The current Australian Federal Police practice of using X-rays of the hand/wrist to assess the age of Indonesian crew members of boats bringing refugees to Australia is based on a method developed in the 1950s. The technique is outdated, inaccurate and raises ethical concerns.
If the crew members, who claim to be under 18, are found to be adults, they must face mandatory five-year jail terms under people smuggling laws.
Hand/wrist X-ray examination was designed as a tool to assess the general physical development of children rather than to determine chronological age, and its degree of error is unknown but likely to be quite large.
Biological variation in human development means that any assessment of age based on analysis of anatomical growth markers is only an estimate, and will inevitably contain a degree of error.
While there’s no foolproof way to determine the exact age of an individual without verified birth date documentation, there are better methods available.
The size of inherent error in determining age depends, in large measure, on the estimation methods used so it behoves the government to use better tools in what is a determination with very serious ramifications.
The inherent fallibility of hand/wrist x-rays
Hand/wrist X-rays rely on an anatomical atlas of standard hand/wrist X-rays taken of a group of 1000 middle-class American children from birth to nineteen years in the 1930s and 1940s.
This atlas displays standard skeletal development at three-month intervals during the first year of life; six-month intervals from one to five years; and one year intervals thereafter, up to 19.
At no point does this technique claim to be able to determine chronological age; it was designed as a tool for health workers to better assess children’s skeletal development and overall growth.
Quite clearly, the major drawback when using this technique today is that individuals included in the atlas grew up on the other side of the world more than 60 years ago. The economic circumstances and general public health then aren’t comparable to what prevails now.
It’s clearly unlikely an American child in the 1940s developed at the same rate at which an Iraqi, Afghani, Indonesian or indeed an Australian child will develop today.
Even if the atlas accurately reflected current growth trends, it would still be of limited use in predicting chronological age.
The variability inherent in the model means that for a boy predicted to be 17 years old, there is a standard deviation of 15.4 months. This means that the boy’s age could have the possible range of 17 plus or minus two-and-a-half years. He could therefore, be aged anywhere from 14.5 to 19.5 years.
Plainly, if the question being asked is whether or not this person is an adult or still a minor, this method cannot possibly provide an accurate answer.
There are also medico-ethical and legal considerations involved in conducting radiological procedures – with no defined medical need – on living people.
These are yet to be adequately addressed by the government.
Comprehensive age assessment of living individuals necessarily involves the use of ionising radiation – X-rays – with unavoidable radiation exposure.
While this exposure is not at a level sufficient to cause immediate harm, it does raise the total lifetime dose of radiation experienced by the individual.
Medical imaging is continually advancing, and development of methods such as ultrasound, MRI, and more focused low-dose CT technology may, in the future, resolve some of these issues.
Unfortunately, the most accurate assessment of a person’s age requires examination of internal skeletal and dental structures and that can’t be done without X-rays.
If age estimation were limited, as some have suggested, to an external physical examination, then the potential for gross error in either over or under estimating age would be great.
But we still need a considered debate about the risks and ethics associated with the use of X-rays for non-medical purposes versus the benefits of more accurate age assessments in the interests of justice.
There are other available, proven methods for assessing age. These include physical examination by a suitably qualified medical practitioner, coupled with assessment of the degree of dental development and assessment of other skeletal elements such as the clavicle (collar bone), which has a far longer development time than the wrist, and can be used to assess age well into the 20s.
Still, many of these methods were developed from research conducted decades ago and on populations which geographically, environmentally, and nutritionally do not accurately reflect current population demographics.
Some of these systems, especially those describing dental development, have proven to be quite adequate for people under 15 years of age and have served the profession well.
But recent research suggests these older methods may not be as accurate as was once thought for young individuals, and it is suspected that they may be substantially more inaccurate for individuals older than 15 years of age.
In recent years, partly as a result of a report into the state of forensic science in the United States, there have been increasing calls for forensic science to improve the evidence basis for its conclusions.
For forensic anthropologists and odontologists, this includes improving age estimation methods, especially developing new reference standards for particular populations, and devising a more scientifically robust definition of the error and range in age estimations.
New research is now focusing on the development of multifactorial methods for age estimation, whereby a combination of methods provides the age estimate with the smallest possible error.
Techniques able to provide age estimates at least twice as precise as any that have been previously used have been developed for older teenage individuals.
The development of the third molar, for instance, probably provides the best way to discriminate between an adult and a child, and should be part of any age assessment procedure for late teenage individuals.
These new methods still have an inherent degree of error due to unavoidable development variability, but at least they are based on modern populations more closely resembling individuals who are most often targeted for age estimation.
And they are far more accurate than the outdated hand/wrist method currently in use, involve no greater radiation exposure, are based on modern population data, have a smaller error margin, and are just as simple to apply in practice.