Appendix A: Details of Intervention Selection

RAND and CDC worked together closely to select the interventions for consideration. Although we began with a list of 16 interventions that met our three criteria—implementable at the state level, demonstrated to be effective, and not already in widespread use—we both added to and subtracted from this initial list. The two main reasons for doing so were the availability of data and the fact that the Countermeasures That Work report occasionally combined multiple interventions that we need to define more precisely to develop cost estimates. The reasons for including or excluding interventions that met the above criteria are provided here:

  • Saturation patrols: Although these were not on the original list, because they can be used in states that forbid the use of sobriety checkpoints, we decided to include them.
  • DWI courts: We dropped this from consideration because evidence for their effectiveness is limited and it is difficult to design a methodologically sound study or determine their costs.
  • Red-light cameras: Despite conflicting literature, including some recent evidence suggesting that they are not as effective as initially thought, we retained this intervention on the grounds that it would be helpful to conduct work on their costs and benefits because many states and localities have considered red-light cameras. In addition, a majority of the literature reports that red-light cameras are effective.
  • Vehicle and license plate sanctions: The Countermeasures That Work report lists five types of sanctions under this heading. (Other sources include alcohol interlocks as a type of sanction, but, because these have been extensively studied, we retained them as a separate intervention.) We considered license plate sanctions separately from vehicle sanctions. Three vehicle sanctions are commonly used: vehicle impoundment, vehicle forfeiture, and vehicle immobilization. Vehicle impoundment has received the most study, so we retained this intervention. Two license plate sanctions are in use: special plates and license plate impoundment. Of the seven states with special license plates, six also have impoundment laws. Although both studies of impoundment laws were carried out in the same state (Minnesota), the more recent was conducted in 2011. Both studies of impoundment laws found positive results, and the most recent study was conducted in 2001.
  • Communications and high-visibility enforcement of seat belt and child restraint laws: All states conduct such enforcement campaigns, but some more extensively than others, making it difficult to determine an appropriate metric for how widely used this intervention currently is. We spoke with Richard Compton, director of the Office of Behavioral Safety Research at NHTSA, on February 16, 2012. In this conversation, he noted that, for high-visibility enforcement to work, both increased enforcement and publicity are needed. In our previous research, we did not find any references to child restraint high-visibility enforcement campaigns conducted independently of seat belt campaigns. The UNC report notes, “NHTSA typically includes child restraint and booster seat use and enforcement as a part of their Click It Or Ticket campaigns” (pp. 2–36). Therefore, we decided to combine these previously separate interventions into one covering all enforcement campaign activity for seat belts and child restraints.
  • Referring older drivers to licensing agencies: We found relatively little literature on the effectiveness of referring older drivers for additional testing to retain a driver’s license, so we did not retain this intervention.
  • Lower BAC limits for repeat offenders: The main study on this intervention (R. Jones and Rodriguez-Iglesias, 2004) identified five states that use this intervention. This is the main source cited in UNC Highway Safety Research Center, 2011. However, our subsequent review of NHTSA’s Digest of Impaired Driving and Selected Beverage Control Laws, 2011, found that only one state retains such a law. We have been unable to determine precisely why this law was apparently repealed in the other states. Given the apparent difficulty in implementing and enforcing this intervention, we decided not to include it in the tool.
  • Bicycle and pedestrian interventions: Although we had hoped to include an intervention that specifically addressed these groups of road users, we could not identify one that met our criteria. Almost all the effective interventions we researched were changes in roadway configuration (e.g., shortened curb radii or raised medians). Although we considered pedestrian countdown signals, we could not develop a sound method for determining how many would be required to be effective.
  • Cell phone and texting bans: These bans have received a good deal of attention as a potential means of countering distracted driving because of device use. However, although some evidence supports their effectiveness (Sampaio, 2010), other studies have found only a short-term or mixed effect (Jacobson et al., 2012; Lim and Chi, 2013). Although these might be considered useful interventions in the future, perhaps combined with specific types of enforcement, at the time we conducted this research, there was insufficient evidence of their general effectiveness.