Is driver education contributing towards road safety? a systematic review of systematic reviews

Abstract: Background: There is a vast amount of literature on the effects of driver education. However, the evidence has become somewhat fragmented, making it challenging to understand driver education's effec-tiveness for improving road safety. The current study aimed to provide the efficacy of pre-LDE and post-LDE interventions aimed at improving the safety of drivers (includes crashes, injuries, or secondary outcomes). Methods: The following online databases were searched up to the 21st of February 2020: Web of Science (WOS), Scopus, PubMed, Cochrane library, and other relevant databases. Systematic reviews (SR) and meta-analyses were selected to investigate the efficacy of driving education in reducing crashes, injuries, or secondary outcomes. Two investigators independently conducted the data extraction and used the assessment of multiple systematic reviews (AMSTAR) tool to conduct a quality assessment of each SR identified. Results: Out of the 229 potential articles, seven SRs were eligible for the current overview of systematic reviews and meta-analyses. This overview showed that pre-and post-license education by people of all ages led to improvements in secondary outcomes, such as performance, self-perceived driving abilities, behind-the-wheel driving performance, and even a small decrease in traffic offenses. However, education was not effective in reducing crashes or injuries, either at the individual or community level. Conclusions: There was no evidence that driver education is an effective approach to reducing crashes or injuries. This negative result might be due to ineffective teaching methods. To improve road safety, it appears necessary to change the method or content of driving education since the current approaches to driving education do not reduce traffic crashes or injuries.

engineering, enforcement, education, evaluation, the speed of emergency responses, as well as improving vehicles and roads. [2][3] Due to the practical and feasibility of training or education, among these central pillars, this is a common approach for improving safety, especially in developing countries that have limited resources. 4 The training and education of drivers can be placed into two main categories, which are pre-license education and post-license driver education (pre-LDE and post-LDE). Pre-LDE comes in two common forms, school-based road safety education and one-to-one driving instruction with professional driving instructors. 5 The procedural and cognitive skills training tend to be combined in pre-LDE but are treated separately in post-LDE. 5 Previous research has found school-based driver education (pre-LDE) has been found to lead to earlier licensing but not to reduce crash involvement among teenagers. 6 Furthermore, pre-LDE using driving simulators was not found to change the driving style significantly. 7 Research has found moderately improved knowledge, driving awareness, and driving performance in a number of systematic reviews (SRs) on post-LDE among older people, but no reduction in crashes and injuries. [8][9] In addition, there is no evidence that education for improving driver behavior and post-LDE has significantly reduced road injuries or crashes, although a small decrease in the occurrence of traffic offenses was reported. 10 In contrast, there is some evidence that education has a somewhat positive impact on outcomes such as crash rates. However, these studies would be considered low on Sackett's levels of evidence. 11 In a study, 12 cognitive training for drivers was found to significantly reduce motor vehicle collision involvement per year and per mile driven. In a group-based education program have been found a reduction in collisions among old-aged drivers by following exposure to a 1 × 90 min session every two years. 13 Furthermore, a systematic review reported that older driver training, with programs tailored to individual participants, improved their self-perceived driving ability, their behind-thewheel performance, and reduced crashes. 14 Therefore, the contribution of driver education towards road safety remains somewhat controversial.
Although there are several systematic reviews (SRs), there are currently no overviews of the systematic reviews and meta-analyses on the efficacy of pre-LDE and post-LDE interventions aimed at improving the safety of drivers (includes crashes, injuries, or secondary outcomes). The current study was conducted to provide a comprehensive overview of pre-LDE and post-LDE efficacy by summarizing the findings from published SRs using data from different countries. And it will answer the question of whether driver training contributes to road safety.

Search Strategy and Inclusion Criteria
The methodology of this study is based on the preferred reporting items for systematic reviews and meta-analysis (PRISMA) checklist. We searched online databases, including the Web of Science (WOS), Scopus, PubMed, Cochrane library, and other relevant databases. Searches were performed up to the 21 st of February 2020. This study is a systematic review of systematic reviews (SRs). In order to identify SRs that assessed the impact of driver education on crashes, injuries, or secondary outcomes, the following search terms were used: (driver* OR driving) AND (driver training* OR pre-driver training* OR learning to drive training* OR post-licensure training* OR driver improvement * OR novice drivers) AND (young* OR youth* OR teen* OR adolescent* OR novice* OR older* OR student) AND (systematic review* OR metaanalysis*). All identified studies were included in the software StArt (State of the Art through Systematic Review). 15 This software used different combinations of the terms "systematic" and "review" in the title or abstract to identify the other studies' systematic reviews. These studies were also screened separately by two other individuals in addition to the software.
SR studies were included that met the following criteria: studies were systematic reviews with or without a meta-analysis that covered the effect of driver education [included pre-license driver education (professional driving instruction, school-based driver education, and simulator training) and post-license driver education (novice drivers education, remedial driver education, advanced driver education, and driver improvement)] on road traffic outcomes [(crash and injury) and secondary outcomes (driving performance or driving awareness or driving behavior and knowledge)].

