Skip to contents

Disproportionality Analysis for Drug-Event Combinations

Source: R/disproportionality_analysis.R
disproportionality_comparison.Rd

This function performs a disproportionality analysis for drug-event combinations in the FAERS dataset, calculating various metrics such as the Reporting Odds Ratio (ROR), Proportional Reporting Ratio (PRR), Relative Reporting Ratio (RRR), and Information Component (IC). Contrary to disproportionality_analysis(), it calculates the association based on number of reports, and not based on the Drug and Reac database. For a more flexible application of other disproportionality measures, you can look at this vignette of the pvda R package.

Usage

disproportionality_comparison(
  drug_count = length(pids_drug),
  event_count = length(pids_event),
  drug_event_count = length(intersect(pids_drug, pids_event)),
  tot = nrow(Demo),
  print_results = TRUE
)

Arguments

drug_count

An integer representing the number of reports for the drug of interest. Default is the length of pids_drug.

event_count

An integer representing the number of reports for the event of interest. Default is the length of pids_event.

drug_event_count

An integer representing the number of reports for the drug-event combination. Default is the length of the intersection of pids_drug and pids_event.

tot

An integer representing the total number of reports in the dataset. Default is the number of rows in the Demo table.

print_results

A logical to control whether the results should also be printed, besides being stored in the results.

Value

This function prints a contingency table and several disproportionality metrics (which are also stored in a list):

  • ROR: Reporting Odds Ratio with confidence intervals.

  • PRR: Proportional Reporting Ratio with confidence intervals.

  • RRR: Relative Reporting Ratio with confidence intervals.

  • IC: Information Component with confidence intervals.

  • IC_gamma: Gamma distribution-based Information Component with confidence intervals.

Details

The function constructs a contingency table for the drug-event combination and computes the following metrics:

ROR

Reporting Odds Ratio: Based on odds ratio

PRR

Proportional Reporting Ratio: The expected probability of the event is calculated on the population not having the drug of interest.

RRR

Relative Reporting Ratio: The expected probability of the event is calculated on the entire population.

IC

Information Component: A measure based on Bayesian confidence propagation neural network models. It is the log2 of the shrinked RRR.

IC_gamma

Gamma distribution-based Information Component: An alternative IC calculation using the gamma distribution. It is more appropriate for small databases

See also

Other disproportionality functions: disproportionality_analysis(), disproportionality_trend()

Examples

# Example usage
disproportionality_comparison(
  drug_count = 100, event_count = 50,
  drug_event_count = 10, tot = 10000
)
#>     E   nE
#> D  10   90
#> nD 40 9860
#> 
#> 
#> ROR = 27.32 (11.81-57.83)
#> PRR = 24.75 (12.74-48.1)
#> RRR = 20 (10.44-38.3)
#> IC = 3.39 (2.31-4.11)
#> IC_gamma = 2.53 (2.36-4.15)