Schaffer Online Library of Drug Policy Sign the Resolution
Contents | Feedback | Search
DRCNet Home
| Join DRCNet
DRCNet Library | Schaffer Library | Drugs and Driving 


Alcohol-Related Traffic Injuries and Fatalities in Developing Countries: A Critical Review of Literature

WO Odero, AB Zwi

Health Policy Unit, Department of Public Health and Policy, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK


A literature review of published studies on alcohol and traffic injuries in developing countries was undertaken to examine evidence of the prevalence of alcohol. 16 studies were identified through electronic database searches from 1966 to 1994. The studies employed different measurement methods and cut-off levels of blood alcohol concentrations (BACs). 8 fatality studies reported varied BACs in drivers ranging from 33.3% to 63.2%, measured by blood analysis. In four of the studies, alcohol prevalence, tested in less than 50% of the study population, varied from 17.3% to 46%. No clear selection criteria were stated, and the representativeness of those tested could not be ascertained. In eight non-fatality studies, the proportion of intoxicated subjects, determined by blood analysis, breath tests and interviews, were considerably lower and varied widely, from 7.7% to 28.4%. Alcohol prevalence was consistently higher amongst drivers (33.3% - 69.2%) than in other road users, and over 95% of intoxicated drivers were male (95%-100%). 50% of alcohol positive subjects were aged between 20 and 30 years.

From this review, evidence of the influence of alcohol in traffic injuries in developing countries is limited. Due to variable measurements and threshold BAC levels applied, direct comparison of results is inappropriate. The true prevalence of alcohol-related traffic injuries remains unknown. There is need for a standardized methodology, reliable BAC measuring devices and a uniform cut-off level.


The contribution of alcohol to increased risk of traffic accidents has been well established and has received international attention for many years. Experimental studies and epidemiological surveys, undertaken in a number of industrialized countries since the early part of this century, have documented consistent and convincing evidence of the direct dose-response relationship between increasing blood alcohol concentrations (BACs) in a motor vehicle driver and the increasing risk of his/her involvement in a road accident. Over the last decade, and on realization of the increasing importance of road traffic accidents (RTAs) as a cause of morbidity, mortality, disability and economic loss; international attention has been focused in developing policies and strategies for the prevention of injuries and fatalities resulting from alcohol-impaired driving. Most industrialized nations have consequently introduced new legislations, intensified enforcement of drinking and driving laws, and also increased research in alcohol-related traffic crashes. However, due to the diversity in demographic structure, socio-cultural characteristics, levels of economic development, as well as road transport infrastructure and traffic mix, the direct transfer of countermeasures successfully applied in the Western world to developing countries may not be appropriate or feasible.

While studies conducted by the UK Transport Research Laboratories (TRL), the World Health Organization, the World Bank and a number of researchers (Downing et al, 1991; WHO 1984, 1989; World Bank 1993; Mohan and Romer, 1991; Feachem et al, 1992; Zwi, 1993) have revealed the increasing importance of RTAs as a major cause of death and a significant public health problem in developing countries, to the magnitude even higher than in industrialized countries, little reliable information is available on the extent of the contribution of the various risk factors. For instance, the relationship between traffic injuries and alcohol consumption has scarcely been examined objectively in Third World countries. Through application of research experiences of industrialized countries, developing countries should endeavour to examine and prioritize the extent of influence of the various country-specific RTA risk factors, and to generate data appropriate for developing as well as monitoring targeted traffic injury prevention interventions.


The aim of this paper is:

  1. To examine evidence of alcohol-related traffic injuries and fatalities in developing countries
  2. To identify methodological problems associated with blood alcohol prevalence studies


Studies were located through systematic identification of published literature on RTAs in developing countries, spanning from 1966 to June 1994, by using the following methods:

  1. CD-ROM MEDLINE search for the period 1966 to June 1994. Developing countries were identified using regional terms - Asia, Africa, Middle-East, South America, Central America, Mexico, West Indies and Pacific Islands. These were then automatically 'exploded' into constituent countries. Articles were identified using key search terms - "accidents or injuries or trauma or alcohol". Titles and abstracts of the identified articles were then scanned, and those with data on alcohol were selected.
  2. EMBASE and SCI-SEARCH databases were searched from 1980 to June 1994 by applying similar string search terms and article selection procedures as for MEDLINE.
  3. LILAC data base for the Latin American and Caribbean Health Science Literature was searched from 1982 to June 1994. By systematically applying consistent key search terms and selection procedures, relevant articles were identified. Help with translation of non-English publications was sought.
  4. By 'snowballing', relevant references cited in the papers selected through the above methods (1-3) were located.
  5. Proceedings of conferences on RTAs were obtained from institutions, organizations and experts carrying out research on traffic accidents in developing countries.


