Date of Award

Spring 5-10-2018

Degree Type


Degree Name

Master of Public Health (MPH)


Public Health

First Advisor

Gerardo Chowell

Second Advisor

Kimberlyn Roosa



Due to a lack of sanitary infrastructure and a highly susceptible population, Peru experienced a historic outbreak of Vibrio cholerae O1 that began in 1991 and generated multiple waves of disease for several years. Though case-fatality was low, the epidemic put massive strain on healthcare and governmental resources. Here we explore the transmission dynamics and spatiotemporal variation of cholera in Peru using mathematical models and statistical analyses that account for environmental conditions favoring the persistence of bacteria in the environment.


The authors use dynamic transmission models that incorporate seasonal variation in temperature, concentration of vibrios in the environment, as well as separate human and environmental transmission pathways. The model is fit to weekly department level data obtained from the cholera surveillance system in Peru. The authors also assess the spatial patterns of cholera transmission and correlations between case incidence, time of epidemic onset, and department level variables. Reproductive numbers are compared across departments.


Our findings indicate that the epidemic first hit the coastal departments of Peru and later spread through the highlands and jungle regions. There was high seasonal variation in case incidence, with three clear waves of transmission corresponding to the warm seasons in Peru. Department level variables such as population size and elevation also played a role in transmission patterns. Finally, basic reproductive numbers most often ranged from one to eleven depending on department and time of year. Lima had the largest reproductive number, likely due to its population density and proximity to the coast.


Incorporating environmental variables into an epidemic model predicts the multiple waves of transmission characteristic of \textit{V. cholerae}, and effectively differentiates transmission patterns by geographic region even in the absence of unique parameter estimates. Mathematical models can provide valuable information about transmission patterns and should continue to be used to inform public health decision making.