-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathREADME.Rmd
150 lines (88 loc) · 8.89 KB
/
README.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
---
output: github_document
---
<!-- README.md is generated from README.Rmd. Please edit that file -->
```{r, include = FALSE}
knitr::opts_chunk$set(
collapse = TRUE,
comment = "#>",
fig.path = "man/figures/README-",
out.width = "100%"
)
```
# Processing Mortality Microdata<img src="logo.png" align="right" style="padding-left:10px;background-color:white;" width="100" height="100" />
<!-- badges: start -->
[](https://www.repostatus.org/#active)
[](https://www.tidyverse.org/lifecycle/#stable)
[](https://doi.org/10.17605/OSF.IO/WN6RH)
<!-- badges: end -->
## The Sheffield Tobacco and Alcohol Policy Modelling Platform
This R package was developed as part of the Sheffield Tobacco and Alcohol Policy Modelling https://stapm.gitlab.io/ by the [School of Health and Related Research (ScHARR) at the University of Sheffield](https://www.sheffield.ac.uk/scharr).
The aim of the research programme is to identify and evaluate approaches to reducing the harm from tobacco and alcohol, with the aim of improving commissioning in a public health policy context, i.e. providing knowledge to support benefits achieved by policymakers.
The two objectives of the research programme are:
- To evaluate the health and economic effects of past trends, policy changes or interventions that have affected alcohol consumption and/or tobacco smoking
- To appraise the health and economic outcomes of potential future trends, changes to alcohol and/or tobacco policy or new interventions
The STAPM modelling is not linked to the tobacco or alcohol industry and is conducted without industry funding or influence.
## Purpose of making the code open source
The code has been made open source for the following two reasons:
- Transparency. Open science, allowing review and feedback to the project team on the code and methods used.
- Methodology sharing. For people to understand the code and methods used so they might use aspects of it in their own work, e.g., because they are doing something partially related that isn't exactly the same job and might like to 'dip into' elements of this code for inspiration.
## Stage of testing and development
The code is actively being used in project work. It is being reviewed and developed all the time; more tests and checks are being added.
The repository is not intended to be maintained by an open source community wider than the development team.
## Data access
The mortality microdata that this R package is designed to process was obtained through data access requests to the Office for National Statistics (for English and Welsh data) and National Records Scotland.
In ScHARR, all data is stored and processed according to the [ScHARR Information Governance Policy](https://www.sheffield.ac.uk/scharr/research/igov/policy00). No mortality microdata is included within this package.
## Data checks
Data checks are brief reports that show the results of the mortality data processing using the mort.tools package. The data checks help to visualise the patterns in the data.
* [Scottish national death rates](https://stapm.shinyapps.io/scottish_death_rates_dashboard/)
## Code repositories
The code on Github (https://github.com/STAPM/mort.tools) is a mirror of the code in a private Gitlab repository where the actual development takes place (https://gitlab.com/stapm/r-packages/mort.tools). The code in the Github repository is linked to a repository on the Open Science Framework, which provides the doi for the package citation (https://osf.io/wn6rh/).
## Citation
Gillespie D, Webster L, Angus C, Brennan A ([YEAR]). mort.tools: An R Package for Processing Mortality Microdata. R package version [x.x.x]. University of Sheffield. https://stapm.gitlab.io/r-packages/mort.tools/. doi: https://doi.org/10.17605/OSF.IO/WN6RH
## Motivation
The motivation for `mort.tools` was to develop a set of functions to clean and do basic processing of the data. The advantage of this is that it ensures we apply a consistent approach across years of data, data from different countries, and across the different projects for which we use the data.
`mort.tools` was created as part of a programme of work on the health economics of tobacco and alcohol at the School of Health and Related Research (ScHARR), The University of Sheffield. This programme is based around the construction of the Sheffield Tobacco and Alcohol Policy Model (STAPM), which aims to use comparable methodologies to evaluate the impacts of tobacco and alcohol policies, and investigate the consequences of clustering and interactions between tobacco and alcohol consumption behaviours.
## Functionality
mort.tools works with data on annual counts of deaths and mid-year population sizes. Counts of deaths are grouped according to the set of tobacco and alcohol related diseases considered in the modelling (see https://osf.io/v945r/). Deaths not falling into one of these classifications are categorised as 'all other causes'. Counts of deaths and mid-year population sizes are stratified by single years of age, sex and Index of Multiple Deprivation quintiles. For England, further stratification by local authority was included in the data request to the Office for National Statistics. The mort.tools R package does not function without this data. However no datasets are provisioned. Data must therefore be obtained by data requests to the Office for National Statistics or National Records Scotland.
The package is primarily designed for users at the University of Sheffield, working off the university's networked drives and in the secure virtual machine data environment. This is where most of the testing has taken place so there might be unexpected issues out of that environment.
What the software does in general and how it relates to data is documented in the vignettes under "Methods and data checks".
## Usage
See [Get started](mort-tools.html) for example workflows. We have written a range of short methods descriptions. There is a [reference list of functions](https://stapm.gitlab.io/r-packages/mort.tools/reference/index.html) in this package. Data processing is done on the University of Sheffield managed secure virtual machine data environment.
The suite of functions within `mort.tools` reads the data for each year, produces cause-specific life tables (focusing on the diseases related to tobacco and alcohol), and gives options for conducting forecasts of cause-specific death rates. The methods for forecasting are still under development and subject to review.
The **inputs** are the counts of death and mid-year population size.
The **processes** applied by the functions in `mort.tools` give options to:
1. Read each year of data.
1. Group the data by the causes related to tobacco and alcohol.
1. Construct cause-specific life tables.
1. Forecast trends in cause-specific mortality.
The **output** of these processes are the trends in cause-specific death rates and population sizes. These dataset can be saved so that you don't need to run the cleaning processes in `mort.tools` each time you want to use the cleaned data.
## Installation
`mort.tools` is publicly available via Github.
By default the user should install the latest tagged version of the package. Otherwise, if you want to reproduce project work and know the version of the package used, install that version.
If on a University of Sheffield managed computer, install the R, RStudio and Rtools bundle from the Software Centre. Install Rtools - using the [installr](https://cran.r-project.org/web/packages/installr/index.html) package can make this easier. Then install the latest or a specified version of `mort.tools` from Github with:
```{r gh_installation, message=FALSE, eval = FALSE}
#install.packages("devtools")
# The tobalcepi STAPM R package is also required
devtools::install_git(
"https://github.com/stapm/tobalcepi.git",
ref = "x.x.x",
build_vignettes = FALSE)
devtools::install_git(
"https://github.com/stapm/mort.tools.git",
ref = "x.x.x",
build_vignettes = FALSE)
# ref = "x.x.x" is the version to install - change to the version you want e.g. "1.2.3"
```
Or clone the package repo locally and use the 'install and restart' button in the Build tab of RStudio. This option is more convenient when testing development versions.
Then load the package, and some other packages that are useful. Note that the code within `mort.tools` uses the `data.table::data.table()` syntax.
```{r pkgs, eval = F}
# Load the package
library(data.table)
library(mort.tools)
library(tobalcepi)
# Other useful packages
library(dplyr) # for data manipulation and summary
library(magrittr) # for pipes
library(ggplot2) # for plotting
```