Increased police patrols for preventing alcohol-impaired driving.

BACKGROUND: Road traffic injuries cause 1.2 million deaths worldwide each year. Alcohol consumption increases the risk of traffic crashes, especially fatal crashes. Increased police patrols aim to increase both the perceived and actual likelihood of being caught driving while alcohol-impaired, potentially reducing alcohol-related driving, crashes and injuries. OBJECTIVES: To assess the effects on injuries and crashes of increased police patrols that target alcohol-impaired driving. SEARCH STRATEGY: We searched the Cochrane Injuries Group Specialised Register (5/2006), CENTRAL (The Cochrane Library 2006, Issue 2), MEDLINE (1966 to 5/2006), TRANSPORT (1968 to 5/2006), C2-SPECTR (2/2005), NCJRS (1/1951 to 5/2006), PsycINFO (1872 to 5/2006), Social Science Citation Index (1974 to 5/2006), SIGLE (1980 to 2/2006), Science Citation Index Expanded (1970 to 5/2006), Dissertation Abstracts (1870 to 5/2006), NTIS (1964 to 12/2004), conference proceedings, and reference lists. We contacted authors of eligible studies. SELECTION CRITERIA: Randomized controlled trials, controlled trials, controlled before and after studies, interrupted time series (ITS) studies, and controlled ITS studies evaluating increased police patrols, either alone or combined with other interventions, targeting alcohol-impaired motor vehicle drivers. DATA COLLECTION AND ANALYSIS: Two investigators independently screened citations, extracted data, and assessed quality criteria. We compared intervention and no-intervention geographical areas or time periods. We re-analyzed study data as required. Results are presented narratively. MAIN RESULTS: The 32 eligible studies included one randomized controlled trial, eight controlled before-after studies, 14 controlled ITS studies, six ITS studies, and three studies with both ITS and controlled before-after analyses. Most interventions targeted only alcohol-impaired driving (69%) and included additional interventions such as media campaigns or special training for police officers (91%). Only two studies reported sufficient information to assess study quality completely. Two-thirds of studies were scored 'not adequate' on at least one feature. Five of six studies evaluating traffic fatalities reported reductions with the intervention, but differences were statistically significant in only one study. Effects of intervention on traffic injuries were inconsistent in the six studies evaluating this outcome, and no results were statistically significant. All four controlled studies evaluating fatal crashes reported reductions with the intervention, which were statistically significant in one study. All 12 controlled studies assessing injury crashes reported greater reductions with the intervention, though effects were minimal or not significant in several studies. ITS studies showed less consistent effects on fatal crashes (three studies) and injury crashes (four studies), and effect estimates were typically imprecise. Thirteen of 20 studies showed reductions in total crashes and about two-thirds of these were statistically significant. AUTHORS' CONCLUSIONS: Studies examining increased police patrol programs were generally consistent in reporting beneficial effects on traffic crashes and fatalities, but study quality and reporting were often poor. Methodological limitations included inadequate sample size, dissimilar baseline measures, contamination, and inadequate data analysis. Thus existing evidence, although supportive, does not firmly establish whether increased police patrols, implemented with or without other intervention elements, reduce the adverse consequences of alcohol-impaired driving.

Single-factor repeated-measures designs: analysis and interpretation.

In this column we discussed the selection and interpretation of appropriate statistical tests for single-factor within-subjects/ repeated-measures designs and provided an example from the literature. The parametric tests that we discussed were the t test for paired or correlated samples and the single-factor repeated-measures ANOVA. We also mentioned four nonparametric tests to be used in single-factor within-subjects/repeated-measures designs, but they are relatively rare in the literature. The Compton et al. (2001) article did not provide effect size measures, but they could be computed from the means and standard deviations. Remember that a statistically significant t or ANOVA (even ifp < .001) does not mean that there was a large effect, especially if the sample was large. In the Compton example, the sample was quite small (N = 14), and the findings do reflect a large effect size.

Selection of inferential statistics: an overview.

This column serves as an introduction to selection of appropriate statistical methods. In the next five columns we will discuss conceptually, and in more depth, these statistical methods. We will use clinical examples and discuss why the author(s) selected a particular statistical method and how the results of the statistical method were interpreted.

General design classifications.

This column described the general design classifications of between-groups, within-subjects, and mixed designs. Remember that in between-groups designs, each participant is in only one group or condition. In within-subjects or repeated-measures designs, on the other hand, each participant receives all the conditions or levels of the independent variable. In mixed designs, there is at least one between-groups independent variable and at least one within-subjects independent variable. In classifying the design, do not consider the dependent variable(s). The classifications and descriptions presented in this column are for difference questions, using the randomized experimental, quasi-experimental, and comparative approaches to research. Appropriate classification and description of the design are crucial for choosing the appropriate inferential statistic, which is the topic of the next column and several to follow.