Graduated driver licensing for reducing motor vehicle crashes among young drivers
L Hartling,N Wiebe,K Russell,J Petruk,C Spinola,T P Klassen
DOI: https://doi.org/10.1002/14651858.CD003300.pub2
Abstract:Background: Graduated driver licensing (GDL) has been proposed as a means of reducing crash rates among novice drivers by gradually introducing them to higher risk driving situations. Objectives: To examine the effectiveness of GDL systems in reducing crash rates of young drivers. Search strategy: Studies were identified through searches of MEDLINE, EMBASE, CINAHL, Healthstar, Web of Science, NTIS Bibliographic Database, TRIS Online, SIGLE, the World Wide Web, relevant conference proceedings, consultation with experts and authors, and reference lists. The search was not restricted by language or publication status. Selection criteria: Studies were included if: 1) they compared outcomes pre- and post-implementation of a GDL program within the same jurisdiction, 2) comparisons were made between jurisdictions with and without GDL, or 3) both. Studies had to report at least one objective, quantified outcome. Two reviewers independently screened searches and assessed the full text of potentially relevant studies for inclusion using a standard form. Data collection and analysis: Data were extracted by one reviewer and checked by a second. Additional data were requested from authors. Results were not pooled due to substantial heterogeneity between studies. Percentage change was calculated for each year after the intervention, using one year prior to the intervention as the baseline rate. Results were adjusted by internal controls. Analyses were stratified by different denominators (population, licensed drivers). Results were calculated for the different crash types (overall, injury, fatal, night-time, alcohol, and those resulting in hospitalization). Results were presented for 16 year-olds alone and all teenage drivers combined. Main results: We included 13 studies evaluating 12 GDL programs that were implemented between 1979 and 1998 in the US (n=7), Canada (3), New Zealand (1), and Australia (1). Programs varied in their restrictions during the intermediate stage: e.g. night curfews (8); limitations of extra passengers (2); roadway restrictions (1). Based on the Insurance Institute for Highway Safety classification scheme, no programs were good, six were acceptable, five were marginal, and one was poor. Reductions in crash rates were seen in all jurisdictions and for all crash types. Among 16 year-old drivers, the median decrease in per population overall crash rates during the first year was 31% (range 26-41%). Per population injury crash rates were similar (median 28%, range 4-43%). Results for all teenage drivers, rates per licensed driver, and rates adjusting for internal controls were generally reduced when comparing within jurisdictions. Reviewers' conclusions: The existing evidence shows that GDL is effective in reducing the crash rates of young drivers, although the magnitude of the effect is unclear. The conclusions are supported by consistent direction of the findings, and the temporal relationship and plausibility of the association. The reviewers have made recommendations for primary research on GDL (e.g. study methods, standardized reporting of outcomes and results, long-term follow-up). The project has also shown what is needed to carry out systematic reviews of observational studies (e.g. quality assessment instruments).