Sixteen studies were identified (Table 1). The subjects were non-fatal motor vehicle casualties in eight studies, while the other eight studies examined fatal RTA victims. The sample size of those tested ranged from 17.4% (1 study) to 100% (4 studies) of the total subjects enroled, with a mean proportion of 74.5%. Methods of assessment of blood alcohol concentrations varied, and included blood analysis (11 studies), breath tests (2 studies) and interviews (3 studies). Cut-off levels of BACs applied in the different studies varied. In nine studies, a threshold BAC of 80 mg% was applied, it was 20 mg% in one study , while in two studies the level was 20 mg%. The overall alcohol prevalence determined ranged from 10% in Singapore (Wong, 1990) to 69.2% in Puerto Rico (Kaye, 1970). However, eight studies that examined BACs by class of road-user further showed considerable variations, with drivers and pedestrians comprising the highest proportions of those intoxicated (Table 2).

Table 1
Studies on Alcohol and Road Traffic Injury and Fatality in Developing Countries, 1966 - 1994

Author & Year Country/ Region Study period/ Year Subjects Sample size Detection method BAC cut-off level Alcohol prevalence
Total No. tested (%)
Aguwa, 1982 Nigeria, Anambra State 1 yr, 1979-80 non-fatal, drivers 32 32 (100) blood analysis 10 mg% 56%
Asogwa, 1980 Nigeria 17 mo. 1975-77 non-fatal, drivers 1296 555 (43) interview alcohol use 7.7%
CAREC, 1979 Suriname 1 mo. Feb. 1987 non-fatal, all 289 252 (87) breath test 50 mg% 9.0%
Fosseus, 1983 S. Africa, Cape Town 1 yr, 1982 fatal, motorcyclists 48 40 (83) blood analysis 10 mg% 62%
Gururaj, 1993 India, Bangalore 6 mo. 1991-92 non-fatal, head injury 1784 not stated self reporting alcohol use 16%
Kaye, 1971 Puerto Rico 1 yr, 1970 fatal, all 451 338 (75) blood analysis 80 mg% 58.9%
Kaye, 1973 Puerto Rico 1 yr, 1972 fatal, all 552 404 (73) blood analysis 80 mg% 63.2%
Kaye, 1974 Puerto Rico 1 yr, 1973 fatal, all 577 489 (85) blood analysis 80 mg% 46%
Myers, 1972 S. Africa 1 mo. Apr-May 1976 non-fatal, all 165 115 (70) blood analysis 80 mg% 35.6%
Patel, 1977 Zambia, Lusaka 1 yr, 1974-75 fatal, all 217 217 (100) blood analysis 80 mg% 26.7%
Sinha, 1981 Papua NG, Port Morseby 5 yrs, 1976-80 fatal, all 171 39 (23) blood analysis 80 mg% 53% -drivers 90% -pedest
Sinha, 1989 Papua NG, Port Morseby 10 yrs, 1976-85 fatal, all 363 63 (17) blood analysis 80 mg% 48% -drivers 66% -pedest
Van de Spuy, 1989 S.Africa, Pretoria not stated non-fatal, drivers 275 275 (100) blood analysis 80 mg% 26.7%
Wong, 1990 Singapore 15 mo. 1986-87 non-fatal, motorcyclists 198 198 (100) interview alcohol use 10%
Wu, 1991 Taiwan, Taipei 10 wks, 1987-88 non-fatal, all 489 449 (92) breath test, blood analysis 50 mg% 8.5%
Wyatt, 1980 Papua NG, Port Morseby 5 yrs, 1976-80 fatal, all 121 85 (70) blood analysis 80 mg% 33.3% -drivers

Table 2
Alcohol Prevalence, in Percentages, by Class of Road User

Author, Country & Year Drivers Pedestrians Passengers Motorcyclists Bicyclists
CAREC, Surinam, 1987 15.4 (26)* 18.2 (44) 13.6 (44) 24.1 (24) 12.5 (24)
Kaye, Puerto Rico, 1971 63.0 (57) 61.0 (106) 26.0 (37) - -
Kaye, Puerto Rico, 1973 69.2 (65) 64.8 (105) 48.8 (21) - -
Kaye, Puerto Rico, 1974 48.0 (-) 34.0 (-) - - -
Myers, S. Africa, 1977 47.0 (34) 28.6 (21) 37.5 (24) 27.7 (36) -
Patel, Zambia, 1977 30.3 (33) 32.1 (109) 14.7 (61) - 33.3 (12)
Sinha, Papua NG, 1981 52.6 (19) 90.0 (20) - - -
Sinha, Papua NG, 1989 48.0 (-) 66.0 (-) - - -
Wong, Singapore, 1990 - - - 10.0 (198) -
WU, Taiwan, 1991 3.9 (23) 13.8 (65) 10.8 (65) 24.0 (242) -
Wyatt, Papua NG, 1980 33.3 (15) 69.0** (-) - - -
* numbers in parenthesis ( ) - these are denominator figures for each category
** this figure is for male pedestrians only

Among intoxicated subjects, males were predominant, accounting for between 87.5% and 100% of those testing positive, though this analysis was limited to only two studies (Kaye, 1971, 1973).