Data Extraction and Quality Assessment
Two investigators independently conducted the data extraction and quality assessment of all SRs include in this study. Discrepancies were resolved by consensus between the two or by discussions with a third researcher (K-BL).
The following data were extracted from all SRs included: first author's name, publication year, study country, number of included primary studies, type of included primary studies, the stage of the driver, education type, age group, whether they studied crashes, injuries or secondary outcomes, and their main findings. The assessment of multiple systematic reviews (AMSTAR) tool was used to assess all systematic reviews' methodological quality. 16 This checklist assessed the quality of the SRs according to the following items: 'a priori' design provided; duplicate study selection/data extraction; comprehensive literature search; status of publication (i.e., grey or unpublished literature); list of studies included/excluded provided; characteristics of included studies documented; scientific quality assessed and documented; appropriate formulation of conclusions (based on methodological rigor and scientific quality of the studies); appropriate methods of combining studies (homogeneity test, effect model used, and sensitivity analysis); assessment of publication bias (graphic and/or statistical test); and conflict of interest statement.

Characteristics of Included Studies
Our search strategy identified 229 articles, of which 49 were duplicates and were therefore removed. After title and abstract screening, 43 records were found to be eligible for full-text assessment (Appendix 1(, but only seven were found to meet all of these overview criteria. The flowchart of the study identification and selection process is presented in Figure 1.
Five of the included studies were systematic reviews, [7][8][9]14,17 with two of them also including a metaanalysis. 6,10 There were 83 primary driver education interventions included in this current overview, five of which were conducted on post-license education [8][9][10]14,17 and two on pre-license education. 6,7 The publication years of the SRs included ranged from 1999 to 2019. Table 1 presents the detailed characteristics of the included SRs.

Pre-license Driver Education (Pre-LDE)
Pre-LDEs are defined as any educational curriculum designed to encourage the development of drivingrelated behaviors and beliefs that are specifically targeted at those who have not yet received their provisional driver's license. 5 Pre-LDE includes many different formats, including professional driving instruction, school-based driver education, and simulator training. 5  Two randomized controlled trials,9,10 two reanalyses of the data from one of the controlled trials,11,12 three ecologic timeseries designs, [13][14][15] and two ecologic analyses of group differences A systematic review of three randomized controlled trials (RCTs), conducted by Roberts and Kwan, 6 found that road safety education in schools contributes to earlier licensing of teenage attendees. Furthermore, they also suggested that these interventions may potentially lead to a significant rise in the proportion of crashes among adolescents who had been through this course, via earlier licensing.
A systematic review conducted by Martín-delosReyes et al. 7 investigated whether the use of driver simulator training in young learner drivers or inexperienced drivers leads to a reduction in road crash or injury and/or traffic offenses, or to the development of safer driving skills when compared with non-simulatorbased training. The results of this systematic review did not support the efficacy of simulator-based training programs for improving the safety of driving styles among young learner or novice drivers.

Post-license Driver Education (Post-LDE)
Post-LDEs are focused on inexperienced drivers (mostly young) with the aim of reducing their greater risk of being crash involved. 5 Although pre-licensing driver training attempts to incorporate cognitive and procedural skills training within a single curriculum, post-licensing driver training programs attempt to address either procedural or cognitive skills. 5 Vernick et al. 17 performed a systematic review of Post-LDEs to investigate whether high school-age individuals participating in driver education had fewer motor vehicles accidents or violations, or were much more likely to obtain a driver's license, compared to those not participating in these programs. They also investigated whether the presence of high school driver training programs was linked to lower rates of motor vehicle crashes among inexperienced drivers. The results of this review of nine studies found that there was no compelling evidence to suggest that high school driver education decreased the number of motor vehicle accidents among teenage drivers, either at the individual or community level. In contrast, there was evidence that these programs were linked with higher rates of involvement in young drivers' crash or injury, as they led to early licensure.
Several systematic reviews among older drivers have indicated that driver training has little or no direct impact on road safety, in terms of crash or injury risk reduction. 8,9,14 However, there is evidence that these programs can have a beneficial effect on several secondary outcomes (e.g. enhancing driving performance, or driving actions and knowledge or driving awareness). 8,9,14 In another systematic review of post-license driver education, 10 which primarily looking at remedial educational programs, the researchers again concluded that there was no clear evidence that these programs reduce crash or injury, and only very poor evidence that they minimize re-offending.

Quality Assessment
The final scoring of the methodological quality of each SR is presented in Table 2. Based on the AMSTAR methodology, the overall score for six SRs was assessed to be high and one was found to be of medium quality.

