Bioanthropologists, forensic practitioners, or radiologists can rely on one or several of hundreds of methods available to estimate the skeletal or dental age of living or deceased juvenile (or sub-adult) individuals. Despite being published in scientific journals or renowned osteology or bioanthropology books, national, institutional or personal preferences and the excessive number of methods or tools mean that only part of them are routinely applied in anthropological practice. Therefore, many of them remain unknown in the anthropological, bioarchaeological and forensic communities where they could be used on a regular basis instead. More importantly, some of these forgotten or omitted methods are sometimes actually better than some of the more routinely approaches adopted for sub-adult age estimation. Indeed, the methodological “quality”, and the scientific and biological validity of the sampling and statistical protocols of juvenile age estimation methods are extremely variable. Therefore, the accuracy and reliability of the resulting age estimates are extremely uneven, and we can even objectively say that some methods are better than others.
Several publications (Cattaneo 2007; Cunha et al. 2009; Ritz-Timme et al. 2000; Rösing et al. 2007; Schmeling et al. 2007) have tried to objectively highlight the need for standardized methodological sampling and statistical protocols in anthropological methods. This objective evaluation can ensure that estimates are objectively evaluated and compared to one another. The authors of these papers provide guidelines to help with valid and standardized method construction: these guidelines concern both the samples the methods are constructed on, which ought to be reference samples of collections, and several statistical parameters that need to be presented, such as reliability, accuracy, and others. The authors present these guidelines as crucial to obtaining scientifically valid methods and, in this context, age estimates.
These methodological guidelines for method construction and application are the main idea behind SAMS, a decisional tool developed to help anthropologists objectively evaluate and select juvenile age estimation methods. SAMS groups over 250 methods for juvenile/sub-adult age estimation. Each method recorded in SAMS was decomposed into 23 criteria relating to the sampling, methodological and statistical parameters presented in the original publications. The criteria are standardized to ensure uniformity and comparability between all methods (Corron et al. 2018).
Cattaneo C. 2007. Forensic anthropology: Developments of a classical discipline in the new millenium. Forensic Science International 165:185-193.
Corron L., Marchal M., Condemi S., Adalian P. 2018. A critical review of sub-adult age estimation in biological anthropology: do methods comply with published recommendations? Forensic Science International 288:328.e1-328.e9
Cunha E. Baccino E. Martrille L. Ramsthaler F. Prieto J. Schuliar Y. Lynnerup N. Cattaneo C. 2009. The problem of aging human remains and living individuals: a review. Forensic Science International 193:1-13.
Ritz-Timme S. Cattaneo C. Collins M.J. Waite E.R. Schütz H.W. Kaatsch H.-J. Borrman H.I.M. 2000. Age estimation: the state of the art in relation to the specific demands of forensic practice. International Journal of Legal Medicine 113:129-136.
Rösing F.W. Graw M. Marre B. Ritz-Timme S. Rothschild M.A. Rotzscher K. Schmeling A. Schroder I. Geserick G. 2007. Recommendations for the forensic diagnosis of sex and age from skeletons. Homo - Journal of Comparative Human Biology;58(1):75-89.
Schmeling A. Geserick G. Reisinger W. Olze A. 2007. Age estimation. Forensic Science International 165:p.178-181.
Go to the Search button on the upper left side of the screen and click on it. This decisional tool is based on three search criteria that you will choose from: the anatomical region the element(s) belong to, the type of skeletal or dental element available, and the type of age indicator the user can obtain from these elements.
Click in each dialog box and select any number of Anatomical Regions and/or Bone elements you wish. The Bone criteria are filtered according to Anatomical Region. For example, if you choose the Anatomical Region Scapular Girdle, you will only be able to select Clavicle and/or Scapula as Bone criteria. If you choose Scapular Girdle and Upper Limb as Anatomical Regions, you will be able to choose Clavicle and/or Scapula and/or Humerus and/or Radius and/or Ulna in the Bone section.
In the same way, you can select one or several Indicator Type. These are the age indicators you could take on the anatomical elements you have. They are either quantitative (e.g. bone measurements), or qualitative (e.g. epiphyseal stages). You can also choose both types or choose not to use this search criteria by unclicking Include.
You can choose to display between one and 20 descriptive criteria in the results, by selecting them in the Descriptive Criteria box. By not selecting any of these criteria, all 20 of them will be displayed in the results.
