FMCSA continues to test small scale IRT models to determine whether IRT is a better scoring and resource prioritization system than the longstanding Safety Measurement System approach, created more than a decade ago for the CSA program. While some IRT models are being run in the background, out of view of the industry and the public, some information is being shared by FMCSA, albeit in a piecemeal fashion. Here’s the latest FMCSA-provided information picked up by STC at a recent conference:

  1. IRT takes approximately 900 violations (from roadside inspections and traffic enforcement) and puts them into approximately 50 different categories (described by FMCSA as 45-55 buckets);
  2. If a carrier has a violation in a category, IRT flags that category by turning it red;
  3. There’s also an IRT algorithm running in the background attempting to determine which categories are best at determining the ‘safety culture’ of a carrier;
  4. FMCSA has been comparing, among other things, the IRT safety culture score(s) of a carrier with its accident frequency and, so far, the IRT safety culture score appears to be a good predictor;
  5. IRT also appears to address some of the problems identified with the CSA SMS model, including enforcement disparities between State truck enforcement programs.

It’s likely in late 2019, FMCSA will publish a Notice in the Federal Register informing the industry and other interested stakeholders of its IRT-related work and findings thus far. The industry should also expect at least one FMCSA public outreach session concerning the IRT findings, a timeframe for Agency decision making, and likely next steps. If you’d like to read a prior IRT update from STC in February 2019, which includes additional FMCSA observations, please click here: