The burden of road traffic crashes, injuries and deaths in Africa: a systematic review and meta-analysis

Abstract Objective To estimate the burden of road traffic injuries and deaths for all road users and among different road user groups in Africa. Methods We searched MEDLINE, EMBASE, Global Health, Google Scholar, websites of African road safety agencies and organizations for registry- and population-based studies and reports on road traffic injury and death estimates in Africa, published between 1980 and 2015. Available data for all road users and by road user group were extracted and analysed. We conducted a random-effects meta-analysis and estimated pooled rates of road traffic injuries and deaths. Findings We identified 39 studies from 15 African countries. The estimated pooled rate for road traffic injury was 65.2 per 100 000 population (95% confidence interval, CI: 60.8–69.5) and the death rate was 16.6 per 100 000 population (95% CI: 15.2–18.0). Road traffic injury rates increased from 40.7 per 100 000 population in the 1990s to 92.9 per 100 000 population between 2010 and 2015, while death rates decreased from 19.9 per 100 000 population in the 1990s to 9.3 per 100 000 population between 2010 and 2015. The highest road traffic death rate was among motorized four-wheeler occupants at 5.9 per 100 000 population (95% CI: 4.4–7.4), closely followed by pedestrians at 3.4 per 100 000 population (95% CI: 2.5–4.2). Conclusion The burden of road traffic injury and death is high in Africa. Since registry-based reports underestimate the burden, a systematic collation of road traffic injury and death data is needed to determine the true burden.


Introduction
Road traffic injuries are among the leading causes of death and life-long disability globally. 1 The World Health Organization (WHO) reports that about 1.24 million people die annually on the world's roads, with 20-50 million sustaining non-fatal injuries. 1,2 Globally, road traffic injuries are reported as the leading cause of death among young people aged 15-29 years and are among the top three causes of mortality among people aged 15-44 years. 1 The Institute for Health Metrics and Evaluation (IHME) estimated about 907 900, 1.3 million and 1.4 million deaths from road traffic injuries in 1990, 2010 and 2013, respectively. 3 In Africa, the number of road traffic injuries and deaths have been increasing over the last three decades. 4 According to the 2015 Global status report on road safety, the WHO African Region had the highest rate of fatalities from road traffic injuries worldwide at 26.6 per 100 000 population for the year 2013. 1,2 In 2013, over 85% of all deaths and 90% of disability adjusted life years (DALYs) lost from road traffic injuries occurred in low-and middle-income countries, which have only 47% of the world's registered vehicles. 2,3 The increased burden from road traffic injuries and deaths is partly due to economic development, which has led to an increased number of vehicles on the road. 5,6 Given that air and rail transport are either expensive or unavailable in many African countries, the only widely available and affordable means of mobility in the region is road transport. 1,2,7 However, the road infrastructure has not improved to the same level to accommodate the increased number of commuters and ensure their safety and as such many people are exposed daily to an unsafe road environment. 1,4 The 2009 Global status report on road safety presented the first modelled regional estimate of a road traffic death rate, which was used to statistically address the underreporting of road traffic deaths by countries with an unreliable death registration system. 5 In the 2009 report, Africa had the highest modelled fatality rate at 32.2 per 100 000 population, in contrast to the reported fatality rate of 7.2 per 100 000 population. 5 The low reported death rate reflects the problem of missing data due to non-availability of road traffic data systems, which has a direct impact on health planning including prehospital and emergency care and other responses by government agencies.
This study aimed to review existing literature on published studies, registry-based reports and unpublished articles on the burden of road traffic injuries and deaths in the African continent to generate a continent-wide estimate of road traffic injuries and deaths for all road users and by road user type (pedestrians, motorized four-wheeler occupants, motorized two-three wheeler users and cyclists).

Methods
We searched MEDLINE, EMBASE, Global Health, Google Scholar, websites of road safety agencies and relevant organizations within Africa for articles published between 1980 and 2015 (Fig. 1). The search strategy and terms are presented in Box 1 (available at: http://www.who.int/bulletin/ volumes/94/7/15-163121). There was no language restriction.

