5.3 Responsible Beverage Service

Overall Effectiveness Concerns: This countermeasure’s effectiveness has been examined in several research studies; however, server training programs are the only segment of responsible beverage service for adults that has been adequately documented and evaluated. Research suggests that server training programs can be effective if they involve intensive, high-quality, face-to-face server training accompanied by strong and active management support (Shults et al., 2001). When server training programs are not intensive and are not supported, they are unlikely to result in greater refusals of service to intoxicated patrons. Despite these positive research findings, the balance of evidence regarding countermeasure effectiveness remains inconclusive.

Responsible beverage service covers a range of alcohol sales policies and practices that prevent or discourage restaurant and bar patrons from drinking to excess or from driving while impaired by alcohol. Server training programs teach servers how to recognize the signs of intoxication and how to prevent intoxicated patrons from further drinking and from driving. Management policies and programs include limits on cheap drinks and other promotions, support for designated driver programs, strong commitment to server training, and strong support for servers who refuse alcohol to intoxicated patrons. NCHRP (2005, Strategy A2) provides an overview of responsible beverage service. See also Wagenaar and Tobler (2007) and Voas and Lacey (2011; pp. 131-137) for reviews and discussion of the research literature on this issue.

Beginning in the early 1980s a major effort was undertaken to encourage alcohol servers to comply with laws prohibiting the sale of alcoholic beverages to visibly intoxicated patrons. Since that time, many “server intervention” programs have been developed as a means of securing more responsible behavior on the part of servers. Some States have mandatory programs that require at least some alcohol retail employees to attend a server training course. Other States have voluntary programs that provide incentives for retailers to participate (e.g., liability protection or insurance discounts). The quality of server training programs can vary considerably. Wagenaar and Tobler (2007) note that many server training laws “are not optimally designed, do not ensure quality training, and do not ensure all servers are consistently trained, or retrained periodically” (p. 158).

Server training programs are the only segment of responsible beverage service for adults that has been documented and evaluated well. Activities directed at people under 21 are discussed separately in the Alcohol- and Drug-Impaired Driving chapter, Sections 6.1 through 6.4.

Use: As of January 1, 2018, there were 23 States that had some form of mandatory server training program in place; another 26 States had voluntary programs (NIAAA, 2018).

Effectiveness: The findings on the effectiveness of server training have been mixed. In their systematic review, Shults et al. (2001) found five high-quality evaluations of server training programs. They concluded that “intensive, high-quality, face-to-face server training, when accompanied by strong and active management support, is effective in reducing the level of intoxication in patrons” (p. 80). When server training programs are not intensive and are not supported, they are unlikely to result in greater refusals of service to intoxicated patrons.

Few studies have examined the effect of server training on alcohol-impaired crashes. An evaluation of a statewide server training program in Oregon found a 23% reduction in single- vehicle nighttime injury crashes following the program (Holder & Wagenaar, 1994). However, Molof and Kimball (1994) reviewed the same Oregon program and observed no decline in alcohol-related fatalities. A recent evaluation of a responsible beverage service program in two US communities examined outcome measures (Fell, Fisher, Yao, & McKnight, 2017; Fell, Fisher, Yao, McKnight, et al., 2017; NHTSA, 2017b). Outcomes were collected in stages at 10 intervention bars (pre-intervention, 6 months after intervention startup, and 1 year after intervention startup). The methods included bar patron breath tests at each bar to determine the proportion of patrons at high BACs and police-reported alcohol-crash involvement in the two treatment and two control communities. There was no significant impact of responsible beverage service on impaired driving related crashes or single vehicle nighttime crashes. However, patrons had lower average BACs, and a smaller proportion of patrons were intoxicated (BACs of .08 g/dL or higher). This effect was sustained when servers were properly trained in responsible beverage service and bar managers and owners were onboard with the program and aware of its enforcement. Another outcome of the study was the recommendation to target training programs toward problem establishments identified by “place of last drink” responses from DWI arrests (Fell, Fisher, Yao, & McKnight, 2017; Fell, Fisher, Yao, McKnight, et al., 2017; NHTSA, 2017b).

Costs: A typical alcohol server course takes about 4 to 8 hours. Course costs can be borne by the servers themselves, their employers, or the State.

Time to implement: Server training courses are offered by several private vendors and can be implemented in a few weeks. A statewide requirement for server training or more general responsible beverage service policies would require time to enact any necessary legislation, establish policies, and provide for program administration.

Other issues:

• Program quality: The quality of responsible beverage service programs can vary enormously, from excellent to abysmal. Management support can vary from enthusiastic to nonexistent. Shults et al. (2001) clearly limit their conclusions to high-quality programs with strong management support. The Alcohol Epidemiology Program (2000) cites several server training program evaluation studies that found no effect and notes that these programs may have been poorly supported or implemented. Grube and Stewart (2004) emphasize that management policy and its implementation may be at least as important as server training in determining responsible beverage service program effectiveness.
• Dram shop laws: As of 2015 there were 41 States that allow people injured by intoxicated drivers to recover damages from the licensed establishment that served or sold the alcohol in at least some situations (so-called “dram shop” laws) (NHTSA, 2017a). The potential threat of legal liability can provide strong encouragement to retailers to adopt responsible beverage service policies and practices. Research shows the implementation of dram shop laws is associated with reductions in alcohol-related crashes and fatalities (Voas & Lacey, 2011; Scherer et al., 2015).
• Enforcement of responsible beverage service: Enforcement of alcohol service laws is key, but largely lacking. Mosher et al. (2009) identified three main reasons for this: (1) a lack of societal and political will to address violations; (2) limited resources for enforcement operations; and (3) statutory provisions that make collection of evidence overly burdensome. As a result, action against licensed establishments (“dram shops”) has historically been limited to case law action involving serious crashes. Although alcohol enforcement by police is almost exclusively directed toward drivers, research has demonstrated that enforcement of alcohol service laws can help ensure alcohol retailers follow responsible serving practices. For example, an enforcement program in Michigan resulted in a three-fold increase in refusals of service to “pseudo-patrons” who simulated intoxication (McKnight & Streff, 1994).
• “Last Drink” programs: The goal of “Last Drink” programs is to determine where someone who was apprehended for impaired driving consumed the last drink prior to the arrest. This information is then provided to licensing authorities who may issue a warning letter to the retail establishment or take disciplinary action. An evaluation of a last drink program in Washington State found mixed results. No change was observed in retail establishment practices, but there were reductions in impaired-driving arrests and lower BACs among arrested drivers in the intervention community (Ramirez et al., 2008). Similar pilot programs have been tried in Australia, Canada, and New Zealand, although effectiveness data are lacking.