The impact of a national alcohol policy on deaths due to transport accidents in Russia

The impact of a national alcohol policy on deaths dueto transport accidents in Russia

William Alex Pridemore1, Mitchell B. Chamlin2, Maria T. Kaylen3 & Evgeny Andreev4

Department of Criminal Justice and Criminology, Georgia State University, Atlanta, GA, USA,1 School of Criminal Justice, Texas State University–San Marcos, SanMarcos, TX, USA,2 Department of Criminal Justice, Indiana University, Bloomington, IN, USA3 and Center for Demographic Research, New Economic School,Moscow, Russia4

ABSTRACT

Aims To determine the impact of a suite of 2006 Russian alcohol control policies on deaths due to traffic accidents inthe country. Design, setting and participants We used autoregressive integrated moving average (ARIMA) inter-rupted time–series techniques to model the impact of the intervention on the outcome series. The time–series began inJanuary 2000 and ended in December 2010. The alcohol policy was implemented in January 2006, providing 132monthly observations in the outcome series, with 72 months of pre-intervention data and 60 months of post-intervention data. Measurements The outcome variables were the monthly number of male- and female-specificdeaths of those aged 15+ years due to transport accidents in Russia. Findings The 2006 set of alcohol policies hadno impact on female deaths due to traffic accidents (ω0 = −50.31, P = 0.27). However, the intervention model revealedan immediate and sustained monthly decrease of 203 deaths due to transport accidents for males (ω0 = −203.40,P = 0.04), representing an 11% reduction relative to pre-intervention levels. Conclusion The implementation of thesuite of 2006 Russian alcohol control policies is partially responsible for saving more than 2400 male lives annuallythat would otherwise have been lost to traffic accidents.

Keywords Alcohol consumption, alcohol policy, Russia, time–series analysis, traffic deaths.

Correspondence to: William Alex Pridemore, Department of Criminal Justice and Criminology, Georgia State University, Atlanta, GA 30302, USA.E-mail: [email protected] 2 October 2012; initial review completed 4 January 2013; final version accepted 18 July 2013

INTRODUCTION

Russia experiences about 30 000 fatalities due to trafficaccidents annually, for a death rate of 21 per 100 000residents. This rate is five times higher than nations withthe safest roads and nearly twice the European Unionaverage [1,2]. Direct and indirect costs associated withtraffic accidents are high. Direct costs include emergencytreatment, long-term care, insurance administration andlegal actions. Indirect costs include loss of productivity atwork and home, property damage and travel delays.These costs accrue both to individuals involved andsociety. As President Medvedev pointed out in 2009: ‘Thenational economy lost US$175 billion from traffic acci-dents over the past five years. That is comparable withoverall health care expenditures of the same period’ [1].In 2005, these traffic accident costs were approximately2.5% of the country’s gross domestic product (GDP),and the traffic fatality burden falls primarily on the

most economically productive group, those aged 15–44years [2].

Discussion of traffic fatalities in Russia is incompletewithout mentioning their strong connection to alcohol.Alcohol consumption was responsible for at least 10% oftraffic deaths in Russia in 2008 [2]. There are reasonsrelated to infrastructure and engineering (e.g. road con-ditions, automobile design, pedestrian safety, emergencymedical response) as to why the proportion of all trafficdeaths defined as due to alcohol is lower in Russia than inother nations. However, alcohol-related traffic fatalitiesare probably substantially underestimated due to manyfactors, including legislative definitions [2]. Further, in acase–control study of more than 48 000 adult deaths inRussia between 1990 and 2001, alcohol consumptionwas associated strongly with deaths from transport acci-dents for all levels of excess drinking, and those with thehighest volume of consumption were more than fourtimes as likely to die from transport accidents relative to

RESEARCH REPORT

bs_bs_banner

doi:10.1111/add.12311

© 2013 Society for the Study of Addiction Addiction

mailto:[email protected]

reference group drinkers [3]. Driving under the influenceof alcohol is a risk factor for traffic accidents and is alsoassociated with other risk factors, including excessivespeeding and not wearing a seatbelt [2]. Young males areparticularly vulnerable. In Europe, male drivers aged21–24 with a blood alcohol concentration of 0.04–0.05 g/dl are almost twice as likely to crash as malesaged 25–29 with the same blood alcohol concentration[4].

