Skip to contents

Facts Over Time – Processing

Source: R/facts_over_time.R
process_fot.Rd

Intakes the output of check_fot in order to apply additional processing. This includes applying a heuristic meant to compare the fact count in a given time period to other time periods around it. For a monthly computation, for example, the heuristic would be month / ((month-1 * .25) + (month+1 * .25) + (month-12 * .5)) which is the value for a given month divided by the weighted average of the value in the previous month, the next month, and the same month in the previous year. The function can also optionally compute a ratio given a total_pt column (added by the user) to be used as a denominator and a ratio_mult value to be used as a multiplier.

Usage

process_fot(
  fot_results,
  target_col = "row_cts",
  add_ratios = FALSE,
  ratio_mult = 10000,
  rslt_source = "remote",
  csv_rslt_path = NULL
)

Arguments

fot_results

tabular input || required

The tabular output of check_fot. This table should include results for all institutions that should be included in the computation of overall / "network level" statistics.

target_col

string || defaults to row_cts

A string indicating the column in fot_results that should be used to compute the primary heuristic. For the ratio computation, row_pts will always be used regardless of what is specified in this parameter.

add_ratios

boolean || defaults to FALSE

A boolean indicating whether incidence ratios / rates should be computed. if TRUE, the user should edit the fot_results table to include an additional total_pt column with counts for the denominator of the user's choosing. We recommend including the desired denominator population as a check passed into the check_fot function to facilitate the computation of patient counts for each time period.

ratio_mult

integer || defaults to 10,000

If add_ratios is set to TRUE, this parameter should represent the numerical multiplier that should be used to compute the incidence ratio. This will result in a ratio that represents the fact rate per (ratio_mult) patients.

rslt_source

string || defaults to remote

A string that identifies the location of the fot_results table. Acceptable values are

  • local - table is stored as a dataframe in the local R environment

  • csv - table is stored as a CSV file

  • remote - table is stored on a remote database

csv_rslt_path

string || defaults to NULL

If rslt_source has been set to csv, this parameter should indicate the path to the result file(s). Otherwise, this parameter can be left as NULL

Value

This function will return a list of up to three dataframes:

  • fot_heuristic: a table with the summary heuristic for each fact type (month / ((month-1) * .25 + (month+1) * .25 + (month-12)*.5))

  • fot_heuristic_summary: a table with summary values (mean, med, sd, q1, q3) based on the heuristic

  • fot_ratios: the fot_results table with an additional column with the incidence ratio (only returned if row_ratio = TRUE)

Examples

# This function should be run after check_fot has been executed for all
# network institutions and results have been combined into a common table

# Once the labels have been applied, the function can be executed
## When results are kept locally:
if (FALSE) { # \dontrun{
my_fot_process <- process_fot(fot_results = my_fot_rslts,
                              target_col = 'row_cts',
                              rslt_source = 'local')
} # }

## When results are kept in CSV files:
if (FALSE) { # \dontrun{
my_fot_process <- process_fot(fot_results = 'my_fot_rslts',
                              target_col = 'row_cts',
                              rslt_source = 'csv',
                              csv_rslt_path = 'path/to/my/results')
} # }

## When results are kept on a remote database:
if (FALSE) { # \dontrun{
my_fot_process <- process_fot(fot_results = 'my_fot_rslts',
                              target_col = 'row_cts',
                              rslt_source = 'remote')
} # }

# You can also optionally compute patient incidence ratios. This computation will use
# the patient count, regardless of what is specified in the `target_col` argument.
# A column with the total patient count for that time period will need to be added to
# the results output. It can reflect any patient cohort, so long as the column is called
# `total_pt`. We recommend computing the denominator as part of the initial check_fot execution
# and extracting the counts from the resulting output.

if (FALSE) { # \dontrun{
my_fot_process <- process_fot(fot_results = 'my_fot_rslts',
                              target_col = 'row_cts',
                              add_ratios = TRUE,
                              ratio_mult = 10000,
                              rslt_source = 'remote')
} # }