| Title: | Flexible Contingency Tables for Epidemiology |
|---|---|
| Description: | Builds contingency tables that cross-tabulate multiple categorical variables and also calculates various summary measures. Export to a variety of formats is supported, including: 'HTML', 'LaTeX', and 'Excel'. |
| Authors: | Stuart Lacy |
| Maintainer: | Stuart Lacy <[email protected]> |
| License: | GPL-2 |
| Version: | 0.2.2 |
| Built: | 2024-10-30 09:16:18 UTC |
| Source: | https://github.com/stulacy/epitab |
A contingency table provides cross-tabulated frequencies between an outcome
of interest and one or more independent variables. This function extends
contingency tables to include summary statistics formed both column-wise
and row-wise, looking at outcomes and covariates respectively in isolation.
This allows for a large amount of flexibility and tables can be drawn for
a variety of situations. By default, the print method fits these
tables to standard R console output, but publication quality tables
can be produced using the neat_table function. See the vignette
for further guidance.
contingency_table(independents, data, outcomes = NULL, crosstab_funcs = NULL, row_funcs = NULL, col_funcs = NULL, marginal = TRUE)contingency_table(independents, data, outcomes = NULL, crosstab_funcs = NULL, row_funcs = NULL, col_funcs = NULL, marginal = TRUE)
independents |
A named list of independent variables, which will be distributed down the table's rows. The variables must be specified by strings, with the item name used as the column header. |
data |
The data set that contains the columns specified in
|
outcomes |
The variables to cross-tabulate by. These will be distributed across the table's columns. Specified as a named list of strings. Must correspond to factor or character variables. |
crosstab_funcs |
A list of functions that are applied to every cross-tabulation
permutation of |
row_funcs |
A list of functions that are applied row-wise to the table,
one independent variable at a time, providing a value for each level of
the factors specified in |
col_funcs |
A list of functions that are applied column-wise to the table,
to every outcome separate from the independent variables.
Examples provided with the package included |
marginal |
Whether to include the counts of each level of |
An S3 object of class contintab, that provides the cell contents
as a matrix of strings.
# This example uses a dummy data set of whether an individual was treated or not treat <- data.frame(age=abs(rnorm(100, 60, 20)), sex=factor(sample(c("M", "F"), 100, replace=TRUE)), variant=factor(sample(c("A", "B"), 100, replace=TRUE)), treated=factor(sample(c("Yes", "No"), 100, replace=TRUE), levels=c("Yes", "No"))) treat$agebin <- cut(treat$age, breaks=c(0, 40, 60, 80, 9999), labels=c("0-40", "41-60", "61-80", "80+")) # Displays a standard contingency table contingency_table(list("Age"='agebin', "Sex"='sex'), outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), data=treat) # Continuous variables can be summarised with respect to the outcome # by using col_funcs contingency_table(list("Age"='agebin', "Sex"='sex'), outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), col_funcs=list("Mean age"=summary_mean('age')), data=treat) # Regression coefficients can be added using row_funcs contingency_table(list("Age"='agebin', "Sex"='sex'), treat, outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), row_funcs=list("Odds ratio"=odds_ratio('treated'), "Adjusted odds ratio"=odds_ratio('treated', adjusted=TRUE)))# This example uses a dummy data set of whether an individual was treated or not treat <- data.frame(age=abs(rnorm(100, 60, 20)), sex=factor(sample(c("M", "F"), 100, replace=TRUE)), variant=factor(sample(c("A", "B"), 100, replace=TRUE)), treated=factor(sample(c("Yes", "No"), 100, replace=TRUE), levels=c("Yes", "No"))) treat$agebin <- cut(treat$age, breaks=c(0, 40, 60, 80, 9999), labels=c("0-40", "41-60", "61-80", "80+")) # Displays a standard contingency table contingency_table(list("Age"='agebin', "Sex"='sex'), outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), data=treat) # Continuous variables can be summarised with respect to the outcome # by using col_funcs contingency_table(list("Age"='agebin', "Sex"='sex'), outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), col_funcs=list("Mean age"=summary_mean('age')), data=treat) # Regression coefficients can be added using row_funcs contingency_table(list("Age"='agebin', "Sex"='sex'), treat, outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), row_funcs=list("Odds ratio"=odds_ratio('treated'), "Adjusted odds ratio"=odds_ratio('treated', adjusted=TRUE)))
Builds a function to calculate cross-tabulated frequencies
freq(proportion = c("column", "row", "none"), display = c("percentage", "ratio"), digits = 3, missing = NULL)freq(proportion = c("column", "row", "none"), display = c("percentage", "ratio"), digits = 3, missing = NULL)
proportion |
Whether to display proportions in brackets after the counts as either row-wise or column-wise, or not at all. |
display |
How to display the proportions if required. |
digits |
The number of digits to specify |
missing |
Specifies covariates that shouldn't be included in calculating proportions for column-wise summaries. |
A function that calculates the frequency of a cell in a contingency table.
