Age estimations of adolescent refugees

Key message

Many western countries deal with the arrival of a considerable number of refugees on a yearly basis.  The age of these asylum seekers is in many cases uncertain.  Many European countries have therefore initiated age estimations to predict the actual age of refugees when there is reasonable doubt about the applicant’s claimed age.  In Norway age estimations of adolescent refugees have been carried out, on behalf of the Norwegian Directorate of Immigration (UDI) since February 2003.  Two methods are currently used; the assessment of bone age (x-rays of the left hand and wrist) performed at the Paediatric Radiological Department at the Ullevål University Hospital, and the assessment of dental age (dental examination including an orthopantomogram (OPG), a panoramic x-ray image of the teeth) performed at the Institute for Clinical Odontology at the University of Oslo.  The accuracy and validity of these methods have been questioned.  The Directorate for Health and Social Affairs asked to conduct a Health Technology Assessment (HTA) of these methods for age estimation of refugees.

The main objective of this HTA was to evaluate the reliability of the two methods used for age estimation in Norway in order to assess whether an asylum seeker with unknown age is above or below 18 years of age.  In addition we evaluated the accuracy of the methods for the purpose of estimating chronological age of individuals in the age range between 16 – 20 years.

The articles that met the inclusion criteria showed:

• (in most cases) a significant advanced skeletal age compared with the G&P standard.
• A mean standard deviation (calculated from all populations) of 11,8 months for girls and 14,8 months for boys.
• That is almost the same as the standard deviations that are used in Norway (12 months for girls, 14 months for boys).
• A significant differences between the (combined) Caucasian population and several populations of different ethnic groups.
• The skeletal maturation can be both accelerated and retarded compared to the Caucasian population.

Factors like study design, purpose of taken X-ray images, number of observers and socio-economic status might also have influenced these differences.

• The dental methods used in Norway (Kullman, Haavikko, and Nortje) are not widely used in age estimation research.
• The Demirjian method was the method of choice in most publications.
• The heterogeneity of methods, study designs, outcomes and presentation of results, made it impossible to carry out statistics on the data.
• There were no obvious differences between sexes, age groups, or populations that could not also be a result of the heterogeneity of the studies.
• The median standard deviation of the Demirjian method was around 2.2 years.
• The medians of the few other methods were in general lower.
• Probabilities of determining whether an individual is above or below 18 years old differ slightly between males and females and the teeth that are assessed.

In general, third molars from the lower jaw lead to higher age estimates than those from the upper jaw.  Females also have a slightly higher probability of having an estimated age of above 18 years when they have fully developed third molars.

• The sensitivity of the Greulich & Pyle method is acceptable but it will never reach a specificity high enough to be acceptable (maximum specificity of about 80%).
• The dental methods in contrary can produce a high specificity but that results in a very low sensitivity.
• Combining both methods increases specificity but as a result sensitivity decreases.

Unfortunately only one article presented sensitivity and specificity for both skeletal and dental method combined.

• Using the methods as they are used today, someone who has been determined to be 19 years of age by both skeletal and dental