The purpose of this repository is to provide reproducible code and raw data inputs to produce the parameters used in the cost-benefit modelling of Individual Placement and Support (IPS) programmes for individuals with mental health and musculoskeletal problems.
Fork the project on GitHub to create your own repository. This project
uses renv to aid
reproducibility of results by ensuring users make use of the same
environment.
When a new user first launches in this project, renv should
automatically bootstrap itself, thereby downloading and installing the
appropriate version of renv into the project library. After this has
completed, they can then use renv::restore() to restore the project
library locally on their machine. Once the project library is restored,
the repository is ready to be used to replicate the analysis once the
requirements below are met.
There are five stages to the analysis
- Defining the eligible population by local authority.
- Obtaining labour market parameters.
- Calibrating labour market transition probabilities for the treatment effect.
- Calculating EQ-5D health utilities.
- Obtaining income, tax, and benefit profiles.
Note that this repository uses the renv R package to produce a reproducible environment for this analysis. see the package website for more information.
When first launching the project, the function renv::restore() will
populate the renv/library folder, installing the R packages saved in
the lockfifle.
The lockfile renv.lock is in the top level of the repository, and
contains the metadata for all R packages used in the project. The
libraries themselves are ignored by git and not uploaded to GitHub, and
so a new user needs to install the packages themselves using the
metadata.
The .Rprofile file in the top level of the repository is automatically
run when the project is opened and this runs the script file
renv/activate.R to set the renv folder as the source R package
library.
The version of R used to produce the analysis is 4.3.1.
Use of the project requires access to various datasets which are
obtained through the UK Data Service. These must be downloaded in tab
delimited form and the raw data files placed in the relevant folder in
the data folder. The datasets required are:
-
Health Survey for England (HSE) 2018
-
Understanding Society (USoc) waves 1-13
-
Family Resources Survey (FRS) 2018/19 - 2021/22
-
Labour Force Survey (LFS)
-
Five-quarter longitudinal datasets from AJ12-AJ13 to JM22-JM23
-
Cross-sectional quarterly datasets from JM15 to OD23
-
The fifth stage of the analysis involves a micro-simulation of income, taxes, and benefits by health condition and employment status based on FRS data. To reproduce this analysis, the user also needs to install the EUROMOD software and obtain the version of UKMOD used (B1.09) used for the micro-simulation.
The analysis requires the use of a number of R packages for cleaning the raw data for the FRS, LFS, USoc, and HSE. These packages are all open source and available on GitHub:
Each of these packages are part of the renv environment and are
restored with the other CRAN packages when using renv::restore().
Please cite this code repository as:
Morris, D. Clarke, H. Brennan, A. and Whitworth, A. (2024) Estimating costs and benefits of alternative IPS beyond SMI scenarios - code repository. University of Sheffield. doi: .