We respectfully submit the following to the … Commissioner …. We trust that our observations will be of help in determining the appropriate management of children who are crew member of boats engaged in people smuggling.
Our experience is limited to Dental Age Assessment of children, adolescents and emerging adults referred to the Dental Age Assessment Team (DAAT) at King’s College Dental Institute.
The DAAT comprises Professor Graham Roberts (research lead) Dr Victoria Lucas (clinical lead) Dr Jayakumar Jayaraman (researcher) and Ms Christine Bell (co - ordinator). In addition to this we are from time to time joined by Master of Science and PhD students who make a significant contribution by the acquisition and addition of data to the research datasets. These contributions are acknowledged in the appropriate places in the text.
We are aware that a number of bodies including the Australian Association of Forensic Odontology submitting evidence. To this end we are utilising the simple approach of Question and Answer to avoid repetition.
Route Map through this Document
We have used an unconventional but we hope helpful way of providing supporting references. Throughout the text we have used the traditional format which is to cite references as:
[Chudasama et al 2012]. The full reference is then given in the bibliography. These have been cited in alphabetical order.
To assist the inquiry we have also provided a digital copy of the reference and to enable readers to obtain the details from the reference we have provided pdf copies of all the references cited in the bibliography. The filenames are provided thus
This streamlines the process of looking up references from the bibliography.
All the efiles are in the zip folder titled Age_Assessment_Inquiry2011_2012_e_files
[Please contact the Commission if you would like to receive any of these documents.]
We have been carrying out a number of age assessments for several decades. For this we have relied on published data in textbooks. In 2003 the number of referrals increased and in anticipation of questions about accuracy and reliability we decided to set up our own research database to enable us to become familiar with the detailed methodology of age assessment. This DAA database stands at 7,600 plus cases. The awareness of the availability of this data has resulted in over 700 DAA requests since 2004. An important outcome is that despite 3 or 4 vigorous challenges through legal services our findings have always been upheld. We continue to provide DAA on 3 or 4 occasions each month. The numbers have reduced because of severe limitation of legal aid funding. Negotiations are currently nearing completion to make funds available for a large number of clients still requiring age assessment.
A preliminary communication of the first 500 cases was presented to the 7th International Conference on Forensic Inference and Statistics, Lausanne, Switzerland in August 2008. The findings were well received and we were complimented on the thoroughness and detail of our work. [ see Lucas VS, Souberbielle A, Roberts GJ 2008 ] – [ efile Lucas_VS_et_al_2008_Lausanne ]
Over a period of 2 years, an explanatory document aimed principally at solicitors, barristers and social workers and other dentists was developed. This was seen by a colleague in the Max Planck Institute in Berlin and we were invited to contribute a chapter in a book on Digital Forensics. The Reference for this is Dental Age Assessment: a Practical Approach. Roberts GJ & Petrie A, Dental Age Assessment: A
Practical Approach. Chapter 11 in Digital Forensic Science. IGI Global USA. 25 March 2011.
For copyright reason we are unable to send you a copy of the chapter without paying the copyright fee of $58 USA. The essential elements of this chapter are in the draft publication, Roberts GJ and Petrie A 2008, a copy of which is attached. This comprises the two files below.
[ efile 2011_DAA_Paper08A_Practical_GuidePDFand efile 2011_DAA_Paper09Appendix06_Table 06 Full summary Data ]
Is the consent process for subjects of unknown date of birth robust?
To carry out a Dental Age Assessment it is necessary to obtain informed consent. The subjects are primed of this process by the referring social worker or solicitor and when she / he attends the DAA clinic at King’s College Dental Institute the consent process is repeated, with the assistance of an interpreter if necessary, and the consent signed. The Consent committee at King’s College Hospital were asked to review our consent process. This was done during 2009 and the formal approval is given in the letter form Dr Alistair Baker, Chair of the Consent committee
[ efile Consent_working_group_kch_Approval_2009 ]
The consent form used with subjects for whom a dental Age Assessment is also included.
Is it legal to expose subjects to ionising radiation for non-clinical purposes?
Under UK legislation it is permissible to expose radiographs for examinations not involving clinical care. The most common example is to enable the preparation of Medico-Legal reports to assess damages following an accident where the quantum of financial compensation is related to the severity of the damage. This is covered in a newsletter of the British Association of Forensic Odontology [Marsden 2011]. The process of exposing subjects to ionising radiation to obtain an image of the teeth for age assessment purposes is a legal procedure.