Eligibility criteria
We included a study in the review if it met the following criteria: (i) conducted between 1980 and 2015 and that the Objective To estimate the burden of road traffic injuries and deaths for all road users and among different road user groups in Africa. Methods We searched MEDLINE, EMBASE, Global Health, Google Scholar, websites of African road safety agencies and organizations for registry-and population-based studies and reports on road traffic injury and death estimates in Africa, published between 1980 and 2015. Available data for all road users and by road user group were extracted and analysed. We conducted a random-effects meta-analysis and estimated pooled rates of road traffic injuries and deaths. Findings We identified 39 studies from 15 African countries. The estimated pooled rate for road traffic injury was 65.2 per 100 000 population (95% confidence interval, CI: 60.8-69.5) and the death rate was 16.6 per 100 000 population (95% CI: 15.2-18.0). Road traffic injury rates increased from 40.7 per 100 000 population in the 1990s to 92.9 per 100 000 population between 2010 and 2015, while death rates decreased from 19.9 per 100 000 population in the 1990s to 9.3 per 100 000 population between 2010 and 2015. The highest road traffic death rate was among motorized four-wheeler occupants at 5.9 per 100 000 population (95% CI: 4.4-7.4), closely followed by pedestrians at 3.4 per 100 000 population (95% CI: 2.5-4.2). Conclusion The burden of road traffic injury and death is high in Africa. Since registry-based reports underestimate the burden, a systematic collation of road traffic injury and death data is needed to determine the true burden. study was done in an African country; (ii) clearly referred to road traffic crashes, injuries or deaths; (iii) referred data came from a population-or registrybased data system; (iv) registry-based hospital data with the underlying cause of death data coded in the International Classification of Disease and Related Health Problems, 10th revision (ICD-10), with codes V01-V89; (v) directly attempted to estimate the number or rate of road traffic crashes, injuries or deaths in a particular African country or the region as a whole; or (vi) provided any other information (e.g. response time, first responders) that may further help to understand the burden and determinants of road traffic crashes and policy response in the African region.
We excluded studies if they: (i) referred to deaths by other means of transportation including water, air and other unspecified transport means; (ii) were mainly reviews, viewpoints and commentaries; (iii) did not have a clearly defined study design, data capture and analysis method; and (iv) had not clearly defined and consistently applied a case definition of a road traffic crash, injury or fatality.
For this study, a crash is defined as a road traffic collision that resulted in an injury or fatality. Injury refers to non-fatal cases from a road traffic crash. 5 Death is defined as a road traffic crash in which one or more persons involved in the crash died immediately or within 30 days of the crash. 2 For non-fatal injuries, the case definition ranges from minor injuries with disabilities of short duration, to severe injuries with lifelong disabilities.

Quality assessment
For each full text accessed, we checked if the study method had flaws in the design and execution. For the registrybased studies, we examined the study design, completeness, the appropriateness of statistical and analytical methods employed and if the limitations were explicitly stated. For each study, we assessed if the reported sample size or study population was appropriate to provide a representative estimate and if the heterogeneities within and between population groups undermine the pooled estimates. Studies not meeting this quality assessment were excluded.

Data collection process
Available data from all selected studies were extracted twice, compiled and stored in a spreadsheet. For each study, data on the country, study period, study design, sample size, mean age and case definitions were extracted (Table 1). Reported road traffic crash, injury and death data for the overall study population and for the various categories of road users were extracted. The data were grouped by study setting and year of study, with corresponding age and sex categories.

Data analysis
All extracted data on road traffic crashes, injuries and deaths were converted to rates per 100 000 population. Studies were subdivided into population-and registry-based studies and analysed separately for all road users and by road user category. A random effects metaanalysis was conducted on extracted road traffic crash, injury and death rates. To give a better understanding of the data distribution and comparisons with the pooled estimates and the confidence intervals, we further presented the range, median and data points within each data set. All statistical analyses were done in Excel 2010 (Microsoft, Redmond, United States of America) and Stata version 13.1 (StataCorp. LP, College Station, United States of America).

