Journal of Forensic Sciences vol 55, issue 5, Sep 2010

J Forensic Sci, September 2010, Vol. 55, No. 5 doi: 10.1111/j.1556-4029.2010.01399.x Available online at: interscience.wiley.com

PAPER PHYSICAL ANTHROPOLOGY Kristen M. Hartnett,1 Ph.D.

Analysis of Age-at-Death Estimation Using Data from a New, Modern Autopsy Sample—Part I: Pubic Bone*, 

ABSTRACT: This research tests the accuracy of age estimation from the pubic bone. Specimens were collected from decedents of known age, sex, and race at the Forensic Science Center (FSC) in Phoenix, Arizona. The collection consists of pubic bones and fourth rib ends from 419 males and 211 females, ranging in age from 18 to 99. Age-at-death was estimated by three observers using the Suchey–Brooks method. The correlation results indicate that there are significant differences in the observed versus actual ages (r = 0.68169, p < 0.001) and that there are significant interobserver differences. No significant differences were found in the intra-observer tests. The FSC pubic bones were sorted based on morphology without knowing age. New descriptions and age ranges were created. A phase seven was described and is comprised of males and females over 70 years of age-at-death.

KEYWORDS: forensic science, forensic anthropology, age estimation, pubic symphysis, pubic bone, osteology

While forensic anthropologists use a number of adult skeletal aging techniques, one of the most commonly used standards is for the pubic bone, which is often referred to as the pubic symphysis method in anthropological literature and practice. The pubic bone method relies on visual assessments of the topography and appearance of the surface of pubic symphysis, which demonstrates predictable degenerative patterns corresponding to increasing age. This research evaluates the accuracy of the pubic bone method developed by Suchey and colleagues (1,2). To perform this research, a large sample of pubic bones (N = 626) of known age was created to test the accuracy and precision of the accepted standards. In addition, sternal ends of the fourth rib from the same individuals were collected at the same time and will be presented in Part II of this research, which is the topic of another article. Major criticisms of the pubic bone age-phase method center on sample size and composition, population specificity, high inter- and intra-observer error rates, and the large age ranges of the phases established in the original studies (3–7). In addition, the Suchey– Brooks (SB) pubic bone sample is currently not accessible for further study. Recent statistical re-evaluations of age-phase methods suggest that the classical linear regression and other models used in the original studies may contribute to the reduced accuracy levels 1 Office of Chief Medical Examiner, 520 First Avenue, New York, NY 10016. *This research was funded by a Wenner-Gren Foundation Dissertation Fieldwork Grant, #7284, the ASU Dean’s Dissertation Writing Fellowship, the ASU Division of Graduate Studies Completion Fellowship, the Ellis R. Kerley Forensic Sciences Foundation Scholarship, an ASU GPSA Research Grant, a Forensic Sciences Foundation Acorn Grant, and an ASU Dept. of Anthropology Research Grant.   Presented at the 59th Annual Meeting of the American Academy of Forensic Sciences, February 22, 2007, in San Antonio, TX. Received 24 Nov. 2008; and in revised form 17 June 2009; accepted 27 June 2009.

 2010 American Academy of Forensic Sciences

because they do not accurately reflect true age phases, but transitions from age phase to age phase in the reference sample (3,4,8– 11). Based on the criticisms of the pubic bone techniques in the literature, three goals were established for this research: (i) to create a new, documented sample for future research and education, (ii) to evaluate the current SB standards on a large, modern, and diverse sample, and (iii) to propose revisions that increase the accuracy and precision of the method. Materials Skeletal specimens were collected from decedents of known age, sex, and race during examination at the Maricopa County Forensic Science Center (FSC) in Phoenix, Arizona, from January 11, 2005 through June 30, 2006. In addition, specimens were collected from Barrow Neurological Institute in Phoenix, Arizona, when available, to supplement the FSC sample. The total collection (N = 630) consists of pubic bones and bilateral fourth rib ends from 419 males and 211 females, ranging in age from 18 to 99 (Fig. 1). Individuals classified by the medicolegal system at the FSC as Asian (n = 4), Black (n = 20), Caucasian (including Hispanics; n = 598), and Native American (n = 8) were represented in the sample. In addition to the demographic information on age, sex, and race, information regarding the drug and alcohol history was obtained when available. Summary statistics such as counts, means, medians, modes, standard deviations, and ranges were calculated to help describe the different aspects of the research sample (Table 1). Because of preservation and preparation problems, as well as pathological or traumatic conditions in some individuals, both pubic bones and fourth rib ends were obtained and analyzed from 582 individuals (387 males and 195 females). Twenty-eight males and 14 females lack scoreable rib ends, yielding a total sample of 584 individuals for the rib analysis. Four males and two females are 1145

1146

JOURNAL OF FORENSIC SCIENCES

FIG. 1—Age and sex distribution of the total sample.

