|
Abstract
|
Statement of purpose To review the methodology of the UL Safety Index, an algorithm-based data science initiative providing people with information to make better policy choices to improve safety.
Methods/approach A literature review was conducted to identify other safety indices and provide a basis for comparative analysis. The OECD Handbook on Composite Indicators was used to conduct a methodological review, including statistical analysis for consistency, sensitivity and robustness.
Results Composite indicators have become commonplace tools to rank country performance in a variety of fields. However, the literature is void of a composite indicator for unintentional injury. The UL Safety Index uses similar methods and techniques as other composite indicators in the areas of variable selection, multivariate analysis, normalisation, weighting, aggregation and presentation. Correlation analysis of the Index reflects that the drivers are associated with reduced injury at a 5% statistical significance level. Cronbach's analysis shows high levels of internal consistency and reliability within the Index, with a standardised C-Alpha of 0.86. Sensitivity analysis by inclusion/exclusion of individual components of the Index and re-ranking reveals that it is stable across country rankings, with highest sensitivity to the safety frameworks construct.
Conclusions The UL Safety Index uses a proven approach to develop a composite indicator that is transparent, methodologically sound, internally consistent and robust. The sensitivity of the model to safety frameworks is appropriate given the use of these societal interventions as tools for policy. The Index is appropriate for use as one source of information for policy makers and practitioners in decision making to improve safety.
Significance and Contributions to Injury Prevention Science There are no examples of comprehensive composite indicators of unintentional injury at the national level. The development and validation of such an indicator may improve the implementation of evidence based policies to improve safety.
Language: en |