When you have selected a combination of these search criteria, click on the Filter… button to filter the database and reveal which publications correspond to your search.
Publications are indicated by a reference (name of first author and original date of publication) and a URL, which you can click to access the online publication hosting site, when available. The publications that correspond the most to your search criteria will have the highest Relevance values (from 0 to 1).
Moreover, another algorithm will also provide a Validity value for each publication that comes out in the results. Validity evaluates how much the publications follow the methodological guidelines published by Cattaneo 2007, Cunha et al. 2009, Ritz-Timme et al. 2000, Rösing et al. 2007, and Schmeling et al. 2007. Publications with Validity values close to 1 comply the most with these recommendations.
A composite Score of Relevance and Validity is also provided, to help with method evaluation and ranking. You can conduct a quick evaluation of the resulting methods with the additional descriptive criteria. These additional features can help justify the application of particular methods in reports by presenting their main sampling and statistical characteristics. This seems particularly useful for forensic anthropologists who need to meet particular levels of scientific validity in order for their reports to be valid in Court.
We hope you enjoy SAMS and find it useful!
The authors and creators of SAMS
The search criteria are the basis of the classification. Users select between one and three parameters relating to the anatomical element(s) at their disposal for age estimation and the type of age indicator (quantitative, qualitative, both or either) they can obtain from the element(s). After selecting the parameters, the user clicks on the “Filter” button to visualise the results.
This criterion acts as a first filter for identifying the anatomical element(s) on which the age indicators will be obtained from. Several anatomical regions can be selected, depending on the number of elements the user has at his/her disposal.
This criterion is essentially the selection of the skeletal and/or dental element from which the variables used for age estimation are obtained. It is either a single bone (e.g. humerus), several bones (e.g. limb bones: humerus, radius, ulna, femur, tibia and fibula), an anatomical zone (e.g. hand/wrist), or specific developmental states (e.g. deciduous teeth, mixed dentition, permanent teeth). First, the user selects the general anatomical region of interest which narrows the choices to the corresponding skeletal or dental elements of the selected region. Several elements can be selected for search.
Indicator type refers to the predictor variables: they are either quantitative data, such as skeletal measurements, or qualitative data, such as skeletal or dental development or skeletal maturation stages. Depending on skeletal or dental preservation, the user can sometimes only use one type of indicator to predict age and methods are often based on one type or the other.
SAMS main result is given as two major components describing existent juvenile age estimation methods.
The Reference is the name of the first author (first and second if they are only two, first et al. if there are more than two authors) and the original year of the publication. The URL takes you to the website hosting the publication.
Descriptive criteria are different sampling, methodological and statistical characteristics presented in the publications that provide additional information to help users with method selection and evaluation.
The geographical origin of the individuals in the sample used to build the method, mainly relating to the country of origin of the individuals, but sometimes corresponding to an “ethnic” group or sub-population affiliation. The methods referenced in SAMS cover 29 countries so far.
This criterion informs the user on the general chronological period the sample used to build the method dates from. The periods included are contemporary (the sample dates from the same time the method was published), modern (18th and 19th centuries), and archaeological (17th century and earlier).
Sample size was arbitrarily subdivided into five categories from less than a hundred individuals to more than a thousand: [<100], [100-200], [200-500], [500-1000], [>1000].
The type of study was either cross-sectional, semi-longitudinal or longitudinal. A cross-sectional study uses different and random individuals, who only contribute once to the sample, at a given point in time, by the developmental state they are in at that moment. One individual belongs to one age group. A longitudinal study follows the same individuals from a starting point to a finish point in time, and data is repeatedly collected at specific moments for each individual between these two dates.
The precise age ranges of the individuals from the study sample on which the method was built. This criterion varies greatly from method to method, covering different ranges of ages from birth to adulthood. This gives an indication on whether the method is relevant to the user or not if the bone(s) available for estimation correspond to the methodological age range or not.
Age and sex
Age (and respectively sex) relates to whether the age (and respectively sex) of the subjects composing the sample on which the method was constructed was known or unknown when the method was constructed.
The age distribution of the study sample used to construct the method. Distribution was either even, meaning all age groups are represented by comparable numbers of individuals, uneven or unknown.
The sex distribution of the study sample used to construct the method. Distribution was either even, meaning the number of male individuals is comparable to the number of females, uneven, or unknown.
