Modeling childhood agricultural injury risk with composite measurement scales.

摘要

Using longitudinal data recently (1999--2001) collected from 4-H youth in central Ohio, this dissertation's objective was to develop Composite Measurement Scales (CMSs) that parents can use to assess injury risk among youth who perform agricultural tasks. Variables available for modeling include several youth-based (e.g., selected physical and neuropsychological characteristics) and parent-based (e.g., parenting style) factors. For up to 13 weeks, daily data were obtained on time spent on each of 52 chores, and on multiple characteristics of injuries experienced. This dissertation focuses on injuries that occurred while youth led/groomed animals.; Multiple logistic regression was used to develop two CMSs: one with all variables and one with only variables readily knowable by a parent (e.g., youth age, gender, height, etc.). Regression coefficients from the fitted models were scaled and rounded to integers for ease of use. For each variable in the model, an integer score was obtained; a total score reflective of injury risk for leading/grooming animals was determined by summing the individual scores. The total scores were converted to probabilities. A nomogram was constructed so parents could easily determine their child's injury risk.; Age, gender, and mean time spent leading/grooming animals per week were in both models. BMI, youth's ability in sports as reported by a parent, mean reaction time, and standing steadiness were also included in the all-variable model. Parental assessment of their child's sports and learning ability, along with their perception of harm people do to themselves when using substances in excess, were the remaining variables in the knowable-variable model, which performed as well as the CMS developed using all variables. This finding has implications for future research because it may be possible to develop useful CMSs with data collected by less sophisticated techniques.; This dissertation serves as a template for future development of empirically derived risk-scoring systems. CMSs need to be developed for the many agriculture-related chores youth perform. The current CMSs need to be validated in a future study, and this must be followed by an evaluation of the effectiveness of the knowable-variable CMS for decreasing agriculture-related injury among youth who lead and groom animals.