Russia: hazardous drinking and the 2006alcohol policy

The volume of alcohol consumption in Russia is amongthe highest in the world, and alcohol researchers place itin the highest risk category based on the pattern of haz-ardous drinking by Russians [5,6]. Adult per capitaalcohol consumption reached nearly15 liters of pureethanol alcohol annually [7,8], and alcohol-relatedmortality was estimated at 500 000–750 000 annually[9,10]. Hazardous drinking has been considered theunderlying cause of wide swings in Russian mortalitysince the early 1990s, which was driven largely byalcohol-related mortality among middle-aged males[11,12]. Leon et al. estimated that 43% of all prematuremortality among working-age Russian males was due tohazardous drinking [13], with other research revealingan association between alcohol and external causes ofdeath other than traffic accidents, including homicideand suicide [14–16], and an association between percapita alcohol consumption and injury mortality inRussia [17].

On several occasions, the Soviet and Russian govern-ments have attempted to decrease levels of alcohol con-sumption via legislation. The most well-known wasGorbachev’s anti-alcohol campaign in 1985 that, amongother mechanisms, restricted sale hours, implementedpurchase quotas, closed distilleries and breweriesand destroyed vineyards [18]. Another comprehensivealcohol policy—the one of interest in this study—wasimplemented in 2006. Levintova [19] provides a detaileddescription of the policy and its creation. In brief, Presi-dent Putin signed a law on the regulation of productionand sale of ethyl alcohol and alcohol containing productson 21 July 2005, effective from 1 January 2006. Amongother items, the law (i) made it illegal to produce, distrib-ute and sell alcohol without a license, the cost of whichincreased substantially over prior licensing costs, (ii)required all production facilities to purchase recordingequipment regulated by law to measure the amount ofethyl alcohol used and produced, (iii) required theamount of alcohol and alcohol products used and pro-duced to be provided to the government, (iv) required allalcohol products to display an excise stamp, (v) prohibited

alcohol sales at certain types of sites such as educationaland athletic facilities and on public transport and (vi)prohibited the sale of alcoholic beverages containingmore than 15% ethanol at or near large public places,such as train and metro stations and wholesale marketsand in kiosks, and any places (e.g. by individuals, out ofautomobiles) not specifically licensed for such sales. Thenew financial investments required by alcohol producersand sellers (e.g. registration fees, equipment costs, excisestamps) led to fewer producers and distributors andincreased consumer prices [19].

Very recent analyses reveal an increase in life expec-tancy in Russia during the last decade accounted for par-tially by a decrease in alcohol-related deaths [20] anddeclines in all-cause mortality and deaths due to cardio-vascular disease and external causes that coincide withthe 2006 alcohol policy [21]. While the authors of thesestudies speculate that the policy may be partially respon-sible for these changes in life expectancy and mortality,this hypothesis has yet to be tested explicitly. Our study isamong the first to do so, and we focus on a specific andimportant type of mortality often found to be sensitiveto alcohol policies: deaths due to transport accidents(including single and multiple vehicle accidents andaccidents involving vehicles and pedestrians).

Prior studies provide reasons to believe that theRussian alcohol policy may have led to a reduction inalcohol traffic fatalities. For example, an analysis of 14European countries found a significant associationbetween per capita alcohol consumption and male- andfemale-specific traffic fatalities [22]. The countries weredisaggregated into three categories based on drinking cul-tures: beer countries of central Europe and the BritishIsles, spirits countries of northern Europe and wine coun-tries of southern Europe. The alcohol–traffic mortalityassociation was not significant in northern countries butwas in the other two groups. Similarly, Ramstedt found asignificant association between aggregate alcohol con-sumption and traffic fatalities for males and females inthe United States, Canada, central Europe and southernEurope, with the strongest associations for males in theUnited States and Canada and for females in Canada [23].The aim of the present study was to determine if the2006 Russian alcohol policy had an effect on deaths dueto traffic accidents.

