Response to “Demirjian’s method is unsuitable for dental age estimation”

Excerpt

We wish to thank Jayaraman and Roberts for their comments to our article. We are rejoiced by the fact that they agree with our conclusion that Demirjian’s dental maturity data cannot be used to estimate age and that this practice should be totally abandoned. This is the sort of impact and support we were expecting from the academic community, and we are glad that they have previously provided such detailed data to support the argument. We apologize if we have failed to cite their earlier and substantial study in support of our own conclusion, but we think that their concern with our findings not being new and with our title not reflecting our goal is unwarranted. As in all published studies (reviewed extensively by Jayaraman et al. [1]), the age estimation accuracy of Demirjian’s method is only assessed by systematically comparing the estimated point age with the real chronological age. In our opinion, this comparison provides a very limited and superficial (if not incorrect) understanding of the forensic age estimation accuracy of Demirjian’s method. In a forensic situation, age cannot be estimated simply as a point value, but rather as a likely age interval. Together with the statistical approach, this is where the application of Demirjian’s method in a forensic situation fails first. Our paper was an attempt to push Demirjian’s method to its limits and use percentiles as a measure of age as a likely interval. This is where we believe our paper provides significant novelty. In addition to simple comparisons between the point estimated age and chronological, a forensic age estimation method should only be considered accurate if the true chronological age falls within the interval for a certain level of confidence or probability of age. We have shown that Demirjian’s method fails in this respect too, perhaps the most important when considering a forensic age estimation method. In this light, our title is but our conclusion summarized in the form of a question and, although we recognize some ambiguity to the title, it was not intended to mislead the reader in any specific direction. Thus, although we thank Jayaraman and Roberts for their suggestion about how to assess population differences in dental development using Demirjian’s method, this was not our goal, nor do we believe that the title is suggesting that Demirjian’s method is being tested on different population groups. The title seems adequate to emphasize our opinion that Demirjian’s method should not be used for forensic age estimation purposes. In this respect, although this opinion is aligned with that of Jayaraman and Roberts, it is not with that of the author’s previous study [1], where it is suggested that Demirjian’s data set should be used only with considerable caution (emphasis added) when estimating the age of group of subjects of any global population. It is our opinion that considerable caution will not overcome many of the problems with Demirjian’s method when it is used for forensic age estimation and thus there is a risk that its use may result in very real misapplications of justice. Finally, we agree that population studies should take into account several factors including sample size, age, and sex of the individuals, as well as socioeconomic and nutritional factors. Our study was in fact conducted in a sample of the Portuguese population, but the fact that we used a Portuguese sample is not the issue, per se. Rather, given Liversidge’s [2] work, where dental maturity curves were shown to be similar in different world regions and population groups, our Portuguese sample was used as illustrative of the already known problems (overestimation of age) and to further support our assertion that Demirjian’s method should not be used for forensic age estimation purposes. …