get_matches: Find matched controls for a set of cases

Description

This function will provide a set of matched controls for a given set of cases.

Usage

get_matches(cases, control_pool, n_controls, match_vars, extra_vars, extra_conditions = NULL, cores = 1, track = TRUE, tracker = function(case_num) ".", method = c("incidence_density", "exact"), diagnosis_date = NULL)

Arguments

cases

dataframe of cases

control_pool

dataframe of potential controls to be used for matching

n_controls

number of controls to match to each case

match_vars

character vector of variables in the dataframes to be used to perform the matching

extra_vars

character vector of other variables to be used in the matching to define other conditions

extra_conditions

a character vector of length 1 defining further restrictions on matching

cores

number of cpu cores to be used by multicore (windows users should leave set to 1)

track

logical should a dot be printed to std.out for each case?

tracker

function to track progress of the function (See details)

method

The method of selection of controls (see details)

diagnosis_date

character the name of the variable in the cases and control_pool datasets containing the date of diagnosis (or other event to base the IDM method on). If there is no diagnosis date for a patient, this should be represented by NA

Details

Setting method to "exact" means that the matched controls are removed from the control pool after each case has been matched. This makes this method not thread safe and so will only run on a single core (and more slowly). Setting method to "incidence_density" is thread safe as the same controls can be used for more than one case. See Richardson (2004) Occup Environ Med 2004;61:e59 doi:10.1136/oem.2004.014472 for a description of IDS matching. Also see the introduction vignette. The tracker variable allows for different outputs to track the progress of the function. This is currently set to ouput a dot for every case matched. A function can be added to the argument For a more verbose tracking, e.g. to track number of cases, set tracker = function() paste0(case_num, ",")