METHODS

Data

The outcome variables were the monthly number ofmale- and female-specific deaths of those aged 15+ yearsdue to transport accidents in Russia. Counts wereappropriate because Russia’s total population was little

2 William Alex Pridemore et al.


changed between 2000 and 2010, with a population ofabout 146 million in 2000 and 143 million in 2010. TheRussian cause-of-death data and the civil death registra-tion system are based on medical death certificates,which are completed by a medical doctor (who eithertreated the deceased and/or established cause of death),the pathologist who completed the autopsy or a forensicmedical expert. For transport accidents, more than 90%of diagnoses come from forensic medical examination.Deaths are classified according to the International Classi-fication of Diseases, 10th revision [24]. For the purposes ofthis study, ‘transport accidents’ were operationalized aspedestrians injured in a transport accident (V01–V09),vehicle occupants injured in a transport accident (V40–V49), occupants of a pickup truck or van injured in atransport accident (V50–V59), occupants of a heavytransport vehicle injured in a transport accident (V60–V69) and bus occupants injured in a transport accident(V70–V79). Data were obtained from anonymous deathrecords collected by the Russian Federal State StatisticsService. The time–series begin in January 2000 andend in December 2010. The alcohol policy was imple-mented in January 2006, providing 132 monthlyobservations in each series, with 72 pre-interventionand 60 post-intervention observations.

Estimation procedures

We used autoregressive integrated moving average(ARIMA) interrupted time–series techniques to model theimpact of the intervention on the outcome series.Because ARIMA procedures are well established in theliterature on policy impact, including alcohol policy [25–28], we offer a discussion of only its most salient features.

A fundamental concern associated with the evalua-tion of the efficacy of legislative and administrative ini-tiatives is distinguishing their effects from other socialprocesses that may influence an outcome series. ARIMAtechniques, unlike simple pre- and post-interventionmean or percentage difference tests, take into account thepotentially confounding effects of other causal mecha-nisms, allowing one to assess the change in the level ofany outcome series independently of ongoing stochasticprocesses [29].

An ARIMA interrupted transfer function model con-sists of two parts. The first, the ‘noise’ component, usesinformation from prior observations of an outcome seriesto model the systematic variation (i.e. autocorrelation)within the series. By applying the appropriate seasonaland non-seasonal differencing, along with estimating theappropriate seasonal and non-seasonal autoregressiveand moving average parameters, one can separate theconfounding influences of other causal processes fromthose associated with the intervention.

Once a satisfactory noise component is identified andestimated, the intervention component is added to thetransfer function equation. If the inclusion of a dummyseries for the intervention (coded zero for the period priorto the intervention and one beginning with the observa-tion in which the intervention occurs) increases theexplanatory power of the model beyond that provided bythe noise component (i.e. Granger causality), we canconclude that the intervention significantly affects theoutcome series [29,30].

Another advantage of ARIMA modeling techniquesover pre- and post-intervention change scores is that theyallow one to examine the functional form of the relation-ship between an intervention and an outcome series.Crude mean and percentage difference tests assume theeffect of an intervention is an abrupt change in the levelof the outcome series that is sustained throughout theremaining observations. While one can estimate thisfunctional form as a zero-order transfer function usingARIMA techniques, one can also examine the relative fitof competing adjustment models. It is possible the effect ofan intervention gradually reaches a new level or that theeffect is instantaneous but short-lived. A first-order trans-fer function can be estimated to model the former pattern,while a pulse function can be estimated to model thelatter [29].

ARIMA model building is an iterative process. By suc-cessively estimating the noise and intervention compo-nents and subjecting them to diagnostic tests, a finalintervention model is derived. For the statistical detailsinvolved in the identification and estimation of the noiseand intervention components of ARIMA interruptedtime–series models we refer the reader to popularpublished sources [29,31].

RESULTS

Monthly male and female deaths due to transport acci-dents are shown in Fig. 1. Mean monthly transportdeaths for the entire period were 2474, 1772 and 702 fortotal population, men and women, respectively. Pre-intervention means were 2612, 1866 and 746. Post-intervention means were 2307, 1658 and 649. Theseraw data reveal decreases in mean monthly transportdeaths of 11% for males and 13% for females followingthe implementation of the 2006 alcohol policy. In orderto determine if these changes were due to the alcoholpolicy or simply to ongoing patterns in these time–seriesdata resulting from other causes, we estimated ARIMAmodels.

The left column in Table 1 provides information aboutthe form and statistical adequacy of the final univariatemodels for male and female time–series. This showsthat both male and female series required (i) first-order

Alcohol policy and traffic deaths in Russia 3


differencing and first-order seasonal differencing toremove drift and (ii) first-order moving average and first-order seasonal moving average parameters to removeautocorrelation. The Q statistics for the final noise modelsfor males and females met the criterion that none of theautocorrelations was significant at 0.05.

