Topic:
TRAFFIC ACCIDENTS; DEATH; DRUNK DRIVING;
Location:
DRUNK DRIVING; TRAFFIC ACCIDENTS;

OLR Research Report


February 9, 1999

 

99-R-0154

BLOOD ALCOHOL LEVELS, FATAL ACCIDENTS, AND SANCTIONS AGAINST DRIVERS WITH HIGH BACS

 
 

By: James J. Fazzalaro, Principal Analyst

You asked for information showing the fatality rates for drivers with high Blood Alcohol Concentrations (BAC) and, if possible, for serious injuries they cause. You specifically wanted to know the comparative number of fatalities at BACs above 0.15 compared to those between 0.05 and 0.10. You also wanted to know what some other states are doing to address drunk driving offenders with BACs above 0.15.

SUMMARY

Several limitations to national databases on accident statistics limit the response to your questions to fatal accidents. Data on accidents producing only injuries are significantly less reliable than for fatal accidents, particularly with respect to the level of alcohol. Also, these statistics cannot be interpreted to imply a causal relationship between alcohol use and other attributes of fatal accidents. The National Highway Traffic Safety Administration (NHTSA), which maintains the most widely used data system for accident analysis, cautions that alcohol-related accident statistics must be interpreted carefully.

We identified more than 120,000 statistical records on drivers involved in fatal accidents in NHTSA's Fatality Analysis Reporting System for which results of alcohol content tests were known and in the database. The data cover 1994 through 1997. Approximately 49% of the drivers were listed as having test results showing .00 BAC. Other distributions were 3% in the .01-.04 BAC range, 5% in the .05-.10 BAC range, 7% in the .11-.15 BAC range, and 18% with BAC results of .16 or above. When the portion of the sample containing drivers with .00 BAC results was factored out, almost 54% of the remaining drivers had reported results of .16 or more. Approximately 21 % of drivers with positive BAC results were in the .11-.15 range, 17% were in the .05-.10 range, and 9% were in the .01-.04 range.

At least seven states currently have special penalties and other sanctions that apply at specific BAC levels above the level determining intoxication under state law. The high BAC level triggering the separate sanctions may be .15, .16, or .20 depending on the state. Although many states allow courts to prescribe ignition interlock devices as a drunk driving sanction, two states require such devices for offenders at or above BACs of .15 (Kansas) or .18 (Montana).

STATISTICAL SOURCES FOR THE ANALYSIS

NHTSA's Fatality Analysis Reporting System

The Fatality Analysis Reporting System (FARS) database is one of NHTSA's two primary data systems for compiling accident statistics. FARS contains statistics only on accidents that result in at least one death to a vehicle occupant or nonmotorist within 30 days of the incident. Because its data comes from specifically trained analysts submitting information on qualifying accidents in a consistent format under cooperative agreements with one agency in each state, it provides a fairly complete statistical record of such accidents and is the most frequently cited source of information on this subject. Data entry includes more than 100 coded information items. Nonfatal injuries occurring during fatal accidents are also entered into FARS, but since FARS is limited to fatal accidents, it cannot provide a comprehensive picture of all injuries resulting from traffic accidents.

NHTSA uses a second system called the General Estimates System (GES) to augment FARS, particularly with respect to accidents involving nonfatal injuries or property damage only. GES estimates are made from a random sample of roughly 50,000 police-reported accidents resulting in property damage, injury, or death. Data is collected from about 400 police jurisdictions at 60 collection sites throughout the United States. National estimates are made from this sample.

Your inquiry requires analyzing the best available data using specific BAC levels entered in the database. Since BAC test results are not included in the GES dataset, it could not be used. Also, since GES relies on statistical methods for estimating data from the random sample, it is less precise and more subject to estimating error than FARS. Because FARS is more comprehensive and the data can be analyzed using BAC-specific measures, this report uses only the FARS data.

Interpreting Alcohol-Related Traffic Accident Data

Alcohol-related data in FARS must be interpreted carefully. NHTSA considers a fatal accident to be alcohol-involved or alcohol-related if either the driver or a nonmotorist (usually a pedestrian) has a measured or estimated BAC of .01 or more. A nonfatal accident is considered alcohol-involved or alcohol-related whenever a police accident report indicates there is evidence of alcohol present. This does not necessarily mean that a driver, passenger, or nonoccupant was actually tested for alcohol.

FARS includes the specific results of blood alcohol tests of drivers involved in fatal accidents when they are available. But no state reports a specific result for every driver and there is considerable variation in the percentage of reporting among the states due to various legal, technical, and economic issues. While a few states report results from more than 90% of fatally injured drivers, several others report results from fewer than 50%. The nationwide average reporting rate is about 70% for fatally injured drivers and only around 25% for surviving drivers. NHTSA uses a statistical model to assign accidents for which actual BAC test results are not known to one of three groups: sober drivers (.00 BAC), drinking drivers (.01-.09 BAC), and intoxicated drivers (.10+ BAC). Since it has no actual information on the presence or absence of alcohol in these instances, NHTSA's model assigns these accidents a place in one of the three categories based on other surrogate characteristics of the accident such as driver age and sex, crash time, accident type, vehicle type, and police alcohol indication. For example, a single-vehicle crash involving a male driver under age 25 and occurring after midnight is likely to be assigned to the intoxicated driver group because these surrogate factors are frequently present in accidents for which known BACs may be at or above .10.

