In Reply We thank Chung and Dang for discussing our article, "Assessment of Firework-Related Ocular Injury in the US." 1 The focus of this study was to characterize ocular firework injuries and inform potential regulatory efforts, especially because fireworks disproportionately affect the pediatric population. 2 We agree that the proportion of unspecified firework types in our study is a limitation; however, it is unlikely that the missing data fully account for the differences in our results compared with those of Chang et al. 3 As mentioned in our introduction, Chang et al 3 used data from a single medical center in the US Pacific Northwest. This is an important distinction, in that patterns of firework use may differ from the much larger National Electronic Injury Surveillance System (NEISS) database. While notable differences include the higher prevalence of mortars causing ocular injuries found by Chang et al 3 (24% vs <5% in our study 1 ), we believe that our shared findings that (1) mild injuries were the most common (67% corneal abrasion 3 vs 63% ocular burn 1 ), and (2) bottle rockets caused the most ruptured globes (35% 3 vs 56% 1 ) are comforting, despite differences in study design, and salient to future policy enactment.It is interesting to note that 2 studies, 2,4 like ours, that used the NEISS database to evaluate firework injuries also reported different proportions of unknown firework types (14.8% and 34.6%), which may be partially explained by the time elapsed between publication dates, differing NEISS casenarrative interpretation, and different study populations. 2 To handle missing data, Chung and Dang suggested using techniques outlined by Kang, 5 some of which we considered a priori. Since the Consumer Product Safety Commission uses their own modeling to prepare NEISS national estimates, we felt that it was most appropriate to present the data as is, to avoid introducing additional uncertainty. We agree that techniques described by Kang 5 could be used in future research, but prospective epidemiologic study of firework-associated injury may not be feasible on a large scale for an extended period of time.We respectfully disagree with Chung and Dang that a lack of ocular history data may be confounding, especially since most cases were mild and involved children who, as a group, are unlikely to exhibit relevant ocular history. Even so, this would not change our assertion that bottle rockets are especially dangerous.While the Consumer Product Safety Commission provides injury data in the NEISS database that would otherwise not be readily obtainable, limitations include missing data and clinical outcome details. These limitations are important to keep in mind when comparing and interpreting results of future studies in the context of previously published literature.