The right column of Table 1 shows the final transferfunction models assessing the impact of the 2006Russian alcohol policy on monthly male and femaledeaths due to transport accidents. The results suggested azero-order response best fitted the data for each series.The Q statistic is the Ljung–Box test statistic for the nullhypothesis that model residuals are uncorrelated. Themodels showed no impact of the alcohol policy onfemale deaths due to transport accidents (ω0 = −50.31,P = 0.27). The intervention model for men showed thatthe alcohol policy was associated with an immediatemonthly decline of 203 deaths due to transport accidentsfor males (ω0 = −203.40, P = 0.04) that was sustained tothe end of the observation series. The pre- and post-intervention means shown in Fig. 1 for males provide avisual display of the impact.

DISCUSSION

The aim of this study was to determine if Russian trafficfatalities decreased following the 2006 implementation ofa suite of policies aimed at reducing alcohol consump-tion. The results of our interrupted time–series analysisrevealed that the law was associated significantly with areduction in male but not female traffic fatalities. Specifi-cally, we found that the policy was associated with animmediate 11% reduction in male deaths due to trans-port accidents that was sustained to the end of the time–series, suggesting that the implementation of the 2006

Figure 1 Number of monthly male and female deaths due to transport accidents: Russia, 2000–2010.Alcohol law was introduced in January2006. Pre- and post-intervention means for males are represented by horizontal lines

Table 1 Final noise and intervention models for the effect of the2006 Russian alcohol policy on the number of monthly maleand female deaths due to transport accidents.

Noise model Intervention model

Male transport accident deathsARIMA (0,1,1)(0,1,1)12 Yt

* = at + ω0It

(1–θ1B)(1–θ12B12)at = (1–B)(1–B12)Yt

θ1 = −0.544, P < 0.001θ12 = −0.621, P < 0.001

It = 0 for observations 1–72It = 1 for observations 73–132ω0 = −203.40, P = 0.04Q = 26.8, df = 18, P = 0.09

Female transport accident deathsARIMA (0,1,1)(0,1,1)12 Yt

* = at + ω0It

(1–θ1B)(1–θ12B12)at = (1–B)(1–B12)Yt

θ1 = −0.442, P < 0.001θ12 = −0.559, P < 0.001

It = 0 for observations 1–72It = 1 for observations 73–132ωo = −50.31, P = 0.27Q = 18.4, df = 18, P = 0.43

θ = moving average parameter; at = current random shock; B = back-ward shift operator; ω0 = zero-order input parameter of a transfer func-tion; It = intervention series; Q = Ljung–Box test statistic for the nullhypothesis that the model residuals are distributed as white noise;Yt = current time series observation. ARIMA = autoregressive integratedmoving average.



Russian alcohol policy was partially responsible forsaving more than 2400 male lives annually that wouldotherwise have been lost to traffic accidents.

This finding is consistent with research in Russia andelsewhere showing an association between alcohol con-sumption and traffic deaths. Ramstedt found that a 1-literincrease in alcohol consumption is associated with anincrease of 4.4 male traffic fatalities per 100 000 males,but found no association for females [23]. He also foundthat alcohol contributed to 41% of traffic fatalities in theUnited States. Skog similarly found that a 1-liter increasein alcohol consumption in Austria is associated with anincrease of 4.9 male traffic fatalities per 100 000 malesaged 15+ [22]. Of the 14 European countries analysed bySkog, the effect of alcohol consumption on fatal trafficaccidents was highest for males in Austria followed byBelgium (3.03), Ireland (2.01) and Italy (1.21). Forfemales the effect was much smaller, with the highestbeing 1.12 female traffic fatalities per 100 000 femalesaged 15+. The World Bank estimated that alcohol con-sumption is responsible for at least 30% of traffic fatalitiesin Croatia, Estonia, Georgia and Slovenia in 2008, and atleast10% in Belarus, Hungary, Latvia, Moldova, Polandand Russia [2].

Our results are also consistent with evaluations ofalcohol policies in other countries. For instance, Changet al. summarized research in the United States thatexamined the effect of alcohol policies on traffic fatalities,concluding that alcohol control policies such blood–alcohol content and Administrative License Revocationreduce traffic fatalities [32]. In their own study, theyfound that beer taxes and zero tolerance laws were effec-tive in reducing alcohol-related traffic fatalities [32]. Spe-cifically, a 1% increase in beer taxes was associated with a0.37% decrease in alcohol-related traffic fatalities per100 000 residents. The difference in our findings formales and females in Russia is not unexpected. First,alcohol consumption rates in Russia are higher amongmales than females [7,13] and have a disproportionateimpact on negative health outcomes for males. Secondly,research from Europe and the United States shows thatthe population-level association between alcohol con-sumption and traffic accidents tends to be stronger formales than females [22,23].