NHTSA acknowledges that these methods result in producing estimates, not exact counts and the possible errors in these estimates is not known precisely. By applying statistical validation tests to the data, NHTSA concludes that the error in any one estimate is relatively small and does not appreciably affect broader comparisons of things like historical trends, but individual comparisons between states must be made carefully.

NHTSA also cautions that its various statistical tabulations, including FARS, cannot be interpreted to imply a “direct causal relationship between alcohol use and any other attribute of fatal crashes. Inferences concerning causality can only be made on the basis of additional information that is independent of the FARS data” (Alcohol Involvement in Fatal Crashes—1996, DOT HS 808 686, February 1998, p.1).

The estimated BAC groupings using NHTSA's statistical model could not answer your questions because (1) specific BACs were not certain and (2) the three groupings were insufficient to separate drivers into the ranges in which you were interested. Instead, we concentrated our analysis on the significant portion of the FARS records containing known alcohol test results. Since four full years of such information was available, we were able to assess more than 120,000 driver records with specific BAC test results which we believe provides a representative sample for analyzing BAC distributions.

GENERAL FARS FATAL ACCIDENT DATA

The fatality data is presented in Table 1. It shows fatal accident statistics for the period from 1994 through 1997. The total number of people involved in fatal accidents is broken out separately by the numbers of drivers, passengers, and all other types of people involved in the accidents.

Table 1: General Fatal Accident Data in the FARS

 

1994

1995

1996

1997

TOTAL

Number of Fatal Accidents

36,254

37,241

37,494

37,280

148,269

Number of Involved Vehicles

54,911

56,524

57,347

56,978

225,760

Number of People Involved

98,945

102,103

103,347

102,023

406,418

Motor Vehicle Operators Involved (includes motorcycles)

54,549

56,164

57,001

56,602

224,316

Vehicle Passengers

36,898

38,252

38,913

38,108

152,171

Pedestrians

5,989

6,150

5,925

5,774

23,838

All Others

1,509

1,537

1,508

1,539

6,093

Fatalities

40,716

41,818

42,065

41,967

166,566

Table 2 provides fatality and injury information broken out by type of person involved (drivers, passengers, and pedestrians only) for the four years from 1994 through 1997.

Table 2: Fatalities and Injuries by Type of Person

 

Fatally Injured

Incapacitat-ing Injury

Nonincapacitating and Minor Injuries

Unknown Injuries and Other Categories

No Injury

Drivers

         

1994

23,691

7,572

9,944

657

12,685

1995

24,390

7,879

10,396

670

12,829

1996

24,534

7,752

10,746

741

13,228

1997

24,644

7,514

10,565

880

12,999

Passengers

         

1994

10,518

9,114

10,291

224

6,751

1995

10,782

9,361

10,926

265

6,918

1996

11,058

9,469

10,993

289

7,104

1997

10,931

9,409

10,762

358

6,648

Pedestrians

         

1994

5,489

266

220

3

11

1995

5,585

293

258

5

9

1996

5,449

243

213

10

10

1997

5,307

259

197

3

8

DRIVERS INVOLVED IN FATAL ACCIDENTS AND KNOWN BACs

We searched the FARS data to identify all motor vehicle operators who were involved in fatal accidents for whom there were known blood alcohol content test results in the accident information in the database. This search identified 224,316 drivers in the FARS data. Chart 1 shows the percentage distribution of drivers within the total group by specific BAC groupings.


Chart 1

Of the total of 224,316 drivers involved in fatal accidents during these four years, 44% were not given any BAC tests, 27% were given tests that yielded results of .00 BAC, 2% had results of .01-.04, 3% had results of .05-.10, 4% had results of .11-.15, and 10% had results of .16 or above. Some drivers, fewer than 1% of the total, were reported as having refused the BAC test and about 10% of the total driver records reflected either unknown data or unknown results of an administered alcohol test.

Chart 2 removes the records for the 101,882 drivers involved in fatal accidents, but not subjected to blood alcohol tests. It shows the BAC test result distributions for the remaining 122,434 drivers. This demonstrates how known BACs aggregate among those drivers for whom actual results have been entered in FARS. For 60,001 drivers involved in fatal accidents (49% of the total) results showed no alcohol present (.00 BAC). Another 3% of the total group (3,616 drivers) showed results in the .01 to .04 range; 5% (6,595 drivers) showed results in the .05 to .10 range; 7% (8,275 drivers) showed results in the .11 to .15 range; and 18% (21,560 drivers) showed results of .16 or above. Data is unknown or incomplete for about 18% of the total group.