Five studies that presented traffic casualty data by day of the week (Table 3), showed that both the numbers of casualties and the proportions with alcohol were greater during weekends - from Friday to Sunday, compared to those occurring during weekdays. The proportions of nighttime traffic deaths, reported in three studies (Fosseus, 1983; Kaye, 1971 & 1972), were also higher compared to those resulting from daytime crashes. However, no figures indicating diurnal distribution of alcohol prevalence in the victims were reported.

Table 3
Percentage Distribution of RTA Casualties and Alcohol Prevalence by Day of the Week

Author, Country & Year Category of subjects Mon Tue Wed Thu Fri Sat Sun
CAREC, Suriname, 1987 All casualties 14 16 13 15 17 15 9
Alc. present 32 23 17 12 24 34 33
Fosseus, S. Africa, 1983 All casualties 9.5 4.7 4.7 4.7 14.3 23.8 28.6
Alc. present -* - - - - - -
Hayes, Puerto Rico, 1971 All casualties - - - - - - -
Alc. present - - - - 13.9 - 52.5**
Hayes, Puerto Rico, 1973 All casualties - - - - - - -
Alc. present 8.5 8.5 7.2 7.2 13 25 18
Myers, S. Africa, 1977 All casualties 10.4 3.5 10.4 9.6 29.6 25.2 11.3
Alc. present 2.4 - - 9.8 36.6 41.5 9.8
* blank place - no data reported
** this value is for both Saturday and Sunday


The main approaches to dealing with the problem of drinking drivers include detection of those intoxicated, institution of deterrent penalties, and education of the public about the use of alcohol and the risks involved in combining drinking and driving. Reliable and accurate data, of the consequences of drinking on the driver and other road users, is therefore necessary. However, collecting information about alcohol in traffic accident victims, as evidenced in this paper, presents a number difficulties and biases which can be summarized under two broad categories:

Measurement Problems

Differences in techniques of measurement of blood alcohol, the variable reliability and validity of testing methods, and the application of different cut-off levels of BACs affect the comparability of data sets from different studies. In addition, lack of defined time interval between accident and collection of blood samples for analysis or breath testing or making subjective assessment render the determined BAC values inappropriate, since blood alcohol concentration diminishes as alcohol is progressively eliminated from the blood with time after ingestion.

Subject Selection

A number of selection factors may affect the determination of the true prevalence of alcohol involvement. Not every RTA victim is tested for alcohol, and more-over, the criteria for selection is often not clearly stated. For instance, due to a number of reasons, such as refusals and severity of injury, completeness of information about alcohol involvement in non-fatal subjects is less likely than in fatal victims. Men are also more likely to be selected for blood alcohol determination at autopsy than women. While the age bracket of those tested may not be specified, comparatively large numbers of adolescents and young adults are more likely to be included, and therefore appear to comprise the majority of those intoxicated. In accident data derived from police records, those perceived to be responsible for the accident are more likely to be evaluated. Drivers involved in fatal RTAs, for example, are more likely to have blood alcohol determined than passengers, pedestrians or cyclists.

Without a clear statement of inclusion and exclusion criteria, the representativeness of the subjects tested cannot be ascertained, consequently the BAC results may not be generalizable to all casualties involved.

In view of these inconsistencies, caution is needed while interpreting blood alcohol data from different studies. Where similar detection methods, case definitions and selection criteria are consistent, within country as well as international comparisons can be made. For example, in fatality studies reported from Zambia (Patel et al, 1977), Puerto Rico (Kaye, 1971,73 & 74) and Papua New Guinea (Wyatt, 1980; Sinha et al, 1981 & 1989), the BACs are comparable in terms of measurement methods and cut-off points (80 mg%), and may be used for monitoring trends and the effectiveness of drink-driving countermeasures in the respective countries over a time period. The variations of alcohol prevalence reported may however reflect actual differences in alcohol use in the different countries. These results also imply that alcohol is likely to be an important contributing factor in road accident fatalities.

Of the eight studies that reported alcohol prevalence in non-fatal accident victims, three (Asogwa, 1980; Aguwa, 1982; de Spuy, 1989) examined only drivers, and the results, as expected, widely varied - 7.7%, 56%, and 28.4% respectively, taking into account the methodological differences. The rest of non-fatality studies combined all RTA casualties in the analysis which further compounded the suitability for comparability of results.

It is therefore clear that direct evidence of the extent of alcohol involvement in traffic injuries and fatalities in developing countries remains inconclusive. Perhaps this is one of the reasons for the relative lack of explicitly designed policies and strategies for the prevention of alcohol-related traffic accidents in the majority of developing countries.


