the purpose of this research is to create an agentbased model (ABM) to simulate the transmission of harmful pathogens that are the primary cause of ECD in much of the developing world. That model can then be used to predict which interventions will be most effective at stopping ECD and to explore the complex relationships related to this transmission.


The ABM, which is inherently transdisciplinary, will be validated and calibrated using field data collected from two adjacent communities in Limpopo Province, South Africa. Those communities are being studied as part of a long-standing partnership between the University of Virginia and the University of Venda in Thohoyandou, South Africa. Data about pathogen sources as well as sanitation and behavioral practices will be gathered using community surveys, microbiological testing and pathogen growth, decay and transmission measurements. These data will inform the development of the geo-spatial ABM in which the “agents” will be the community members who can make decisions about their adherence to sound water and sanitation practices. Varying input parameters such as water source pathogen levels will elucidate which interventions may be most effective at reducing diarrhea.

Expected Results:

Preliminary field data have suggested that pathogen levels in household drinking water storage containers may be significantly higher than community water sources. This suggests that point-of-use water treatment interventions such as the introduction of household ceramic water filters might be the most beneficial. However, ABM modeling frequently reveals unexpected, emergent outcomes that could not have been predicted based on model inputs. Therefore, a full behavior space analysis will be conducted to search for such emergent outcomes.

Potential to Further Environmental / Human Health Protection

The model can be an easy-to-use tool for researchers, government agencies and humanitarian organizations worldwide and will provide valuable insight into the complex pathogen transmission chain that leads to ECD. This insight will inform the development of intervention strategies that can be broadly applied to similar settings throughout the developing world.

Supplemental Keywords:

diarrhea, agent-based modeling, ABM, coliform, water, sanitation, hygiene, Africa, health, transdisciplinary

Progress and Final Reports:

  • 2012
  • 2013
  • Final