Date of Award


Degree Type


Degree Name

Doctor of Philosophy



Major Professor


Committee Members



Dental comparison of antemortem and postmortem records provides one of the best avenues for establishing personal identification in the forensic sciences. The types of antemortem dental evidence are extensive (including treatment notes, odontograms, radiographs, casts, photographs, etc.) and in many instances a positive identification can be established strictly on a dental comparison. Perhaps the best form of antemortem dental evidence is the radiograph, which provides a detailed odontoskeletal record of a specific individual at a specific point in the past. Unfortunately, antemortem radiographic evidence is not always available during forensic comparisons. For example, at the U.S. Army Central Identification Laboratory, Hawaii (CILHI), dental radiographs are not commonly available when performing antemortem/postmortem comparisons of military personnel missing from past conflicts, especially those missing from WWII or the Korean War. In these instances, as well as some modern forensic cases, antemortem dental information may only be available in the form of handwritten charts and notes derived from the missing individual’s health documents. While these charts are susceptible to human error (not generally a concern with radiographs), dental information of this type that accurately documents an individual’s dental condition can be essential for establishing a link to an unidentified set of remains. Obviously, documents that are incomplete or inaccurate, on the other hand, will not assist in the identification process and could actually hinder the effort.

The goal of this dissertation is to validate the use of non-radiographic dental evidence for identification purposes. Statistically, there are trillions of possible combinations of missing, filled, and unrestored teeth within the adult mouth. This quantity of possible combinations suggests that an individual’s dental health pattern should often be of sufficient uniqueness to be used for identification. While the statistical model of possible combinations is mathematically plausible, it does not necessarily represent reality. Each of the 32 teeth in the adult dentition cannot be considered to be at the same risk for loss or disease. Dental morphology will dictate that molars, based on their large surface area, will be more susceptible to decay than other teeth, such as canines or incisors. Furthermore, all dental patterns are not equiprobable, signifying that some patterns will occur more frequently than others and statistical calculations of the total number of possible combinations of dental characteristics are not useful and are potentially misleading. Thus the theoretical number of possible dental health permutations should not be cited to justify the diversity of dental patterns for identification purposes.

In order to adequately address the issue of diversity in dental patterns, large datasets are needed for analysis. As part of this dissertation, four datasets were compiled that represent distinct demographic or temporal groups from the United States. These datasets are referred to as WWII-Korea, Southeast Asia, Modern Military, and Modern Civilian. With the exception of the Modern Civilian data, all other datasets consist of U.S. military personnel. The WWII-Korea and Southeast Asia datasets are composed of individuals missing in action from these conflicts, while the Modern Military and Modern Civilian data were originally collected as part of large dental health studies (the 1994 and 2000 Tri-Service Comprehensive Oral Health Surveys and the 1988-1994 National Health and Nutrition Examination Survey). Only permanent teeth were considered during this research, excluding third molars.

Initially, it was necessary to explore the accuracy of the dental evidence, specifically the military dental charts from WWII, the Korean War, and the Southeast Asia Conflict. In order to gauge the accuracy of the dental records, the Decayed, Missing, and Filled Teeth (DMFT) index was used to compare the WWII-Korea and Southeast Asia datasets with published results from temporally and demographically similar populations. The DMFT (Klein and Palmer 1937) is a popular index that is reported in many studies of dental health. Distinct variation between the published DMFT scores and those derived from the datasets used in this dissertation is likely indicative of incomplete/inaccurate recordation of treatment within the military dental records. As another test of the accuracy of antemortem dental records, a sample of dental charts was gathered from identified service members who were originally missing from WWII, the Korean War, or the Southeast Asia Conflict. The identification cases had been processed through either the CILHI or the CILTHAI (Central Identification Laboratory, Thailand) and were not part of the datasets used in this dissertation. The antemortem dental records were compared with the postmortem dental findings and the accuracy was assessed as a ratio of corresponding characteristics. It was found that the WWII and Korea records had an overabundance of individuals with “perfect teeth” (defined as the absence of decay and extraction throughout the mouth). In general, the WWII and Korea records were found to either be thoroughly documented or very poor, with the poor records lacking any documentation of treatment. The Southeast Asia cases were found to show excellent antemortem-postmortem congruence.

Next, the overall diversity of dental patterns formed by missing, filled, and unrestored teeth was explored for each of the datasets. As part of this process, the four datasets were transformed into two formats regarding the coding of fillings. Each dataset was coded in a detailed format in which all fillings were designated by the affected tooth surface. In the generic format, fillings were treated as either present or absent with a single code (i.e. there was no surface information coded). The diversity of dental patterns in both the detailed and generic formats was compared to the diversity found in mitochondrial DNA (mtDNA) sequences. The results of this dissertation show that the diversity of dental patterns, regardless of the data format, is on a scale that is comparable, if not superior, to mtDNA. Dental patterns were validated as an excellent means of forensic identification.

At this point it was essential to explore the diversity of specific dental patterns and to derive a method for quantifying the frequency that a specific pattern could be expected to occur. It was found that a method of empirical comparison to a relevant reference dataset is the most useful approach to the quantification of dental pattern frequency since this removes subjectivity and standards based on arbitrary points of concordance. This technique is nearly identical to the manner that mtDNA sequence frequencies are reported. Based on empirical comparison, it is possible to compare dental patterns formed by any combination of teeth and their characteristics. Postmortem loss is not a hindrance to the technique. It was found that very common dental treatment would often form a very unique dental pattern when all of the evidence is analyzed as a whole. This may be counterintuitive to many dentists. Furthermore, if numerous teeth are available in the postmortem analysis, the generic format of the data is sufficient to create very individualistic dental patterns. In situations of extensive postmortem loss, the detailed format will be critical to the establishment of individualistic patterns.

Prior to this research, forensic odontologists did not have a technique for assessing the strength of an antemortem-postmortem match between non-radiographic dental evidence. Up to this point, the comparison has usually been based on the subjective judgment of the dentist, which cannot be statistically quantified. Through empirical comparison with a large, representative dataset, dental patterns can now be objectively assessed. Patterns that may be initially hypothesized to be common in the general population could actually be shown to be extremely rare and individualistic based on empirical comparison to a reference dataset. By attaching an empirically derived probability value (the expected frequency that a specific pattern would be found in the population), matches based on dental patterns can be quantified in a manner that is easily defensible in a court of law. Two important points need to be understood as part of this research: 1) The end result of this research is not to create a database that can be used to match a dental pattern to a specific individual. Rather once an association to an individual has been made, the technique allows the significance of the dental pattern match to be quantified. 2) The use of non-radiographic dental evidence alone, asdiscussed in this dissertation, is not sufficient to establish a positive identification. It is anticipated that the evidence, in conjunction with other circumstantial evidence, can be used to form a very strong association between a missing individual and an unidentified set of remains that is beyond reasonable doubt.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Included in

Anthropology Commons