  1. Adoption and application of practical, valid and reliable alcohol measurement techniques and devices that are affordable for use in developing countries.
  2. Standardization of a case definition of alcohol involvement, and criteria for inclusion and exclusion of subjects.
  3. Development and introduction of a reliable accident data recording system that would provide more complete information on road traffic casualties, including objective assessment of alcohol involvement.
  4. Design and execution of country-specific research examining the relative contribution of alcohol in both traffic and non-traffic injuries, and the establishment of relevant comprehensive countermeasures.
  5. Design and execution of cost-effectiveness studies to help decide on which drink-driving policies are most appropriate and affordable in developing countries.
  6. The need for policy analysis to establish which policy options have a greater chance of successful implementation.


Aguwa C.N., Anosike E.O. and Akubue P.I. (1982) Road Accidents in Nigeria: Level of alcohol in the blood of automobile drivers. Central African Journal of Medicine. 28, 171-174.

Asogwa S.E. (1980) Some characteristics of drivers and riders in road accidents in Nigeria. East African Medical Journal. 57 (6), 399-404.

CAREC Surveillance Reports. (1987) Blood alcohol levels in traffic accident victims- Surinam February 1987. CAREC Surveillance Reports. 13, 1-4.

Downing A.J., Bagunley C.J. and Hills B.L. (1991) Road safety in developing countries: an overview. Transport and Road Research Laboratory, Crowthorne, Berkshire. U.K.

Feachem R.G.A., Kjellstrom T., Murray C.J.L., Over M. and Phillips M.A. (1992) The Health of Adults in the Developing World. Oxford University Press.

Fosseus C.G. (1983). Motorcycle accidents and alcohol. A survey of fatalities in the Cape Peninsula. South Africa Medical Journal 64 (5), 159-160.

Gururaj G., Channabasavanna S.M., Das B.S., Kaliaperumal (1993). Road accidents and head injury in Bangalore, need for prevention. Proceedings of the Conference on Asian Road Safety. Kuala Lumpur, Malaysia, October 24-28 1993.

Kaye S. (1971) The drunk pedestrian and driver on our highways in Puerto Rico (1970). Bulletin of Association of Medicine of Puerto Rico. 63 (7), 170-7.

Kaye S. (1974) Sudden drop in alcohol and drug-related traffic accidents in Puerto Rico - 1974. Bulletin of Association of Medicine of Puerto Rico. 67 (12), 369-71.

Kaye S. (1973) Influence of alcohol on traffic deaths in Puerto Rico, 1972. Bulletin of the Association of Medicine of Puerto Rico. 65 (6), 135-9.

Mohan D. and Romer C.J (1991). Accident mortality and morbidity in developing countries. In: Accidents in childhood and adolescence. The role of research. Manciaux M. and Romer C.J (Editors). World Health Organization, Geneva, pp31-38.

Myers R.A., Taljaard J.J., Penman K.M. (1977) Alcohol and road traffic injury. South Africa Medical Journal. 52 (8), 328-330.

Patel N.S. and Bhagwatt P.P. (1977) Road Traffic Accidents in Lusaka and blood alcohol. Medical Journal of Zambia. 11, 46-69.

Sinha S.N., Sengupta S.K. and Purohit R.C. (1981) A Five-Year Review of Deaths following trauma. Papua New Guinea Medical Journal. 24, 222-228.

Sinha S.N. and Sengupta S.K. (1989) Road Traffic Accident Fatalities in Port Morseby: A 10-year survey. Accident Analysis and Prevention. 21, 297-301.

Van der Spuy J.W. (1989) More about trauma. Continuing Medical Education. VMO.7; 6, 721-726.

Wong T.W., Phoon W.O., Lee J., Yin I.P., Fung K.P. and Smith G. (1990) Motorcyclist traffic accidents and risk factors: A Singapore Study. Asia and Pacific Journal of Public Health. 4 (1), 34-38.

World Bank. (1993) Investing In Health. World Development Report 1993. Oxford University Press.

World Health Organization. (1984) Road Traffic Accidents in Developing Countries. World Health Organization, Technical Report Series. 703.

World Health Organization. (1989) Analysis of achievements of traffic safety in Industrialized and Developing Countries. World Health Organization, Technical Report Series. 781.

Wu S.I., Yang., Chou., Tao C.C. and Chen K.T. (1991) An analysis of traffic injuries in Taiwan in relation to alcohol use and economic loss. Injury. 22 (5), 357-361.

Wyatt G.B. (1980) Epidemiology of road accidents in Papua New Guinea. Papua New Guinea Medical Journal. 23, 60-65.

Zwi A. (1993) The public health burden of injury in developing countries: A critical review of literature. Tropical Diseases Bulletin. 90 (4), R1-R45.