A New System of Mobile Evidential Breath-alcohol Testing has Improved the Fight Against Drunk Driving in Sweden C Bergman K Berglund L Andersson AW Jones Swedish National Laboratory of Forensic Science - SKL, Breath Alcohol Control Unit, SE- 581 94 Linköping, Sweden Background Law enforcement regarding drunk driving has developed progressively over many years. Punishable limits of blood-alcohol concentration were introduced in Sweden in 1941 and were set at 0.8 per mille (80 mg/100 ml). Over the years the legal alcohol limit for driving has been lowered and now stands at 0.2 per mille (20 mg/100 ml) in blood or 0.10 mg alcohol per litre breath (1, 2). Measuring the concentration of alcohol in blood or breath provides a more objective way to assess over-consumption of alcohol and fitness to drive safely. Until 1981 a physician was required to take a blood sample and to conduct various clinical tests to judge whether the suspect was under the influence of alcohol or drugs. A typical drink driving case meant that police resources were occupied for several hours including transporting the suspect and waiting for the doctor to complete the examination and take a blood sample. Agreement between the person s blood-alcohol concentration and the degree of impairment reported by the physician was not always satisfactory. Accordingly, the mandatory medical examination of every drunk driving suspect was abolished in 1981. Today, the apprehension and arrest of a drunk driver starts with a roadside breath-alcohol screening test and if positive, the suspect is transported to a police station for an evidential breath-alcohol test (EBT), or to a hospital where a blood sample is taken for laboratory analysis. EBT was introduced in Sweden in 1989 (2), and the evidential breath analyser (EBA) Intoxilyzer 5000S was approved for legal purposes. This meant added responsibility for the police to produce the evidence necessary for prosecution. Transport of suspects from the place of arrest at the roadside to the nearest police station requires considerable police time and resources, especially in sparsely populated areas of the country. To permit more efficient EBT, it became clear that a more flexible system was needed. The Swedish police authorities requested the possibility to conduct the evidential breathalcohol test at the roadside, closer to the time of driving. This change would improve the prosecution rate by reducing the number of positive roadside screening tests that were below the legal limit for driving by the time the EBT was made. Objectives Here we describe the development of a new system of EBT with Evidenzer, which can be used at a police station and in a police vehicle at the roadside. The instrument has several new analytical features to enhance reliability when used in-the-field closer to the time of driving. The mobile use of the new EBA at the time of writing is still in a developmental stage.
Stationary Evidential Breath-alcohol Testing From 1989 to 2003 about 163 000 evidential breath-alcohol tests have been performed with the stationary breath-alcohol instrument Intoxilyzer 5000S. During this time, both advantages and disadvantages of the EBT system have emerged. First, introduction of the Intoxilyzer simplified and speeded-up the testing because sampling of blood was not necessary (2). Second, obtaining an on-the-spot result of the EBT meant that suspects knew whether they were over the legal limit of 0.10 mg alcohol per litre breath. This allowed the police to decide whether the licence should be taken into custody. The Intoxilyzer was installed at 145 police stations scattered throughout the country. (Figure 1). This number of instruments should be considered in relation to the size of the country and the population of 8.9 million inhabitants (1, 2). SWEDEN Population: 8.9 million Area: 449 964 km 2 EBA location Stockholm Gothenburg Malmo Figure 1. Map of Sweden showing location of the 145 Intoxilyzer instruments and the three largest cities (Stockholm, Gothenburg and Malmo). According to calculations made by the Swedish National Road Safety Administration, there are roughly 16,000 drunk drivers travelling Swedish roads daily. Even though the Intoxilyzer 5000S made it considerably easier and faster to test drunk drivers, there were still problems in effectiveness owing to the long distances to travel when suspects were apprehended in some parts of the country. The long distance between police stations and instruments was one of the reasons for the need of blood tests. Furthermore, blood samples will always be needed because a person cannot be forced to blow into a breathalcohol analyser. Some drunk drivers refuse to co-operate with the breath test procedure and fail to provide two successive samples. Other suspects are too drunk to provide an appropriate specimen, and those involved in road-traffic crashes where the driver requires emergency hospital treatment also provide blood for analysis of alcohol. The need to transport a suspect to a location where there is a stationary breath-alcohol instrument has implications for the number of prosecutions and convictions. The vast majority of apprehended drivers are in the elimination phase of alcohol metabolism and studies have shown that drunk drivers eliminate alcohol at a rate of 0.019 per mille per hour (19 mg/100 ml) on average (3). The time difference between the positive roadside screening test and the EBT brings many suspects below the legal limit for driving by the
