Synthetic populations (demes) for epidemiological research
For a simple scenario run, have Java SE Platform 8+ installed, along with Maven 3+ and Git, or your favorite IDE which includes them, e.g. Eclipse EE or NetBeans,= (possibly extended with tooling for YAML, Docker, Gradle, etc.) then follow these steps:
git clone https://github.com/krevelen/epidemes.git
cd epidemes/java
mvn install -Dmaven.test.skip=true
java -jar epidemes-demo/target/epidemes-demo.jar
DOCKERIZE!
For the purpose of creating a census data driven synthetic Dutch population fr epidemiological research and policy support, this project applies recent advances with respect to simulating household demographics (Geard et al., 2013) and spatial contact patterns in large-scale environments (Zhang et al., 2016). Furthermore the Four C model of individual vaccination hesitancy behavior (Betsch et al., 2015) has been formalized and implemented to study the effects of national vaccination programs under various conditions.
Related work includes for instance, POHEM (Hennessy et al., 2015), FRED (Grefenstette et al., 2013), EpiSimS/OPPIE (Del Valle et al., 2012), and EpiFast (Bisset et al., 2009).
The following model is derived loosely from definitions given by Betsch et al. (2015) and validated with data vaccination attitudes and behaviors obtained from several studies, including a large-scale serosurveillance study held among the Dutch population (ISRCTN 20164309, Van der Klis et al., 2009).
Data from these studies show that for the Dutch situation, convenience of vaccination events is considered to have little effect on vaccination behaviors, since the events are conveniently combined with regular visits to municipal health services aimed at tracking and supporting individual welfare of newborns and infants. Furthermore, the data show that mental barriers against vaccination are strongly increased by certain protestant religious convictions, whereas alternative medical and health convictions (e.g. homeopathy as inspired by Hahneman, and anthroposophy as inspired by Steiner) appear to have a more nuanced correlation with individual vaccination hesitancy.
The Four C model formalization applied here is as follows. Given an individual i having several opinions and/or experiences j from the social network (including self), with weight w representing the individual's appreciation of the opinion's source exceeding the individual's minimum calculation threshold for inclusion, the mental barrier resulting from a weighted product of each source's confidence in the vaccine system and/or the source's complacency in disease risk absence will lead to hesitancy (i.e. vaccination delay) iff it is not sufficiently relieved by the convenience of the vaccination event e at hand:
hesitancy(i,e): product_j(w_j ^ confidence_i,j[w_j > calculation_i] * w_j ^ -complacency_i,j[w_j > calculation_i]) > convenience_i,e