The 2006 Russian alcohol law was actually a suite ofseparate policies that varied in the success and timelinessof implementation. For instance, there was a 6-monthextension on the requirement of imported alcohol prod-ucts having excise stamps and there were problems withthe use of the alcohol monitoring system [19]. Gil et al.found that in several Russian cities in the first half of2007 there remained ready access to cheap non-beveragesources of ethanol [33]. In spite of delays in some com-ponents of the policy, our results show the policy reduced

traffic fatalities. Nevertheless, this uneven implementa-tion suggests that an alternative functional form of thepolicy impact on deaths due to transport accidents isplausible. Specifically, instead of the immediate andenduring step change (i.e. a zero-order model) we found,the impact may have been more gradual. We estimated afirst-order model to test this possibility. This model pro-vided no evidence for such an effect, however, and thezero-order model fitted the data better. We also estimateda model containing two intervention points, January andJuly 2006, but found no impact of the latter. Further,government data reveal that while overall alcohol salesdid not begin to decline until after 2007, there was anearlier decline in sales of vodka and distilled spirits, andothers have shown a decline in overall alcohol consump-tion in Russia beginning in 2006 that was driven bydeclines in recorded and unrecorded consumption ofspirits [21]. These declines in sales and consumptionprovide further support for our findings, as do the con-comitant declines in suicide [34], liver cirrhosis [35] andcardiovascular disease and overall mortality found byothers [20,21]. This suggests that with all aspects of thepolicy completely in place, together with the implemen-tation of additional recent alcohol policies in Russia, wemight expect further reductions in male deaths due totraffic accidents and perhaps an impact upon femaledeaths.

Limitations

We used traffic fatality counts instead of standardizeddeath rates (SDR). It is not possible to standardize thesemonthly data because the corresponding population sizecannot be estimated reliably. There was only a smalldecline in the population (from 146 to 143 million)during this period and, more importantly, the populationof those most affected by the alcohol policy (i.e. thoseaged 15+) grew by about 2 million (from 119 to 121million). Further, as a sensitivity test we calculatedannual SDRs for transport deaths, and we found simulta-neous declines for transport death SDRs and counts.Finally, aside from population size another key contribu-tor to the denominator here would be number of cars andmiles driven. The decline in traffic deaths we foundoccurred during times of increasing motorization andmiles driven in Russia, although there was also a declinein the age of the car fleet and an increase in importedvehicles that increased automobile safety in the country[36].

A key threat to validity of the interrupted time–seriesdesign is history: the occurrence of other events aroundthe time of the intervention that may also influence thelevel of the outcome series. The Russian governmentimplemented the Federal Targeted Program ‘Improving



Road Safety in 2006–2012’ during the period of ourstudy. The goal was to improve traffic safety via road mod-ernization, education, information campaigns, increasedfines for driving without a seatbelt and anti-alcohol cam-paigns [1]. Unfortunately, we were unable to obtainmonthly data series that would allow us to examinealcohol-related traffic deaths specifically (and alcohol-related relative to non-alcohol-related traffic deaths) or tocontrol for the new road safety program. While wecannot rule out the effect on our outcome series of thisroad safety program, several items provide confidence inthe significant impact of the alcohol policy. First, whilethe road safety initiative may have resulted in the reduc-tion of traffic deaths, due to the type and pace of changesimplemented it is unlikely to have had the immediateeffect revealed by our analysis. For example, Russia’s roadsystem at that time was characterized as ‘intrinsicallyunsafe’, with ‘fragmented institutional processes’ toaddress its ‘growing road safety crisis’ [36]. Many of thekey road safety actions required improving infrastruc-ture, automobile safety and training, all of which taketime to carry out. Secondly, we found no significantdecline in female traffic deaths associated with this inter-vention point, which we would expect from road safetyimprovements. Thirdly, Russian police data show declinesof 35% in deaths due to excess alcohol crashes and 59%in excess alcohol traffic offenses between 2004 and 2008[36]. While the road safety program included actionsrelated to drinking and driving, the lesson is that alcoholpolicy reduces traffic-related harm. Finally, in analysesnot shown here we found immediate and sustainedreductions accompanying the alcohol policy for suicide[34] and alcoholic liver cirrhosis [35], and there is noreason for either of these two causes of death to be asso-ciated with the road safety program. Nevertheless, thisthreat to the validity of our findings must be kept in mindwhen interpreting our results, and thus we temper ourlanguage as to both a causal relationship and to the pro-portion of decline in transport deaths due specifically tothe alcohol policy.

