Age estimation using the sternal end of the clavicle: a test of the Falys and Prangle (2014) archaeological method for forensic application
Price, Meghan D.
MetadataShow full item record
Age estimation is a critical component of the biological profile in forensic and bioarchaeological contexts. The majority of age estimation methods are most accurate for individuals of younger age cohorts, typically those under 40 years of age. Skeletal degeneration can vary greatly between individuals, making age estimation less accurate for adult individuals. While there are some methods that attempt to age older individuals accurately and precisely, more research must be conducted to expand the range of methods available. Falys and Prangle (2014) developed a method for estimating age in individuals over the age of 40 using three degenerative characteristics of the sternal end of the clavicle: (1) surface topography, (2) porosity, and (3) osteophyte formation. In order to test their method, a sample of 1,510 individuals of known sex and age, ranging from 20 to 101 years of age (males: n = 1112, mean = 50.57, SD = 18.015; females: n = 398, mean = 53.065, SD = 20.358), were drawn from the McCormick Collection and the William M. Bass Donated Skeletal Collection at the University of Tennessee. Due to the paucity of remains of other ancestries, only individuals of reported White ancestry were used in this study. The two estimation methods proposed in Falys and Prangle (2014), regression equation and composite score, were tested to see how well they perform when applied to a different sample population than the populations used to develop the method. When applied to the collected data, the regression equation produced age estimations that fell within the 95% confidence interval in 47.6% of the male sample and 57.4% of the female sample. Composite scores were calculated and compared to the corresponding age ranges provided in Falys and Prangle (2014). The composite scores of the male sample estimated the age of an individual more accurately than the composite scores of the female sample (male = 65.9%; female = 58.8%). The lowest estimation accuracy for both males and females was between 70-79 years of age (male = 46.0%; female = 51.4%). From 80-89 years of age, the accuracy increased for males (76.4%) and females (69.4%). The sample also included individuals under the age of 40 in order to test whether the inclusion of clavicles with recent epiphyseal union would affect the applicability of the Falys and Prangle (2014) method. Multiple regression equations were generated: (1) individuals over 20 years of age, (2) individuals over 30 years of age, and (3) individuals over 40 years of age. The results from the multiple regression analyses show comparable Pearson’s coefficients for the above mentioned equations (r = 0.690, r = 0.632, and r = 0.611, respectively). Spearman’s rank correlation coefficients indicated a correlation significant at the 0.01 level for all three components individually, as well as the composite score. Of the three components, surface topography was most strongly correlated with age for both males (r = 0.643) and females (r = 0.590). Unlike the findings of Falys and Prangle (2014), porosity was found to be the least correlated with age for both males (r = 0.474) and females (r = 0.514). In addition, when broken down into ten year intervals (40-49, 50-59, etc.), the correlation coefficients increase with advancing age. This suggests that the method becomes more accurate as the age of an individual increases. The inter-observer and intra-observer agreement tests produced very low agreement values. The low observer agreement indicates that the current scoring method is not a reliable, repeatable technique. However, when examined further, the observed trait values that differed between the tests primarily differed by one score. These results suggest that condensing the scores in order to account for more variation would likely increase the observer agreement. However, condensing the scores would result in larger age intervals, which nullifies the purpose of this method. The findings in the present study indicate that the sternal end of the clavicle has potential for use in age estimation in older individuals. Although the present study produced lower correlation coefficients than proposed by the original study in 2014, the correlations and age-at-transition test results suggest that the sternal end of the clavicle deteriorates in a predictable manner that, with more observation and understanding, could be used to accurately age older individuals more precisely than the large age intervals currently in use. Despite the correlations between the degeneration of the sternal end of the clavicle and the age-at-death, the error rates suggest it is not a reliable alternative to the current methods used.