/pcv-impact

An ensemble model for estimating health impact from time series data

Primary LanguageR

PCV Impact Analysis Using Data from Manhica DSS

Basic approach: interrupted time series analysis with Bayesian model averaging

Instructions:

  1. Put all data files in [some directory]/data/input
  2. Clone this repository to your local computer
  3. Set current working directory (in R) to the location of this readme
  4. source('impact_analysis.r')
  5. impactAnalysis(root='[some directory]', rePrepData=TRUE)
    • If no root is specified, it will assume the data are in ./data/input
    • The default is to not re-prep the data, after running it once with rePrepData=TRUE, you shouldn't need to do that again (unless you want to re-prep the data)
Code structure:

  1. impact_analysis.r

    Main script that executes full analysis, has a battery of arguments to control behavior (see script) Outputs:

    • .rdata files in /data/output
    • .pdf graph files in /visualizations

  2. prepData.r

    Code that loads/preps data. Inputs:

    • dir - directory where data live

    Outputs:

    • data - all datasets in 'prepped' format (below)

  3. its.r

    Function that carries out interrupted time series analysis.
    Inputs:

    • data - data table object in 'prepped' format (see below)
    • outcome - character. name of the outcome variable
    • cutpoint - date object containing the time point or points (up to 2) of intervention
    • slope - logical. TRUE indicates that an interaction term (or terms) should be used to estimate a different slope before/after intervetion

    Outputs (in a list):

    • data - the input data object with six new columns: [outcome]_pred, [outcome]_pred_upper, [outcome]_pred_lower, [outcome]_cf, [outcome]_cf_upper, [outcome]_cf_lower,
    • outcome - character. name of the outcome variable
    • cutpoint - date object containing the time point or points (up to 2) of intervention
    • effect size - a data frame containing the intercept shift associated with intervention, including uncertainty
    • gof - goodness of fit based on BIC

  4. bma.r

    Function that carries out Bayesian model averaging over multiple interrupted time series analyses.
    Inputs:

    • itsResults - list of lists. standard output from multiple runs of its.r

    Outputs (in a list):

    • data - the input data object with six new columns: [outcome]_pred, [outcome]_pred_upper, [outcome]_pred_lower, [outcome]_cf, [outcome]_cf_upper, [outcome]_cf_lower,
    • outcome - character. name of the outcome variable
    • cutpoint - date object containing the time point or points (up to 2) of intervention
    • effect size - a data frame containing the intercept shift associated with intervention, including uncertainty

  5. cpbma.r

    Function that carries out change-point Bayesian model averaging following Kurum 2016. Inputs:

    • data - data table object in 'prepped' format (see below)
    • outcome - character. name of the outcome variable
    • cutpoint - date vector or matrix (up to 2 columns) containing the time points of intervention to include in BMA
    • slope - logical. TRUE indicates that an interaction term (or terms) should be used to estimate a different slope before/after intervetion

    Outputs (in a list):

    • data - the input data object with six new columns: [outcome]_pred, [outcome]_pred_upper, [outcome]_pred_lower, [outcome]_cf, [outcome]_cf_upper, [outcome]_cf_lower,
    • effect size - a data frame containing the intercept shift associated with intervention, including uncertainty

  6. graph.r

    Function that produces a time series graph from the results of its.r, bma.r or cpbma.r. Inputs:

    • itsOutput - list. standard output from its.r

    Outputs:

    • p - a ggplot graph

  7. sensitivity_analysis.r

    Script that assembles numerous model outputs and re-graphs them together Output:

    • file at /visualizations/sensitivity_analysis.pdf
Data formats:

  1. 'prepped' format FILL IN

  2. 'output' format FILL IN