Discussion
To the best of our knowledge, this current systematic review of systematic reviews is the first study to investigate the efficacy of pre-LDE and post-LDE aimed at improving the safety of drivers. Based on 7 SRs (5 SRs and 2 Meta). The present study found that pre-and post-license driver training improved driving performance, self-perceived driving ability, and may also result in a small decrease in traffic offenses. However, these educational interventions were not effective in reducing crashes and injuries, either at the individual or community level. Driver education is a popular approach for improving road safety, 44 but our study indicates that driver education has not been effective in achieving its main outcome, which is reducing crashes, injuries, and deaths. This negative finding may be due to the use of ineffective teaching methods, ineffective course content, failure to understand the needs of adult learners, or not targeting the correct risky driving behaviors.
Educational interventions may not always have the intended effect on all individuals, which needs to be taken into consideration. It has been found that raising awareness of risk may reduce the subject's attractiveness for some learners, while for others it may be more attractive and more likely to lead to action. 18 Furthermore, the systematic review by Roberts et al. 6 reported that school-based driver education interventions can even lead to more harm, by reducing the age at which they obtain a license and thus leading to an increase in the proportion of teenagers involved in traffic crashes.
Some researchers have argued that road safety education interventions frequently focus on vehicle control skills, which are not often the most important issue when it comes to crash involvement. 18,19 Perhaps road safety educations focuses too strongly on a small number of driving behaviors, such as speed and neglect other important behavioral aspects, such as smoking, alcohol, and drug use. In such a case, training recipients, even if they show improvements in their attitudes and driving speed, the negative impact of other behaviors, such as alcohol consumption will continue to increase the number of crashes and their related injuries. 18 In fact, some educational interventions satisfy a number of goals but do not necessarily target all of the important causes, which reduces the efficiency of such interventions. 19 Furthermore, most educational interventions only provide general information about safety, without providing drivers a clear understanding of their own behavior as a road user. Many education interventions focus on improving knowledge and attitudes, rather than an individual's actual behavior, which reduces these interventions' efficiency. 19 In other words, drivers may have an awareness of the risks and the appropriate knowledge, but still engage in unsafe behavior even after education program. 18 The ineffectiveness of education programs may be caused an implementation without appropriate scientific evidence or being based upon educational theory. Some researchers believe that designing an educational intervention without regard to guiding theory is like prescribing a medical drug to a patient regardless of the patient's physiology. 18,20 Nasvadi and Vavrik suggested that before making decisions regarding driving education, it is important to carefully assess the potential benefits and contents in order to optimize the efficacy of the interventions. 21 Furthermore, another problem with driver education-based interventions is that designers do not include appropriate behavior change techniques. [22][23] In the case of safe traffic behavior, it has been argued that the resources allocated to educational programs may fail to improve traffic safety since traffic collisions and their related injuries do not decrease and drain already stretched budgets and hinder the development of more creative and innovative solutions. 24 The ineffectiveness of training and education programs have also been reported in other areas, prompting the British Medical Association to conclude that many educational interventions are politically attractive, but have been found to be ineffective. 25 The implementation method of educational programs can also influence its effectiveness. For example, it has been shown that educational activities are often implemented in short time periods. 26 Short-term education includes the length of the session, the number of sessions, and the interval between sessions. The use of short-term interventions may reduce the long-term effectiveness of education or training for many individuals. 18 Continuous driver education should be conducted in an evidence-based holistic framework, using road safety interventions such as the enforcement of traffic safety and risk reduction by decreasing and controlling early exposure. [27][28] Our conclusion should be interpreted with caution on the effects of education on road safety, because first we should be ensured that the evaluation of interventions across primarily included studies has conducted with a scientific method, appropriately. 17 For example, when the driver education cannot reduce the crash or death rate, this failure may be manifested in the inability of young drivers' function, inexperience, or interaction of other characteristics, instead of reflecting inabil-ity education interventions. 21 Other study by Zhang et al., 29 reported that driving educational interventions were effective only among careful drivers. And also, the period between getting safe behavior and reducing road crash and injury in the community is a very long and time-consuming phenomenon. So measuring the rate of road traffic injuries and education's effect seems not to be a rational manner. Instead, the effect of education interventions should be measured on safe traffic related behaviors. On the other hand, when educational interventions accompanied by other important factors such as road safety; vehicle quality and safety; and even signalization, policy, legislation, law enforcement, etc, it could improve safe behavior. 14 In brief, our findings are important to consider for stakeholders and policymakers. Driving education was not found to be an effective approach for improving the main outcomes, which are crashed and injuries. The reasons for this need to be considered in future research.
This study examines the efficacy of pre-LDE and post-LDE interventions to improve the safety of drivers (includes crashes, injuries, or secondary outcomes). A review of systematic review studies can give us a complete result. One of the strengths of this study was the qualitative review of the studies, most of which were of high quality. Also, a review of several large databases demonstrates the comprehensiveness of this study. How-ever, the limitations of the study should not be overlooked. Among other things, numeric data was not possible to perform meta-analysis and comment in numerical data. Also, there may be overlap between the initial studies included in the systematics understudy and have achieved similar results in all respects.

Conclusion
There is no evidence the pre-LDE or post-LDE driver education leads to a reduction in crashes or injuries. This negative finding may be due to the use of ineffective teaching methods, the absence of effective techniques for changing attitudes or future behaviors, failure to understand the needs of adult learners, or not targeting the correct risky behaviors.