Verifying Applicability of the Logistic Regression Formula for Gender Estimation through Odontometry of the Lower Canine in a Bogotan sampl
Objective: Evaluate the predictive validity of the logistic regression formula proposed by Mesa et al. for gender estimation in Colombian population. Methods: A sample of 98 mandibular canines extracted from human corpses was gathered at the National Institute of Legal Medicine and Forensic Sciences...
| Autor Principal: | Casas Martínez, Jorge Alfonso; Pontificia Universidad Javeriana / Instituto Nacional de Medicina Legal y Ciencias Forenses |
|---|---|
| Otros Autores: | Pinzón Jaime, Neyla Dayana; Pontificia Universidad Javeriana, Romero Díaz, Lilián Catalina; Pontificia Universidad Javeriana, Sánchez Cortés, Sandra Milena; Pontificia Universidad Javeriana |
| Formato: | info:eu-repo/semantics/article |
| Idioma: | spa |
| Publicado: |
Editorial Pontificia Universidad Javeriana
2009
|
| Acceso en línea: |
http://revistas.javeriana.edu.co/index.php/revUnivOdontologica/article/view/671 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: |
Objective: Evaluate the predictive validity of the logistic regression formula proposed by Mesa et al. for gender estimation in Colombian population. Methods: A sample of 98 mandibular canines extracted from human corpses was gathered at the National Institute of Legal Medicine and Forensic Sciences (INMLCF) in Bogota. Each tooth was carefully measured using an odontometer and repositioned to the alveolar bone. A database was created to collect the data that were analyzed with SPSS 12.0. Normal distribution of data was analyzed through the Shapiro-Wilk and Kolmogorov-Smirnov tests and the association between tooth measures and gender were determined through Pearson correlation. For the second part of the study, the modification proposed by Mesa et al. was used to estimate gender in the samples through diagnostic concordance; operative characterisitics such as predictive sensitivity, specificity, predictive values and diagnostic accuracy including confidence intervals were determined (Win-Episcope 2.0, α=0.05). Results: 62 samples belonged to deceased males of which 54 (87.8%) were adequately classified through the logistic regression model. On the other hand, 36 samples belonged to deceased females of which 17 were properly classified with the model. Of the total samples, 74.5% were properly classified. Conclusions: The logistic equation showed a greater predictive capacity in the estimation of gender in males than in females, failing to estimate 25% of the cases. |
|---|