This criterion informs us on whether the method was built separately for males and females or not. In the case of sex-specific methods, sex must be determined prior to age or already known.
The specific variable type used for age prediction (e.g. bone measurement, dental mineralisation stage, etc.) in the corresponding method. Again, this criterion is extremely variable, and we advise the user to verify the methods’ applicability in the original publication depending on the preservation of the elements available.
Medium of study
The medium used for acquisition of the predictor variables in the corresponding method, such as dry bone, radiography, computed tomography, biochemical analysis, etc.
This criterion is a general description of how the method is applied and how the results are obtained. It provides additional information on the method for the user, prior to its application. Results can either be based on age per stage/score, based on descriptive statistical parameters (mean, standard deviation, frequencies, etc.), based on regression equations, based on atlas, abacus, tables, diagrams or growth curves, based on probabilistic inference, or unknown.
Accuracy or validity is the degree of conformation of a measured or calculated value to its (actual) true value. It can be presented as the percentage of successful correct estimations, i.e. the percentage of estimated values that do not differ significantly from the real values for a given error risk α.
The exact accuracy value of the age estimates of the corresponding method, when known.
Standard Estimation Error (SEE)
Standard error (SEE) or standard deviation (SD) is the degree to which further calculations (or estimations) give similar results (or predicted ages). It is the standard deviation of the estimation errors. Standard error is estimated following the normal distribution of the estimation errors. The standard estimation error multiplied by 1.96 is the precision of the method associated to 95% reliability. It is often indicated under the mathematical expression +/- X years, with X = 1.96SEE.
The exact value of the standard estimation error of age of the corresponding method, when known.
Testing intra- and inter-observer errors of the predictor variables is a prerequisite for any methodology. Indeed, if variable acquisition is observer-dependent, the variables cannot be considered objective and there is a risk of error during acquisition, which can bias the results. Different types of tests for repeatability (intra-observer error) and reproducibility (inter-observer error) exist to test variable acquisition (t-tests, Wilcoxon tests, Cohen’s kappa…). Seven modalities were identified for this criterion: not tested, presence of intra- and inter-observer error, presence of inter-observer error, presence of intra-observer error, absence of intra-observer error, absence of inter-observer error, absence of intra- and inter-observer error. This criterion was assessed when the concerned method included it, or when it was tested in different publications.
Methodological validation is the last statistical criterion selected for method evaluation. Testing the method is a prerequisite for methodological validation. It aims to verify the applicability of the method on different samples, and normally, it is the way to calculate the accuracy of the method. Several validation techniques can be used by the authors: cross-validation is a way to test the method on a subset of the study sample that is not used to construct the method. It is also a way to construct the method by alternatively including and leaving-out several individuals to construct and test the method. Validation can also be done using an independent test sample by the authors themselves or by different researchers in an independent study either to validate, invalidate it/point out inconsistencies, modify it or confront it with other methods. The question remains as to whether or not this can be considered as validation, but since several methods are in fact used because of repeated independent testing, it was considered as a legitimate way of assessing methodological validity. If no validation technique was found in literature (in the original article or any article mentioning it) but the method or study is mentioned, validation was indicated as “not done” in the database. If no reference was made of validation either in the original article or literature, validation was marked as “unknown”.
Reliability is the degree of conformity between a real value and an estimated value. For an age estimation method, reliability is expressed as the confidence interval of the estimated age. It is typically set at 95%, and is associated with a standard estimation error. The size of a 95% confidence interval is equal to twice the standard error associated with the 95th percentile multiplied by 1.96.
SAMS search algorithm will classify the publications according to the users criteria input using two parameters:
Relevance corresponds to how much the methods are close to the features selected for the three search criteria selected: Anatomical region, Bone and Indicator type. The higher the Relevance (i.e. the closer it is to 1), the closer the method is to the user’s selected search criteria.
Validity is how much the methods respect what are considered valid sampling and statistical criteria by the authors of methodological guidelines cited previously. Publications with high V values (i.e. closer to 1) comply the most with these methodological guidelines. In this context of juvenile age estimation, the valid criteria are the following:
A final Score (S) is computed based on these parameters. This composite value, given by the product of relevance and validity, gives an overall ranking of the publications according to R and V values. Publications with high S values are the ones that correspond to the users’ choice and respect methodological guidelines the most.