CONCLUSION

A key goal of the 2006 Russian alcohol policy was todecrease consumption and hazardous drinking and toreduce the burden of alcohol-related harm. Further,Shkolnikov et al. [20] found recently that life expectancyat birth increased in Russia during the last decade, andNeufeld & Rehm [21] found that all-cause mortality anddeaths due to cardiovascular disease and external causesdeclined in the latter half of the last decade. The authorsof both studies hypothesize that the 2006 alcohol lawwas partially responsible for these improvements. Ourstudy is among the first of its kind to evaluate systemati-

cally the effect of this law, and we conclude that itwas associated with an annual reduction of more than2400 male deaths due to traffic accidents, indicating thatalcohol policy is among the levers that can be manipu-lated to reduce traffic fatalities in Russia and elsewhere.

Declaration of interests

None.

Acknowledgements

We thank Svetlana Nikitina of the Russian FederationFederal State Statistics Service for her help with organiz-ing the special data tabulation required for this study.

References

1. Marquez P. V., Bliss A. G. ECA Knowledge Brief: Dangerousroads: Russia’s safety challenge. Europe and Central AsiaKnowledge Brief, 27. 2010.

2. World Bank. Confronting ‘death on wheels’: Making roads safein Europe and Central Asia. 2009. Available at: http://siteresources.worldbank.org/INTECA/Resources/DeathonWheelsWeb.pdf [accessed on 15 August 2013] (Archived athttp://www.webcitation.org/6ItXg8OM8 on 15 August2013).

3. Zaridze D., Brennan P., Boreham J., Boroda A., Karpov R.,Lazarev A. et al. Alcohol and cause-specific mortality inRussia: a retrospective case–control study of 48 557 adultdeaths. Lancet 2009; 373: 2201–14.

4. Sethi D., Racioppi F., Bertollini R. Preventing the leadingcause of death in young people in Europe. J EpidemiolCommun Health 2007; 61: 842–3.

5. Gmel G., Rehm J., Frick U. Methodological approaches toconducting pooled cross-sectional time series analysis: theexample of the association between all-cause mortality andper capita alcohol consumption for men in 15 Europeanstates. Eur Addict Res 2001; 7: 128–37.

6. Rehm J., Monteiro M., Room R., Gmel G., Jernigan D., FrickU. et al. Steps towards constructing a global comparativerisk analysis for alcohol consumption: determining indica-tors and empirical weights for patterns of drinking, decidingabout theoretical minimum, and dealing with different con-sequences. Eur Addict Res 2001; 7: 138–47.

7. Treml V. G. Soviet and Russian statistics on alcohol con-sumption and abuse. In: Bobadilla L., Costello C. A., MitchellF., editors. Premature Death in the New Independent States.Washington: National Academy Press; 1997, pp. 220–38.

8. Nemtsov A. V. Estimates of total alcohol consumption inRussia, 1980–1994. Drug Alcohol Depend 2000; 58: 133–42.

9. Nemtsov A. V. Alcohol-related human losses in Russia in the1980s and 1990s. Addiction 2002; 97: 1413–25.

10. MacKellar L., Andrichina E., Horlacher D. Policy Pathways toHealth in the Russian Federation. Vienna, Austria: Interna-tional Institute for Applied Systems Analysis (IIASA); 2003.

11. Chenet L., McKee M., Leon D., Shkolnikov V., Vassin S.Alcohol and cardiovascular mortality in Moscow; new evi-dence of a causal association. J Epidemiol Commun Health1998; 52: 772–4.

12. Bobak M., Marmot M. Alcohol and mortality in Russia: is itdifferent than elsewhere? Ann Epidemiol 1999; 9: 335–8.



http://siteresources.worldbank.org/INTECA/Resources/DeathonWheelsWeb.pdf



http://www.webcitation.org/6ItXg8OM8

13. Leon D. A., Saburova L., Tomkins S., Andreev E., KiryanovN., McKee M. et al. Hazardous alcohol drinking and prema-ture mortality in Russia: a population based case–controlstudy. Lancet 2007; 369: 2001–9.

