optic

This tool can help researchers assess how co-occurring policies and confounding can affect the performance of statistical models commonly used in state policy evaluations. Specifically, the tool helps users compare the performance of various causal inference models using their own longitudinal data. Users can select from a variety of simulation options to explore how different state policy evaluation methods perform. Although the tool was initially created to examine data related to opioids, its framework can be used with longitudinal data on any topic.

Recent research on difference-in-differences (DID) models revealed issues with traditional DID models, and there has been an explosion of new methods in this area for researchers to consider. Researchers found it difficult to evaluate the relative performance of different causal inference methods using longitudinal outcome data on opioid mortality and opioid prescribing rates; thus, they designed a series of simulations to study the performance of various methods under different scenarios for any type of repeated measures outcome data.

The tool's introductory vignette provides a working example of how to use the package. This example uses the sample overdoses dataset provided with the package. Users need the R software environment (version 4.1.0 or above) to use the package.