Chart 2

Chart 3 provides a third level of refinement relating to known BAC results by presenting a distribution of them among the 40,046 drivers for whom there were positive BAC test results (ranging from .01 to more than .94 BAC). Within this positive BAC records group, almost 54% (21,560 drivers) had BAC test results of .16 or more. Approximately 21% (8,275 drivers) had BAC results in the .11 to .15 range; almost 17% had results in the .05 to .10 range; and 9% had results in the .01 to .04 range. The stacked bars in the chart show the number of drivers at each BAC individually by year. The overall height of the bars shows the cumulative number at each BAC for all four years combined.











SANCTIONS IN OTHER STATES FOR PEOPLE WITH HIGH BACs

Some states have recently established more stringent sanctions for drunk drivers whose BACs exceed levels defined as excessive. Traditionally, license suspensions, fines, and other penalties have been based on the number of prior offenses of the same type. According to information compiled by the National Conference of State Legislatures, a few statesΧFlorida, Idaho, Minnesota, Maine, New Hampshire, New Mexico, and Washington—have augmented their penalties by specifying a BAC at or above which higher or additional sanctions are deemed to apply.



Florida defines a BAC of .20 as excessively high. Violators with BACs at or above this level face higher criminal penalties than offenders below .20. These high BAC offenders also must get a mandatory clinical assessment and evaluation to determine the nature and extent of any possible alcohol problem. These offenders apparently also fall under the prohibitions of an anti-plea bargaining statute and may also face additional sanctions with respect to substance abuse education and treatment, community service, and vehicle impoundment (FSA 316.193).

Idaho law also establishes special sanctions for offenders with BACs at or above .20. A first offense at this level is considered a misdemeanor and carries additional mandatory jail time, a $2,000 fine and a driver's license suspension for a minimum of one year during which no driving privileges can be granted. These sanctions are in addition to the regular penalties for drunk driving. A second conviction at .20 or above within five years of the first constitutes a felony and may result in a five-year jail sentence, a $5,000 fine, and up to a five-year license suspension in addition to any other standard drunk driving penalties that apply (IC 18-8004C).

Minnesota classifies drunk driving offenses at a BAC of .20 or above as a gross misdemeanor or enhanced gross misdemeanor, depending on the number of previous convictions for drunk driving. Convictions at these higher levels result in (1) a license suspension that is double what would otherwise apply, (2) the offender having to wait longer to qualify for limited license privileges, and (3) additional fines and higher mandatory minimum jail sentences. These apply in addition to the standard penalties (M.S.A. 169.121).

In Maine, special sanctions are triggered at the .15 BAC level. A first offense at this level requires mandatory imprisonment for 48 hours, a fine of at least $400, and a 90-day license suspension. A second such offense within 10 years carries a mandatory term of imprisonment of seven days, a fine of at least $600, and an 18-month license suspension. A third such offense within 10 years results in a 30-day mandatory jail term, a fine of at least $1,000, and a four-year license suspension. A fourth or subsequent such offense within 10 years results in a six-month mandatory jail sentence, a fine of at least $2,000, and a six-year license suspension (29-A MRSA 2411).

New Hampshire considers a BAC of .16 or higher as a separate offense of aggravated drunk driving. This results in a higher fine (a $500 minimum instead of a $350 minimum) and a longer minimum license revocation (one year instead of 90 days). A second such offense within seven years carries a special penalty of not less than 10 consecutive days incarceration (the first three 24-hour periods in a county correctional facility and the next seven 24-hour periods in the state-operated multiple DWI offender intervention detention center established by law for this purpose (RSA 265-82a).

New Mexico considers offenses at .16 or above as the separate felony offense of aggravated drunk driving. Offenders must serve a mandatory period of 48 hours in jail in addition to all of the standard incarceration, fine, community service, and screening and treatment sanctions that otherwise apply. A second aggravated drunk driving offense carries a mandatory 96-hour incarceration period, a third carries a period of 60 days, and a fourth carries a mandatory period of six months. DWI offenses at below the .16 level are not classified as felonies until the fourth conviction (NMSA 66-8-102).

Washington applies special penalties at or above the .15 BAC level. A first offense results in imprisonment for two days to one year, a fine of $500 to $5,000, and a 120-day license suspension. For a second such offense within five years the term of imprisonment is from 45 days to one year, the fine is from $750 to $5,000, and the license suspension is for 450 days. For a third such offense the penalties are from 90 days to one year imprisonment, a fine of $1,000 to $5,000, and a two-year license suspension. These offenders are also subject to special alcohol assessment and treatment requirements (RCWA 46.61.5055).

Besides Florida, Maine, New Mexico, and Washington, the states of California, Delaware, Iowa, Nevada, and South Dakota require that drunk drivers above certain specified high BACs undergo alcohol abuse evaluation or participate in mandatory treatment for problem drinking. Although numerous states allow courts to prescribe installation of ignition interlock devices as a sanction for some drunk driving offenders, Kansas and Montana make such devices mandatory for convictions at or above the .15 and .18 levels respectively.

JJF:pa