14. Pridemore W. A. Vodka and violence: alcohol consumptionand homicide rates in Russia. Am J Public Health 2002; 92:1921–30.

15. Pridemore W. A. Heavy drinking and suicide in Russia.Soc Forces 2006; 85: 413–30.

16. Pridemore W. A., Chamlin M. B. A time–series analysis ofthe impact of heavy drinking on homicide and suicide mor-tality in Russia, 1956–2002. Addiction 2006; 101: 1719–29.

17. Landberg J. Population drinking and fatal injuries in EasternEurope: a time–series analysis of six countries. Eur AddictRes 2010; 16: 43–52.

18. White S. Russia Goes Dry. Cambridge: Cambridge UniversityPress; 1996.

19. Levintova M. Russian alcohol policy in the making. AlcoholAlcohol 2007; 42: 500–5.

20. Shkolnikov V. M., Andreev E. M., McKee M., Leon D. A.Components and possible determinants of decrease inRussian mortality in 2004–2010. Demogr Res 2013; 28:917–50.

21. Neufeld M., Rehm J. Alcohol consumption and mortality inRussia since 2000: are there any changes following thealcohol policy changes starting in 2006? Alcohol Alcohol2013; 48: 222–30.

22. Skog O. Alcohol consumption and mortality rates fromtraffic accidents, accidental falls, and other accidents in 14European countries. Addiction 2001; 96(Suppl. 1): S49–S58.

23. Ramstedt M. Alcohol and fatal accidents in the UnitedStates—a time series analysis for 1950–2002. Accid AnalPrev 2008; 40: 1273–81.

24. World Health Organization. International Statistical Classifi-cation of Diseases and Related Health Problems, 10th revision.Geneva, Switzerland: World Health Organization; 2004.

25. Chamlin M., Myer A., Sanders B., Cochran J. Abortion ascrime control. Crim Justice Policy Rev 2008; 19: 135–52.

26. Bloomfield K., Rossow I., Norström T. Changes in alcohol-related harm after policy changes in Denmark. Eur AddictRes 2009; 15: 224–31.

27. Humphreys D. K., Eisner M. P., Wiebe D. J. Evaluating theimpact of flexible alcohol trading hours on violence:an interrupted time series analysis. PLoS ONE 2013; 8:e55581. doi: 10.1371/journal.pone.0055581.

28. Pridemore W. A., Snowden A. J. Reduction in suicide mor-tality following a new national alcohol policy in Slovenia:an interrupted time–series analysis. Am J Public Health2009; 99: 915–20.

29. McDowall D., McCleary R., Meidinger E., Hay R. Jr.Interrupted Time Series Analysis. Beverly Hills, CA: Sage;1980.

30. Granger C. Testing for causality: a personal viewpoint. JEcon Dyn Control 1980; 2: 329–52.

31. McCleary R., Hay R. Jr. Applied Time Series Analysis for theSocial Sciences. Beverly Hills, CA: Sage; 1980.

32. Chang K., Wu C., Ying Y. The effectiveness of alcoholcontrol policies on alcohol-related traffic fatalities in theUnited States. Accid Anal Prev 2011; 45: 406–15.

33. Gil A., Polikina O., Koroleva N., McKee M., Tomkins S.,Leon D. A. Availability and characteristics of nonbeveragealcohol sold in 17 Russian cities in 2007. Alcohol Clin ExpRes 2008; 33: 79–85.

34. Pridemore W. A., Chamlin M. B., Andreev A. Reduction inmale suicide mortality following the 2006 Russian alcoholpolicy: an interrupted time series analysis. Am J PublicHealth; in press; 2013; 103.

35. Pridemore W. A., Chamlin M. B., Kaylen M. T., Andreev A.The 2006 Russian alcohol policy and alcohol-related mor-tality: an interrupted time series analysis. Alcohol Clin ExpRes; in press; 2013; 37.

36. International Transport Forum. Road Safety Performance.National Peer Review: Russian Federation. Update 2010. Paris:International Transport Forum; 2010.



Documents

The impact of a national alcohol policy on deaths due to transport accidents in Russia