Is the UK reference dataset sufficiently large to provide age estimates?
This reference dataset has been created from over 2,700 cases of UK Caucasians and over 850 Afro – Trinidadian subjects. This is the largest dataset with specific ethnic derivation in the world. [see Roberts and Petrie 2011].
What system of assessing teeth has been used in the DAA database?
The first part of the Demirjian age assessment method has been used. This is important as there is wide spread support for the use of the very clear and easily identifiable tooth Development Stages [Roberts and Petrie 2008]. A consequence of this simple approach is that the same dentist assessing Tooth development Stages on 2 separate occasions and a different dentist assessing the same developing teeth have a very index of accurate reproducibility. The most important reference to support this is Olze et al 2005.
This high level of reproducibility is shown in our own work [Mitchell et al 2009; Yadava et al 2011; Chudasama et al 2012]. This issue of high level of reproducibility is important because it underpins the accuracy of age assessment using the Tooth Development Stages.
(Please note that the 2nd part of the Demirjian DAA system is NOT used because there is considerable concern about the statistical handling of the primary data). Many papers relating to use of the Demirjian system on different populations [s]how that there is a large discrepancy between the Chronological Age and the estimated Dental Age [Jayaraman J et al 2011] This is a common finding in the many papers that have used the Demirjian system in its entirety in population groups with a different ethnic background from the original French Canadian children studied by Demirjian in the late 1960s.
How is the data from Tooth Development Stages used to calculate a Dental Age?
The system used is based on the Average Age of Attainment of each of 256 Tooth Development Stages. For a given case of unknown age the Tooth development Stages present are identified and the information in the reference Dataset is extracted. For example, a case is presented with a Lower Left Second Permanent Molar (LL7) stage G = 14.31 years. LL8 Stage D = 14.47 years. These are added and divided by 2 to give the age of 14.39 years. This principle is applied at any age, the maximum number of developing teeth is 16 and the minimum number is 1.
It is important to understand that only teeth are still developing are used, that is Stages A to G. Once a tooth attains Stage H it is not possible to use in growing children because here is no upper boundary for the age of attainment for Stage H. An example of how this may mislead assessors is that a 12 year old child may have UL5 Stage G as 12.6 years and LL6 Stage Has 10.30 years. This would give a misleading but small error when averaging these 2 values. If the LL8 Stage G is present the average age will be 17.2 years yet the LL6 age is the same as that at the time of attaining the stage and is at 10.3 years. It can be seen by combining these 2 that the error increases dramatically. For these reasons any teeth which have reached Stage H are excluded from the analysis.
There is one exception to this exclusion rule which is the attainment of Stage H in 3rd permanent molars which will be covered later.
How accurate is this dataset in estimating the age of UK Caucasian subjects?
Publications from the Dental Age Assessment Team at King’s College Dental Institute show relatively small differences between the average Chronological Age (the Gold Standard) of study samples and the average Dental Age of the same subjects. It is helpful to focus on the publication which is closest to the 18 year threshold to illustrate this, which is the study at the 16 year threshold (Mitchell et al 2009). From this study it is possible to see that on average there was an under estimate of 0.27 years for females and 0.23 years for males.
At the 18 year threshold, for a subject who presents with a completely formed (mature) Lower Left 3rd Molar, it is possible to estimate the probability that the subject is over or under 18 years old.
Readers are referred to the DAA dataset of UK Caucasians (Roberts and Petrie 2011)
[efile 2011_DAA Paper08Appendix06_Table 06 Full Summary Data]
Is it possible to estimate the probability of a subject with Stage H being over 18 years old?
Yes, it is possible to calculate this probability using a simple routine in Microsoft Excel in the functions dialogue – this is the NORMDIST function. It is necessary only to use the Lower Left Third Molar.
In this file, 2011_DAA Paper08Appendix06_Table 06 Full Summary Data on page 3, is the data for the Lower Left Third Molar Stage H for females – LL8Hf. This shows a value of n-tds = 75, x-tds = 18.89, and sd-tds = 2.31. This gives a probability of being less than18 years of 0.3500 or in percentage terms 35%. The corollary to this is that probability of the subject being over 18 years old is 0.6500 or in percentage terms 65%.
Also on this page is the data for males. For LL8Hm the values are n-tds = 61, x-tds = 19.04, and sd-tds = 2.36. This gives a probability of being under 18 years old of 0.3297 or 32.97%. The corollary to this is that the probability of a male subject being over 18 years old is 0.6703 or 67.03%.