Reported rates
From all registry-based studies, Nigeria recorded the highest and lowest total crash rate at 716.57 per 100 000 population and 2.9 per 100 000 population, in 1990 and 2011, respectively. 39,42 Ethiopia recorded the highest death rate at 81.6 per 100 000 population in 2011, 11 while   43 From the available populationbased studies, Nigeria reported the highest number of road traffic injury and death rates at 4120 per 100 000 population and 160 per 100 000 population, respectively. The road traffic injury rate is the highest recorded in any single study in Africa. Algeria and Ghana also reported high road traffic injury rates at 700 and 938 per 100 000 population, respectively. 19,34 Only six studies reported male and female road traffic crash estimates, 14,19,21,22,29,31 with Algeria and South Africa recording the highest number of casualties. Table 2 presents the estimated pooled rates for the African continent. For total crashes, the pooled rate was 52.8 per 100 000 population, with the median at 39.7 per 100 000 population. The pooled fatal crash rate was estimated at 9.6 per 100 000 population with a median at 4.8 per 100 000 population. Pooled crash injury and death rates were estimated at 65.2 injuries and 16.6 deaths with medians of 38.9 injuries and 7.9 deaths per 100 000 population, respectively ( Fig. 2 and Fig. 3).

Pooled rates
From 1990 to 2015, road traffic injury rates increased from 40.7 to 92.9 per 100 000 population (Table 3). In contrast, death rates decreased from 19.9 to 9.27 per 100 000 population ( Fig. 4 and Fig. 5). Applying these figures and using the United Nations (UN) population estimates 47 for the region, the pooled estimate came to 106 000 road traffic deaths and 1.1 million injuries in 2015, compared with 126 000 deaths and 260 000 injuries in 1990.

By road user category
From individual studies, road traffic death rates among pedestrians ranged from 0.26 per 100 000 population in Nigeria in 2007 to 13 per 100 000 population in South Africa in 2003. 43,24 The death rate among motorized four-wheeler occupants was lowest in Nigeria in 2007 and highest in South Africa in 1999 at 0.74 and 63 per 100 000 population, respectively. 43 study from the United Republic of Tanzania reported the highest, at 0.12 and 3.12 per 100 000 population, respectively. 8,13 The pooled rates showed that motorized four-wheeler occupants had the highest road traffic death rate, closely followed by pedestrians. The pooled road traffic injury and death rates among pedestrians were 10.8 and 3.4 per 100 000 population, respectively. Among motorized four-wheeler occupants, the pooled road traffic injury and death rates were 37.2 and 5.9 per 100 000 population, respectively. Among motorized twothree wheeler occupants and cyclists, the pooled injury and death rates were 16.1 and 1.3 per 100 000 population, respectively ( Table 2).

Discussion
Our study reflects the difficulties that many experts have noted in describing the extent of road traffic crashes, injuries and deaths in Africa, for which modelling based on scarce and variable information, may not necessarily provide a reliable estimate. 48 Moreover, registrybased reports may grossly underestimate the burden of road traffic crashes. Population-based studies consistently report a higher fatality rate. 19,34,40 For example, a population-based survey conducted in Ghana in 1998 reported an injury rate of 940 per 100 000 population, 34 while another registry-based study in the same country for the same year estimated 32 per 100 000 population. 32 The Nigerian Federal Road Safety Corps estimated 3.7 deaths per 100 000 population for Nigeria in 2009. 39 In contrast, a popu-  Note: In the box plot, the boxes represent the interquartile range of road traffic injury rates where the middle 50% (25-75%) of data are distributed; the bars represent road traffic injury rates outside the middle 50% (< 25% or > 75%); the dots represent specific road traffic injury rates which were a lot higher than normally observed over the study period (outliers) and the lower, middle and upper horizontal lines represent the minimum, median and maximum road traffic injury rates (excluding outliers), respectively.
Road traffic crashes, injuries and deaths in Africa Davies Adeloye et al. Note: In the box plot, the boxes represent the interquartile range of road traffic injury rates where the middle 50% (25-75%) of data are distributed; the bars represent road traffic injury rates outside the middle 50% (< 25% or > 75%); the dots represent specific road traffic injury rates which were a lot higher than normally observed over the study period (outliers) and the lower, middle and upper horizontal lines represent the minimum, median and maximum road traffic injury rates (excluding outliers), respectively. Table 3.  Note: In the box plot, the boxes represent the interquartile range of road traffic injury rates where the middle 50% (25-75%) of data are distributed; the bars represent road traffic injury rates outside the middle 50% (< 25% or > 75%); the dots represent specific road traffic injury rates which were a lot higher than normally observed over the study period (outliers) and the lower, middle and upper horizontal lines represent the minimum, median and maximum road traffic injury rates (excluding outliers), respectively.