time the Intoxilyzer test is made. After making a deduction for uncertainty (0.14 per mille or 0.07 mg/l), about 20 % of all positive screening tests are below the legal limit for driving. These 20 % are therefore classified as false positive roadside screening tests and are troublesome for the police. The number of blood samples compared with EBT and the total number of apprehended drivers during the past 13 years is shown in Figure 2. Figure 2. Development in number of blood and breath samples from suspects apprehended in Sweden 1990-2003. 30000 25000 20000 15000 10000 Total Breath-tests Blood samples 5000 0 Advantages With the New Evidenzer Instrument In May 2002, the Swedish police started to use a new EBT instrument, Evidenzer (Figure 3). About 6000 tests have been made so far up to December 2003. There are 21 police districts in Sweden and 200 Evidenzers are currently used throughout the country. The instrument combines several features that make it unique. Figure 3. The new evidential breath-alcohol analyser used in Sweden (Evidenzer). Mobility One of the important new features with Evidenzer is that it is intended both for stationary and mobile use. This quantitative infrared analyser is smaller and more robust than the Intoxilyzer. Hopefully the Evidenzer will also work properly on snowmobiles enabling EBT in the remote mountains in the north of Sweden. Detection of interfering substances and mouth alcohol Both Intoxilyzer and Evidenzer employ the same measuring principle for determination of alcohol, namely absorption of infrared light at fixed wavelengths (1, 2, 4). The Intoxilyzer incorporated three filters with 3.80 µm being used as a reference, whereas ethanol molecules are registered at 3.48 µm, and the filter at 3.39 µm is used to detect potential interfering substances, e.g. acetone, ether and methanol (4).
The measuring principle in the Evidenzer is also infrared but with five filters (3.37 µm, 3.41 µm, 3.47 µm, 3.52 µm and 3.80 µm), which select the wavelengths reaching the detector for quantitative and qualitative determination of ethanol. This gives a high selectivity for identifying ethanol without disturbances from interfering substances, which is an important requirement when testing in different environments eg at the roadside, where there is risk of interference from organic solvents, such as windscreen cleaning fluids, petrol fumes etc. Furthermore, the Evidenzer also includes improved algorithms for detecting mouth alcohol because a test might be made closer to finishing the last drink. The ambition with use of mobile testing is to eliminate the need for a 15 minutes observation and deprivation period. However, the main safeguard against mouth alcohol is a 6 minutes time gap between two breath samples. The police officers can use this time for completing various forms in the computer necessary for any legal proceedings. National database Besides the enhanced analytical capacity the possibility of storing data from each EBT in a national database is of great importance for quality assurance purposes. For example, each breath ethanol profile is stored resolved into 50 data points each. The stored data makes it possible to reconstruct the most vital parts of every EBT, which is valuable, particularly when the reliability of the EBT result is questioned or challenged. The system also includes software for authentication of operators and search options for retrieving statistics from the EBTs. Time saving Before 1981, when a blood sample and clinical examination was made, the time between a positive breath screening test and the evidential test was on an average 2.5 hours. This time was reduced to about 1.5 hours when nurses were allowed to take the blood sample, and the clinical examination was abolished. When evidential breath-alcohol testing was introduced in 1989, the time between arrest and the evidential test was reduced to about 1 hour. In 2003 mobile testing with Evidenzer meant that the first EBT could be done as soon as possible (5-10 minutes) after a positive screening test result, which has increased the prosecution rate because fewer offenders are now below the legal limit. Furthermore, thanks to new computer software, the time needed to process each drunk driving case and the paper work required has been reduced to a minimum. User-friendliness The improved breath-sampling characteristics of the Evidenzer has meant that fewer suspects fail to provide a breath sample as they do not have to blow as hard into the Evidenzer compared the Intoxilyzer 5000, something that caused problems for some suspects, especially elder women (Figure 4). Insufficient samples (%) 20 15 10 5 0 <40 40-60 >60 Intoxilyzer, males Intoxilyzer, females Evidenzer, males Evidenzer, females Age class (year) Figure 4. Proportion of insufficient samples for Intoxilyzer 5000S (n = 13,143) from 2002 and Evidenzer (n = 10,310) from June 2002 to February 2004.
