dice-package: DICE: Dynamics of Interacting Community Epidemics

In predsci/DICE: DICE: Dynamics of Interacting Community Epidemics

Description

An arbitrary scale global Modeling Framework for modeling the spread of influenza and other infectious diseases.

The DICE package implements a global model of influenza within civilian populations with arbitrary spatial resolution, with the objective of making best use of available data at different spatial scales. DICE uses a deterministic compartmental S-I-R model, applied on various interacting spatial scales, and a robust Markov-Chain-Monte-Carlo (MCMC) fitting procedure that can quickly characterize Influenza-Like-Illness (ILI) incidence profiles, of either coupled or uncoupled regions, in real time. Upon completing a run DICE provides estimates for the epidemic transmissibility and severity.

Publicly available weekly ILI data is included in the package (from the CDC and Google Flu Trends), along with weekly averaged specific humidity and school opening/closing schedules. The last two sets of data are needed for three of the five different models for the time dependence of the basic reproduction number that the user can choose from. These data sets are provided on spatial scales of various, increasing, resolutions.

Upon completion of the MCMC fitting procedure, the DICE package analyzes the results and produces an extensive set of publication-quality plots (in a png or pdf format) and tables (in csv format). The complete history of the run and the fitting procedure is saved as an RData file which the user can later load and further analyze.

Details

Package:	DICE
Type:	Package
Version:	2.0
Date:	2017-12-24
License:	Currently Used by PSI and Leidos Only

Author(s)

Predictive Science Inc.

Maintainer: Michal Ben-Nun <[e-mail]>

References

Ben-Nun, M., Riley, P., Turtle, J., Bacon, D. P., and Riley, S., DICE: Dynamics of Interacting Community Epidemics, to be submitted to PLOS – Computational Biology, 2018.

Riley P, Ben-Nun M, Armenta R, Linker JA, Cost AA, et al. (2015) Early Characterization of the Severity and Transmissibility of Pandemic Influenza Using Clinical Episode Data from Multiple Populations. PLoS Comput Biol http://dx.doi.org/10.1371/journal.pcbi.1004392

Riley P, Ben-Nun M, Armenta R, Linker JA, Eick AA, et al. (2013) Multiple Estimates of Transmissibility for the 2009 Influenza Pandemic Based on Influenza-like-Illness Data from Small US Military Populations. PLoS Comput Biol 9: e1003064. doi:10.1371/journal.pcbi.1003064. http://dx.doi.org/10.1371/journal.pcbi.1003064

Centers for Disease Control and Prevention Influenza Surveillance: http://www.cdc.gov/flu/weekly/fluactivitysurv.htm

Google Flu Trends: http://www.google.org/flutrends/us/

See Also

The DICE manual

Examples

For a DICE run of the 2015-2016 cdc national data using the ten HHS regions with coupling between the regions use:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2015, mod_level = 2, fit_level = 3, isingle = 0)

For a DICE run of the 2015-2016 cdc national data using the ten HHS regions without coupling between the regions use:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2015, mod_level = 2, fit_level = 3, isingle = 1)

For a DICE run of the 2014-2015 GFT data for HHS region number 9 RegState = 9 using state level data with coupling between the states in region 9 use:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2014, mod_level = 3, fit_level = 4, RegState = 9, isingle = 0)

To control which model is used for the basic reproduction number R0, set the parameter model in your call. Default value is 5:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2014, mod_level = 3, fit_level = 4, RegState = 9, isingle = 0, model = 3)

To control the number of MCMC chains that the code will run set the parameter nreal in your call, default is 1:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2015, mod_level = 2, fit_level = 3, isingle = 0, nreal = 3)

To control the number of MCMC steps/trial in each chain set the parameter nMCMC in your call, default is 1e5:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2015, mod_level = 2, fit_level = 3, isingle = 0, nMCMC = 1e6)

To control the name of the sub-directory where all the output files and plots are saved use the keyword subDir, default is output:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2015, mod_level = 2, fit_level = 3, isingle = 0, nMCMC = 1e6, subDir = 'test')

To control the file format for the plots (pdf, png or x11) set the parameter device:

output <- runDICE(diseae = 'flu', data_source='cdc', year = 2015, mod_level = 2, fit_level = 3, isingle = 0, nMCMC = 1e6, device = 'pdf')
(The package can accept an array of file formats, i.e. device = c('pdf','png'), in which case more both 'png' and 'pdf' files will be created.)

To run DICE in a forecast or predictive mode you can set the number of weeks the code uses in the fit to be lower than the number of weeks in the season.
(Note that for the current season it is always running in a predictive mode because the season is not yet completed.)

output <- runDICE(disease = 'flu', data_source='gft', year = 2013, mod_level = 3, fit_level = 4, isingle = 1, nMCMC = 1e6, nfit = 35)

To select only  a few HHS regions and run them coupled (for example using the three east coast regions: 1, 2 and 3) use:
output <- runDICE(disease = 'flu', data_source="cdc", year=2015, mod_name=c(NAME_2="US"), mod_level=2, fit_level=3, fit_names=c("R1","R2","R3"), isingle = 0)

To select only a few states and run them coupled  use for example:
output <- runDICE(disease = 'flu', data_source="cdc",year=2014, mod_name=c(NAME_2="US"), mod_level=2, fit_level=4, fit_names=c("WA","OR","CA"), isingle = 0)

predsci/DICE documentation built on Aug. 9, 2019, 9:41 a.m.