Dear Sir, I happened to notice the article entitled ‘Temperamental and socioeconomic factors associated with traumatic dental injuries among children aged 0–17 years in the Swedish BITA study’ by Oldin et al. (1) among the articles published online in your journal, Dental Traumatology. As a final report of the comprehensive, long-term BITA study, the authors concluded that no association existed between traumatic dental injury (TDI) and temperamental reactivity, whereas the more social and active children at 11 and 15 years of age were less likely to sustain TDI. Although attention deficit hyperactivity disorder (ADHD) and child temperament are regarded as two different constructs, strong correlations between the symptoms of ADHD and dimensions of childhood temperament were demonstrated before (2). As the findings of the above-mentioned study apparently contradict the ADHD model for TDIs (3), there is a need to clarify what factors influenced the results. When I reviewed the article, I had a hard time understanding the authors’ conceptualization of the problem, and the way data analysis was performed. As the authors’ hypothesis was that the risk of having TDI in the primary and mixed/permanent dentition is affected by temperamental reactivity, they need to compare the temperament data of children with TDI to children without TDI. However, statistical analysis was not carried out with this perspective. As can be seen in Table 1 and Table 2, the databases of children having no TDI and having a single episode of TDI have been merged, forming a disproportionately large group in the total sample. In other words, normal and pathological conditions have been treated as the same, which is apparently incorrect and unacceptable. Comparison of this ‘merger’ group with children having multiple episodes of TDI, what the authors have done in Table 1 and Table 2, may reveal nothing logical and meaningful. The surprising end result of the study, no association between TDI and temperamental reactivity, may simply originate from this methodological issue. Likewise, the relevance of ‘0 or 1 TDI Group’ in Table 3 and Table 4, which aim to display the influence of socioeconomic factors, is of question. What I would like to suggest is to initialize data analyses by forming no TDI group and single–multiple TDI group first and, then, to compare temperament data of both groups. Further, the authors may consider comparison of multiple TDI group with single TDI and no TDI groups. Consequently, the results must be reinterpreted and the Discussion section must be rewritten in light of correct findings. In sum, this letter aims to enable better handling of the database of an important project in which significant time and effort were invested. Constructive feedback among peers let scientific knowledge advance from which children benefit.
Download Full PDF Version (Non-Commercial Use)