| Objective By investigating the injury among primary and secondary students in acertain pilot County of Hefei, we screen out children with multiple hurt,describe the distribution characteristics of injury and multiple injuries, andexplore the application of negative binomial regression and modifiedPoisson regression analysis in studying the factors influencing injuryfrequency, together with the risk factors leading to the increase of injuryfrequency, so as to offer some references and basis for formulating injuryprevention and interference measures and policies.Methods Based on epidemiological investigation, we make questionnaires amongrandomly cluster sampled2917primary and secondary students of a certainpilot County in Hefei, which then is fitted to the modified Poissonregression and negative binomial regression model. With the result, weselect the risk factors for increasing accidental injury frequency of juvenilestudents, so as to explore the efficiency of the three models in studying thefactors influencing injury frequency.Results⑴Injury situation: There were426damage occurred during one years in2917subjects. Total injury rate (person-based rate) was14.60%.507people werehurt,and hurting event rates (event-based rate) was17.38%. There were364(85.45%) single injuries occurred and the single injuries report rate was 12.48%. The single injuries report rate of boys (8.74%)were higher than thegirls’(10.68%, P <0.01). There were62people (14.55%) occurred multipleinjuries and multiple injuries report rate was2.13%.Among multipledamage group, there were no gender differences among all age groups.⑵The composition of the percentage between single and multiple injury wasdifferent in age, grade and parents divorce, and the difference wasstatistically significant (P <0.05).⑶55.8%of the children don’t remember the single injury occurred time,others’ respectively in July (5.2%), in August (6.9%), in September(6.0%).62.9%of the children don’t remember happen week, others’ wereWednesday (7.1%), Friday (7.1%), Saturday (5.2%).49.7%of the multipleinjury children don’t remember occurred time, others’ respectively in July(6.0%), in August (7.4%), in November (6.7%).57.7%of the children don’tremember happen week, others’ respectively was Monday (6.7%),Wednesday (6.7%), Friday (7.4%), Saturday (6.0%).The single and multipleinjury occurred highest at home (39.6%,47.5%). The single and multipleinjures largest injured area were fingers or toes and limb, injured propertiesmainly were skin, major damage intent were hurt.⑷The Lagrange multiplier indicated that the data of the Poisson modelexisted overdispersion (P <0.0001). Therefore, we analyzed the frequencydata with overdispersion using modified Poisson regression and negativebinomial regression, respectively. The results revealed that the modifiedPoisson regression model fitted better. Both of the methods showed that male,younger, father work outside the hometown，the guardian with junior highschool above, and smoking may result in the increase of injury frequency.Conclusion⑴Injury among rural children from a County of Hefei Anhui Province is frequent, and have a high incidence of injuries.⑵Through the three regression model fitting results all showed that boy,the smaller age, father staying out, guardian cultural degree for juniorhigh school above, smoking increased risk of injury frequency. Butwether divorce, and parents both staying out increased risk of injuryfrequency or not remains to be further research.⑶The rural children injury effects by the situation of staying out, soleft-behind children injury problems should be taken seriously.⑷For injury-frequency data with clustering tendency (ie, the larger ratioof the variance and mean), both of the modified Poisson regressionanalysis and negative binomial regression analysis can be used. In thisstudy, the modified Poisson regression fits better and would give a moreaccurate interpretation for relevant factors effecting injury frequency. |
PDF Full Text Request