Systematic reviews
Road traffic crashes, injuries and deaths in Africa Davies Adeloye et al.
lation-based study in the same country reported a higher estimate of 160 deaths per 100 000 population. 40 The subgroup analysis showed that injury rates increased and death rates decreased between 1990 and 2015. A high road traffic injury number may reflect the effect of economic growth on the burden of road traffic injury in the region, which may be associated with increased travel and exposure to a hazardous traffic environment. 49,50 However, death figures may be decreasing due to a relatively improving prehospital and emergency response system, 51 as noted in Ghana, South Africa and Zambia. 52,53,54 It is important to note that many deaths may be missed or not recorded, as many of the road safety agencies tend to only record crashes, leaving the recording of deaths to health agencies. 55,56 Our findings further revealed that the highest rates of casualties are among motorized four-wheeler occupants and pedestrians. A WHO report shows that 43% and 38% of road traffic deaths in the African Region occurred among motorized four-wheeler occupants and pedestrians, respectively. 2 In Africa, most of these motorized four-wheeler occupants are passengers of commercial vehicles which is the commonly used means of transport. 4 The high death rate among motorized four-wheeler occupants may also be due to the fact that crashes involving motorized four-wheeled vehicles are often recorded, while pedestrian crashes may be missed. 4,35 However, we agree with some authors who have reported that pedestrians may be more affected in Africa due to bad road infrastructure, lack of pedestrian-friendly road signs, the way traffic is mixed with other road users and a general disregard for pedestrians by drivers. 1 Meanwhile, a major challenge for the response to road crashes in Africa is the lack of reliable information and data that can inform an evidence-based public health response. 49,57 Underreporting especially of vulnerable road users, poor or absent links between reporting agencies, exemptions from reporting, poor sampling techniques and varying case definitions have been indicated as limitations of reported data. The different rates of road traffic crash, injury and death reported in this study may be mostly related to surveillance system reporting errors and biases. In many African countries, there are no effective vital registration and active surveillance systems to capture the outcome of a road traffic crash 1 and police data is the main source of traffic crash data. 1,2 However, data from police sources tend to underreport injuries and deaths due to poor traffic police response and follow up on injured victims and varying traffic fatality definitions for real-time and chronologic data capture. 4 Our study has the following limitations. Population-based studies on road traffic crashes in Africa, which would have been more reliable than registry-based studies, were not available. Population-based studies may have given insights on the extent of road users' exposure to traffic risk, mode and frequency of road travel, distance travelled, number of road commuters and the conditions of the road. In the absence of such information, we have not based our estimates on an appropriate travel exposure denominator, thus limiting an understanding of the reasons behind the reported road traffic crash, injury and death rates and trends.
The available registry-based studies varied in their quality. They reported questionable values and trends and provided uncertain estimates. Lack of appropriate case definition for road traffic fatalities and incomplete breakdown of road traffic crash estimates by road user type were major limitations. Additionally, the non-fatal injury figures reported by the different studies varied with respect to severity and outcome. These variations could have affected our meta-analyses.
While we applied the UN population data for Africa to estimate rates where relevant national reference population data were unavailable, there were no comparable data to use for subnational studies. In addition, the data employed for this analysis were generated only from 15 countries, which is relatively small to accurately reflect the overall situation in the region. Hence, our estimates should be interpreted against these limitations.
In conclusion, our study suggests that the burden of road traffic injuries in Africa is high and there is an underestimation of road traffic fatalities. Improved road traffic injury surveillance across African countries may be useful in identifying relevant data gaps and developing contextually feasible prevention strategies in these settings. ■ Competing interests: None declared. Year range 1980-1989 1990-1999 2000-2009 2010-2015 Note: In the box plot, the boxes represent the interquartile range of road traffic injury rates where the middle 50% (25-75%) of data are distributed; the bars represent road traffic injury rates outside the middle 50% (< 25% or > 75%); the dots represent specific road traffic injury rates which were a lot higher than normally observed over the study period (outliers) and the lower, middle and upper horizontal lines represent the minimum, median and maximum road traffic injury rates (excluding outliers), respectively.

Systematic reviews
Road traffic crashes, injuries and deaths in Africa Davies Adeloye et al.