|
Article Excerpt
THIRTY STATES LOWERED THE LEGAL DRINKING age to 18 during the 1970s. In response to evidence of increased traffic fatalities (Wagenaar, 1983; Williams et al., 1974), however, almost all states raised their minimum drinking age between 1978 and 1987. In 1984, federal legislation was enacted that threatened withholding a portion of federal highway funds from states with a minimum drinking age lower than 21. All states eventually complied (Chaloupka et al., 2002). Evidence indicates that the uniform higher drinking age reduced alcohol-related car crashes among young people (General Accounting Office, 1987), as well as deaths from suicide, pedestrian injuries, homicide, and other unintentional injuries (Birckmayer and Hemenway, 1999; Jones et al., 1992; Parker and Rebhun, 1995).
Despite minimum-purchase-age laws, young people continue to have access to alcohol, and many of them drink. In 2001, more than half of high school seniors reported having drunk alcohol in the last 30 days; 36% reported consuming at least five drinks in one session during the month, a frequently used indicator of a heavy drinking day (Grunbaum et al., 2004).
Alcohol use by youth continues to lead to a substantial number of hospitalizations, disabilities, and premature deaths (Chaloupka et al., 2002; Perkins, 2002). Foremost among short-term consequences of alcohol use are motor vehicle crashes, the leading cause of death for teenagers (Centers for Disease Control and Prevention [CDC], 2005). Other acute problems are suicides, rapes, assaults, and unintentional injuries as well as alcohol poisonings, psychoses, dependence, and fetal alcohol syndrome. In addition, youth alcohol use can lead to property crimes and risky sexual behavior (Perkins, 2002).
Little has been known about the magnitude and costs of problems associated with underage drinking. Previous studies of the costs of alcohol use (e.g., Harwood et al., 1998; Manning et al., 1991) did not report costs by age. We estimated the magnitude of the general problem and of specific problems resulting from drinking by those under the legal drinking age of 21 and provided the first peer-reviewed estimate of the resulting costs. We compared those costs with associated alcohol sales to provide a yardstick that helps interpret the size of the problem.
Method
General approach
Although evidence is accumulating that underage drinking initiation increases alcohol-related problems in later life (DeWit et al., 2000; Grant and Dawson, 1998; Hingson et al., 2003), we limited costs to alcohol-related problems that can be directly tied to immediate or acute use to develop conservative estimates. We separately examined injuries and acute illnesses associated with underage drinking. For each problem, we first estimated fatal and nonfatal cases involving underage drinking. We then attempted to isolate those alcohol-involved cases actually attributable to or caused by alcohol use, as opposed to other contributing factors. Table 1 summarizes the sources of incidence data and the percentages used for alcohol involvement and attribution. We multiplied the number of alcohol-attributable cases by the costs per case to obtain total costs. The methodology for estimating attributable cases and costs is detailed below for each cause category.
Costs may be direct, indirect, or intangible. Medical costs are direct. In addition, alcohol-involved traffic crashes and crime lead to property loss, the need for emergency and other victim services, and criminal justice expenses. Indirect costs are the work loss from an individual's present and future inability to work as a result of injury or premature death. Intangible costs value the pain, suffering, and lost quality of life. This article places a monetary value on those losses, an increasingly common practice (Cutler and Richardson, 1998; French et al., 1996; Miller and Levy, 2000; Tolley et al., 1994). Because valuing quality-of-life losses is controversial, we also offer a nonmonetary estimate of these losses in quality-adjusted life years (QALYs; Gold et al., 1996).
Our cost estimates adopted a societal viewpoint and a cost framework similar to those of Manning et al. (1991). Unlike Manning et al., however, following several studies (e.g., Miller et al., 1998, 2001), we decomposed the value people place on their health and safety (their willingness to pay to prevent morbidity and mortality) into work-loss and quality-of-life components, which also provides readers with values in a cost-of-illness framework (Harwood et al., 1998) and a QALY framework (Gold et al., 1996).
In this article, medical care includes payments for hospital and physician care, rehabilitation, prescriptions, allied health services, medical devices, insurance-claims-processing costs, and costs associated with emergency medical transport. Costs for alcohol treatment are included only under that category.
Property losses arise in traffic crashes and crimes. They include costs to repair and replace lost or damaged property, as well as to process insurance claims. Public programs involve costs of police, fire, and victim services; criminal justice (adjudication, probation, incarceration, etc.); and foster care.
Lost work places a value on economic losses in the workplace or home due to mortality or impaired functioning (e.g., due to injury). It is measured in terms of the monetary value of lost wages plus the value of lost housework.
We offer nonmonetary and monetary estimates of the pain, suffering, and quality-of-life losses experienced by substance users, their victims, and their families due to illness, injury, and death. Extensive theoretical literature supports the inclusion of these costs (Manning et al., 1991; Miller and Levy, 2000), and the Office of Management and Budget (1989) requires including these costs whenever a regulatory cost-benefit analysis places a dollar value on saving human lives.
For fatalities, the value of pain, suffering, and lost quality of life was computed based on what people actually and routinely pay for small reductions in their chance of dying. An extensive literature containing more than 65 sound studies estimated that the value of a statistical life is at least $3.5 million 2001 U.S. dollars (Miller, 1990, 2000; Viscusi and Aldy, 2003).
The QALY loss to nonfatal injury was valued in four steps (Miller et al., 1995). First, physician ratings of the functional capacity typically lost over time by victims of every injury diagnosis cataloged in a common diagnosis system (Association for the Advancement of Automotive Medicine, 1985) were obtained. The ratings cover six dimensions of functioning: cognitive, mobility, bending/grasping/lifting, sensory, cosmetic, and pain. Second, data were added about the probability of permanent work-related disability by diagnosis. Third, guided by surveys of the general population, the functional capacity losses were converted into estimates of the percentage loss in quality of life, measured on a QALY scale. Fourth, the QALY losses were valued at $113,150 per QALY, computed using a 3% discount rate from the value of statistical life. To avoid double counting, this value and the associated QALY loss estimate excluded work loss.
We converted all costs to 2001 dollars using a health expenditures index for medical costs, a wage index for wage and quality-of-life loss, and the Consumer Price Index for other items. Future costs are stated in present value at the 3% discount rate recommended by the Panel on Cost-Effectiveness in Health and Medicine (Gold et al., 1996).
Traffic crashes
Although the sale of alcohol to minors under 21 is now illegal in all states, almost 20% of traffic crashes with a driver under age 21 involved youth drinking (Miller et al., 1998). Tabulating Fatality Analysis Reporting System census data (National Highway Traffic Safety Administration, 2003) on fatal crashes in 2001 suggested that alcohol-involved underage drivers and nonoccupants accounted for 12.8% of alcohol-impaired driving costs. Substantial evidence indicates that alcohol use increases motor vehicle occupant, pedestrian, and cyclist risk (National Institute on Alcohol Abuse and Alcoholism [NIAAA], 1997; Zador, 1991). An estimated 91% of crashes involving drivers with a .10 blood alcohol concentration (BAC) or higher, 43.5% of crashes involving drivers with a BAC of .08 to .099, and 24% of crashes involving drivers with a BAC level of .01 to .0799 were directly attributable to drinking (Levy and Miller, 1995; Miller et al., 1999; Reed, 1981)....
|