Builds a function to fit a Cox model to the outcome survival object and returns the hazard ratios for each covariate level.
hazard_ratio(outcome, adjusted = FALSE, relevel_baseline = FALSE, digits = 2, ci = TRUE)hazard_ratio(outcome, adjusted = FALSE, relevel_baseline = FALSE, digits = 2, ci = TRUE)
outcome |
The dependent variable, specifies a |
adjusted |
Whether to adjust for the other covariates, specified by
|
relevel_baseline |
Whether to use the largest level as the baseline. |
digits |
The number of digits to display. |
ci |
Whether to include a confidence interval in parentheses after the estimate. |
See documentation for contingency_table and vignette
for usage.
A function that is used to calculate hazard ratios.
This function provides a default means of converting a contingency table
into HTML or LaTeX for publishing. By default, multiple column and row
spanning cells are formed to accentuate the hierarchical nature of the data.
The output of this function is a kable object and so can be further
manipulated.
neat_table(table, format = c("html", "latex"), ...)neat_table(table, format = c("html", "latex"), ...)
table |
A |
format |
A string specifying output format passed to |
... |
Other arguments passed to |
A kable object.
# This example uses a dummy data set of whether an individual was treated or not treat <- data.frame(age=abs(rnorm(100, 60, 20)), sex=factor(sample(c("M", "F"), 100, replace=TRUE)), variant=factor(sample(c("A", "B"), 100, replace=TRUE)), treated=factor(sample(c("Yes", "No"), 100, replace=TRUE), levels=c("Yes", "No"))) treat$agebin <- cut(treat$age, breaks=c(0, 40, 60, 80, 9999), labels=c("0-40", "41-60", "61-80", "80+")) tab <- contingency_table(list("Age"='agebin', "Sex"='sex'), outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), col_funcs=list("Mean age"=summary_mean('age')), data=treat) # For use in an Rmarkdown that outputs to HTML neat_table(tab, 'html') # When outputting to PDF, the \code{booktabs} option produces well-formatted tables neat_table(tab, 'latex', booktabs=TRUE)# This example uses a dummy data set of whether an individual was treated or not treat <- data.frame(age=abs(rnorm(100, 60, 20)), sex=factor(sample(c("M", "F"), 100, replace=TRUE)), variant=factor(sample(c("A", "B"), 100, replace=TRUE)), treated=factor(sample(c("Yes", "No"), 100, replace=TRUE), levels=c("Yes", "No"))) treat$agebin <- cut(treat$age, breaks=c(0, 40, 60, 80, 9999), labels=c("0-40", "41-60", "61-80", "80+")) tab <- contingency_table(list("Age"='agebin', "Sex"='sex'), outcomes=list('Treated'='treated'), crosstab_funcs=list(freq()), col_funcs=list("Mean age"=summary_mean('age')), data=treat) # For use in an Rmarkdown that outputs to HTML neat_table(tab, 'html') # When outputting to PDF, the \code{booktabs} option produces well-formatted tables neat_table(tab, 'latex', booktabs=TRUE)
Builds a function to run logistic regression on the outcome variable of interest and return the odds ratios for each covariate level.
odds_ratio(outcome, adjusted = FALSE, relevel_baseline = FALSE, digits = 2, ci = TRUE)odds_ratio(outcome, adjusted = FALSE, relevel_baseline = FALSE, digits = 2, ci = TRUE)
outcome |
The dependent variable as a string. |
adjusted |
Whether to adjust for the other covariates, specified by
|
relevel_baseline |
Whether to use the largest level as the baseline. |
digits |
The number of digits to display. |
ci |
Whether to include a confidence interval in parentheses after the estimate. |
See documentation for contingency_table and vignette
for usage.
A function that is used to calculate odds ratios.
Prints the contingency table as an ASCII table
## S3 method for class 'contintab' print(x, ...)## S3 method for class 'contintab' print(x, ...)
x |
An object of class |
... |
Options passed to |
None. Prints a table to standard output.
Builds a function to calculate the mean of a continuous variable for each level of an outcome.
summary_mean(var)summary_mean(var)
var |
A continuous variable name as a string. |
A function that calculates the mean value of var
for each outcome level.
Builds a function to calculate the median of a continuous variable for each level of an outcome.
summary_median(var)summary_median(var)
var |
A continuous variable name as a string. |
A function that calculates the median value of var
for each outcome level.