represented only by rib ends, yielding a total sample of 620 individuals for the pubic bone analysis. Methods When a decedent arrived at the Maricopa County FSC and fit the study criteria, the legal next of kin of the deceased person was contacted to request consent to include the decedent in the study. The phone conversation was conducted in compliance with the protocol previously submitted to and approved by the Arizona State University Human Subjects’ Institutional Review Board and was recorded to ensure proper consent was obtained. After each specimen was collected, a copy of the phone transcript and a letter with the contact information of the researcher was mailed to the next of kin for their records. Once permission to use the individual in the study was granted from the legal next of kin, two fourth rib segments and the pubic bones were removed with an oscillating saw or large clipper tool during examination. Each rib was cut at approximately 2 inches from the sternal end, and the pubic bones were extracted with two cuts on each side through the superior and inferior rami. Soft tissue was manipulated with a scalpel and forceps to facilitate extraction of the bone segments. The bone segments were macerated and then dried on racks in the laboratory. Each set was labeled with a random code number and the sex of the individual, according to the protocols set forth in the approved Institutional Review Board submission. Age was estimated using the SB casts and written phase descriptions for the pubic bone (2). Two volunteers, in addition to the author, participated in an inter- and intra-observer error study. One volunteer was a forensic anthropologist with a Ph.D. and over 20 years of experience in the field, whereas the second volunteer was a physical anthropologist with an M.A. and approximately TABLE 1—Basic statistics describing the ages of individuals in the FSC collection by sex.

FSC males FSC females BNI males BNI females Total males Total females Sample total

Count

Mean

Median

Mode

St. Dev.

Range

409 197 10 14 419 211 630

52.2 57.9 70.7 75.4 52.6 59.2 54.8

52.0 54.5 73.0 74.0 52.0 58.0 53.0

48.0 75.0 – 75.0 52.0 75.0 48.0

21.50 21.45 20.0 19.97 19.0 21.4 20.1

18–97 18–99 42–93 58–93 18–97 18–99 18–99

FSC, Forensic Science Center.

4 years of experience in the field. Each volunteer estimated the age of 50 males and 50 females using the SB pubic bone casts. After several days, each observer was asked to rescore a random selection of 15 males and 15 females that the author culled from the main sample to assess intra-observer reliability. All skeletal analyses in this research were performed with the knowledge of the sex of the individual, but without any prior knowledge of age. Next, the pubic bones were segregated by sex and sorted and seriated separately based on observed morphological characteristics of the symphyseal face as well as bone quality. Two observers jointly inspected both symphyseal faces of each individual and placed them in broad groups with similar morphology. Initially, the elements were placed into the broad and general categories of young, middle, and old adult based on visual inspection of the features of the symphyseal face (i.e., rim formation, appearance of face, bone quality). After this initial sort into broad groups, both observers further subdivided the groups using more specific character states and additional morphological characteristics. Together, both observers inspected every individual in each group to confirm that they fit the morphology of the group. Written descriptions of the general characteristics for each group were created, the actual ages were recorded, and the mean, standard deviation, and range were calculated. For the pubic bone, the morphological features that were considered to be the most discriminatory by the two observers were the extent of the ridge and furrow system on the face, the degree of symphyseal rim formation ⁄ deterioration, and bone quality. An individual was placed into the young group if he ⁄ she had extensive ridges and furrows on the face, an unformed or incomplete rim, and good bone quality. A person was placed into the middle-aged group if the pubic bones had slight ridges and furrows or remnants on the face, a partial or complete rim, and good bone quality. Finally, an individual was included in the older adult group if there were no ridges and furrows on the face, the rim was complete, and the bone quality was fair to poor. Several diverse statistical tests were run on the data collected from the bone segments, ranging from simple summary statistics to complicated Bayesian analyses. The summary statistics, correlation, and linear regression procedures are discussed later. The Bayesian transition analysis results will be discussed in a future study. All statistical analyses were performed using the SAS Statistical Software version 9.1.2. Descriptive statistics were calculated for the different types of data collected, and Spearman’s coefficient of rank correlation was used to examine the relationship between the observed age-phase estimates and the actual phases for the three observers. In addition, Spearman’s correlation was used to test intra-observer reliability. For comparison with the SB aging method, least-squares linear regression analyses with resulting r values were conducted on this new sample. Generally, traditional regression analyses are inappropriate for the categorical data produced in this study because regression is typically used when the data are continuous. In addition, regression has been shown to underage the elderly, and overage young individuals (11) and is very sensitive to the age composition of the reference sample (3). Strictly for ease of comparison with previous studies, however, regression analyses were performed. Results The results of the correlation analyses describing the relationship between the known age-at-death and SB phase (actual) and the estimated phase (observed) are presented in Table 2. The correlation

1147

HARTNETT • AGE-AT-DEATH ESTIMATION USING THE PUBIC BONE TABLE 2—Correlation of estimated versus actual Suchey–Brooks phases for each observer in the combined sex sample.

Author Observer 1 Observer 2

Count

Spearman’s r

p Value

620 100 100

0.64439 0.69936 0.58048