The level of proof required is on the balance of probability. It is clear that the probability estimates for subjects for whom dental development of the third molar is complete is well above the threshold of 50% for being ‘Over 18 years old’.
Other investigators have also provided data from different ethnic groups. Of note is the work from the German team that has provides the following probabilities for a subject with Stage being over 18 years.
We are sometimes asked why only one third molar is used?
First, the upper 3rd molars complete development at an earlier time than lower 3rd molars (Boonpitakthasit et al 2011). This would therefore lead to an artificial lowering of the estimated age if UL8H and LL8H were combined.
Second, it is inappropriate to combine data from left and right when the developmental stage is the same. This artificially increases the apparent total N (for subjects) and is inappropriate.
(there is an exception in DAA cases where the Tooth Development Stages on the left and right are different as occasionally happens in individuals. In such a circumstance it is usually helpful to combine the two stages as it results in a younger (below 18years of age) and gives the subject the benefit of the doubt.
[ efile 2011_DAA_Paper08A_Practical_GuidePDF pages 37 and 38 ]
Dental Age Assessment using Tooth Development Stages (TDSs) discernible on a Dental Panoramic Tomograph is a legal, simple and reliable method of estimating the age of subjects without birth records.
A unique feature of the work we have cited is that the Reference Data Set has been independently validated by deriving an independent study sample of radiographs separate from the Reference Data Set Radiographs. The Gold Standard of Chronological Age is compared with the estimated Dental Age. This approach has been used in the papers by Mitchel et al 2009, Yadava et al 2011, and Chudasama et al 2012.
Boonpitaksathit T, Hunt N, Roberts GJ, Petrie A, Lucas VS. Dental Age Assessment of adolescents and emerging adults in United Kingdom Caucasian using censored data for stage H of third molar roots. European Journal of Orthodontics 2011; 33: 503-508.
Chudasama P N, Roberts GJ, Lucas VS. Dental age assessment (DAA): s study of a Caucasian population at the 13 years threshold. Journal of Forensic and Legal Medicine 2012;19:22-28. DOI:10.1016/j.jflm.2011.09.008.
Jayaraman J, King NM, Roberts GJ, Wong HM. Dental age assessment: are Demirjian’s standards appropriate for southern Chinese children. Journal of Forensic Odontostomatology. 2011; 29: 22-28.
Jayaraman J, Roberts GJ, King NM, Wong HM. Dental Age Assessment of Southern Chinese using the United Kingdom Caucasian Reference Data Set. Forensic Science International. Epub September 2011 & In Press.
Lucas VS, Souberbielle A, Roberts GJ. Dental Age Assessment: Practical application of Meta-Analysis to Age Estimation in Children, Adolescents and Emerging Adults. 7th International Conference on Forensic Inference and Statistics. Lausanne, Switzerland 20-23 August 2008.
Mitchell JC, Roberts GJ, Donaldson NA, Lucas VS. Dental age assessment (DAA): Reference data for British caucasians at the 16 year threshold. Forensic Science International 2009;189:19-23. DOI:10.1016/jforsciint.2009.04.002.
Olze A, Bilang D, Schmidt S, Wernecke K-D, Geserick G, Schmeling A. Validation of common classification systems for assessing mineralization of third molars. International Journal of Legal Medicine 2004;119:22-26. DOI:10.1007/s00414-004-0489-5.
Olze A, Schmeling A, Taniguchi M, Maeda H, van Niekerk P, Wernecke K-D, Geserick G. Forensic age estimation in living subjects: the ethnic factor in wisdom tooth mineralization. International Journal of Legal Medicine. 2010; 124: 119-123.
Roberts GJ, Petrie A. 2011. DAA Paper08 Dental Age Assessment: a practical approach.
and DAA_Paper08_appendix06_Table – comprising the full data set.
Roberts GJ, Petrie A. Dental Age Assessment: A Practical Approach. Chapter 11 in Digital Forensic Science. IGI Global USA. 25 March 2011.
Yadava M, Roberts GJ, Lucas VS. Dental age assessment (DAA): reference data for British children at the 10- year- old threshold. International Journal of Legal Medicine 2011;125:651-657. DOI 10.1007/s004-010-0491-z.
Zeng DL, Wu ZL, Cui MY Chronological age estimation of third molar mineralization of Han in southern China. International Journal of Legal Medicine 2010; 124: 119-123.