Skip to contents

Cohort Attrition

Source: R/ca_process.R
ca_process.Rd

This is a plausibility module that will assess attrition counts and how counts change from each step to the next to help identify the most impactful and any potentially anomalous attrition criteria. The user will provide a precomputed attrition. A sample attrition_tbl with the expected formatting can be found using cohortattrition::.

Usage

ca_process(
  attrition_tbl,
  multi_or_single_site = "single",
  anomaly_or_exploratory = "exploratory",
  start_step_num = 0,
  var_col = "num_pts",
  p_value = 0.9
)

Arguments

attrition_tbl

tabular input || required

A table or CSV file with attrition information for each site included in the cohort. This table should minimally contain:

  • site | character | the name of the institution

  • step_number | integer | a numeric identifier for the attrition step

  • attrition_step | character | a description of the attrition step

  • num_pts | integer | the patient count for the attrition step

multi_or_single_site

string || defaults to single

A string, either single or multi, indicating whether a single-site or multi-site analysis should be executed

anomaly_or_exploratory

string || defaults to exploratory

A string, either anomaly or exploratory, indicating what type of results should be produced.

Exploratory analyses give a high level summary of the data to examine the fact representation within the cohort. Anomaly detection analyses are specialized to identify outliers within the cohort.

start_step_num

integer || defaults to 0

The step_number from the attrition_tbl that should be considered the "start" for the analysis. This will drive comparisons of later steps to the start.

var_col

string || defaults to num_pts

The name of the column that should be used to conduct the analysis for the Multi-Site, Anomaly Detection, Cross-Sectional check. The options are:

  • num_pts: raw patient count

  • prop_retained_start: proportion patients retained from the starting step, as indicated by start_step_num

  • prop_retained_prior: proportion patients retained from prior step

  • prop_diff_prior: proportion difference between each step and the prior step

p_value

numeric || defaults to 0.9

The p value to be used as a threshold in the Multi-Site, Anomaly Detection, Cross-Sectional analysis

Value

This function will return a dataframe with all the original attrition information, plus columns examining the difference between each step and others. If the Multi-Site, Anomaly Detection, Cross-Sectional check is run, this output will also include some descriptive statistics about the chosen var_col and an indication of which sites are outliers at each attrition step.

Examples


#' Build mock study attrition
sample_attrition <- dplyr::tibble('site' = c('Site A', 'Site A', 'Site A', 'Site A'),
                                  'step_number' = c(1,2,3,4),
                                  'attrition_step' = c('step 1', 'step 2', 'step 3', 'step 4'),
                                  'num_pts' = c(100, 90, 70, 50))

#' Execute `ca_process` function
#' This example will use the single site, exploratory, cross sectional
#' configuration
ca_process_example <- ca_process(attrition_tbl = sample_attrition,
                                 multi_or_single_site = 'single',
                                 anomaly_or_exploratory = 'exploratory',
                                 start_step_num = 1) %>%
  suppressMessages()
#> ┌ Output Function Details ─────────────────────────────────────┐
#> │ You can optionally use this dataframe in the accompanying    │
#> │ `ca_output` function. Here are the parameters you will need: │
#> │                                                              │
#> │ Always Required: process_output, var_col                     │
#> │ Optional: log_scale                                          │
#> │                                                              │
#> │ See ?ca_output for more details.                             │
#> └──────────────────────────────────────────────────────────────┘

ca_process_example
#> # A tibble: 4 × 9
#>   site   step_number attrition_step num_pts prop_retained_prior ct_diff_prior
#>   <chr>        <dbl> <chr>            <dbl>               <dbl>         <dbl>
#> 1 Site A           1 step 1             100              NA                NA
#> 2 Site A           2 step 2              90               0.9              10
#> 3 Site A           3 step 3              70               0.778            20
#> 4 Site A           4 step 4              50               0.714            20
#> # ℹ 3 more variables: prop_diff_prior <dbl>, prop_retained_start <dbl>,
#> #   output_function <chr>

#' Execute `ca_output` function
ca_output_example <- ca_output(process_output = ca_process_example,
                               log_scale = FALSE,
                               var_col = 'prop_retained_start')

ca_output_example[[1]]

ca_output_example[[2]]



  
Attrition Step Reference
    
Step Number
      Description
    
1
step 1
2
step 2
3
step 3
4
step 4

#' Easily convert the graph into an interactive ggiraph or plotly object with
#' `make_interactive_squba()`

make_interactive_squba(ca_output_example[[1]])