Calibration check A calibration check was made before and after every EBT with Intoxilyzer 5000. However, with the improvements made to the Evidenzer software, it has been deemed practical to conduct the control test once a week. The check comprises 8 single tests of wet-gas (0.50 mg/l) from a modified Guth 34C simulator. The average of the five last results is used as a measure of instrument calibration. Results between 0.47 to 0.53 mg/l are considered as acceptable. A wet-gas simulator is not suitable for use in police vehicles because the simulator is sensitive to changes in temperature, which has required developing another method to check calibration. During an EBT with the Evidenzer instrument an optical filter is introduced to reduce the infrared light to the same amount as 0.50 mg alcohol per litre of air. This filter serves to check the calibration when every suspect is tested. Calibration control with a wet-bath simulator is done weekly by a special trained police officer. Running-in Problems The running-in problems have been acceptable considering that Evidenzer is a part of a new and different concept. A majority of the police officers working with traffic control are satisfied with the new system. However, a few officers have expressed the opinion that they preferred the Intoxilyzer system, because it was in some ways simpler and that the new weekly calibration checks necessary with Evidenzer takes too much time. Their complaints are understandable although the new system saves time it is also more complex, and a simplified calibration check procedure needs to be developed. The Evidenzer measuring unit needs to be connected to a desktop or laptop PC to be functional. The PC must use the operating system Windows NT4 or Windows XP. This concept has caused some trouble during the running-in period. Different police stations use PCs of different brands and software, and the computers are connected to the nationwide police network. Some police districts hesitated for a long time to release the new software into their network, a fact that delayed the introduction of the Evidenzer system. Software problems have been solved but there are, and will probably always be, plenty of room for software improvements to make the system user-friendlier. Implementing this new system required a carefully planned training programme lasting two days at each EBA location with instructions and demonstration of the instrument capabilities and computer software. A few policemen were reluctant to learn the new more complex system and felt unsure when dealing with computers. Deduction for Uncertainty The enforcement of concentration per se statutes and the penalties imposed for those found guilty requires a compensation for random and systematic variations in the methods of alcohol analysis. With the Intoxilyzer 5000S, a constant deduction of 0.07 mg/l (0.14 per mille) was made from the average of two tests. The same deduction is currently used in Evidenzer but this will shortly be changed. Instead, the deduction will be a function of the measured alcohol concentration and will contain a component for undetected interfering substances and undetected mouth alcohol. With a lower deduction for uncertainty the number of suspects liable to prosecution will increase. Conclusions Sweden s new portable instrument for evidential breath-alcohol testing (Evidenzer) incorporates several novel features and has rationalised police procedures in the battle against drunk driving. The most important new feature is the fact that testing can be done at the roadside in close proximity with the driving. Hopefully, the Evidenzer will work
properly also on police patrol boats and snowmobiles. Transport of a suspect from the place of arrest to the nearest police station often required considerable police time and manpower, especially in sparsely populated parts of the country. This time-delay meant that many suspects above the legal limit according to the roadside test were below the legal limit for driving by the time the EBT was made. The vast majority of apprehended drivers are in the elimination stage of alcohol metabolism and blood-alcohol drops by 0.19 per mille per hour (3). Conducting the EBT closer to the time of driving will increase the prosecution rate of drunk drivers. Introduction of Evidenzer has also simplified the paperwork, which is an additional time saving for the police. The improved infrared analysis gives a higher selectivity for identifying presence of interfering substances. Furthermore, new algorithms and a sample time space of 6 minutes have been introduced to detect mouth alcohol, which is necessary if the 15 minutes observation period is eliminated. The enhanced analytical precision also permits a smaller deduction for uncertainty, which will now be a function of the person s breath-alcohol concentration. The improved breath-sampling requirement of the Evidenzer has meant that fewer suspects fail to provide a sample. The new Evidenzer instrument has been well received so far by the majority of the police working with traffic controls. With the exception of some technical problems in the beginning, related to computers, software, instruments and simulators the police have confidence in the new instrument. With a more efficient instrument, time and money can be saved, more suspects can be prosecuted and fewer victims will hopefully be affected by the crime of drunk driving. References 1. Bergman C, Berglund K, Andersson L, Jones AW. Ten years experience of evidential breath-alcohol testing in Sweden. 15th International Conference on Alcohol, Drugs and Traffic Safety, T-2000 Stockholm 2000, pp 271-277. 2. Jones AW, Andersson L, Kopp I. Enforcement of drunk driving laws in Sweden with blood and breath alcohol concentrations as evidence for prosecution. Kriminalistik und forensische Wissenschaften 82;11-26, 1994. 3. Jones AW, Andersson L. Influence of age, gender, and blood-alcohol concentration on the disappearance rate of alcohol from blood in drinking drivers. J. Forensic. Sci. 41;922-926, 1996. 4. Jones AW, Andersson L Comparison of ethanol concentrations in venous blood and end-expired breath during a controlled drinking experiment. Forensic Sci Intern. 132 18-25, 2003.