Assessing the applicability of the Greulich and Pyle (1959) skeletal age estimation method to South African children between 0 and 13 years

The assessment of skeletal age (SA) in both clinical and forensic settings serves to estimate the degree of development and biological maturation of a child. This is a suitable and frequently used technique, as skeletal age is a valuable indicator which can be analysed from birth until the attainment of the fully-developed adult body form. Radiographs of the hand-wrist region are most often used to estimate skeletal age, by comparing a sample radiograph against an established standard. The most internationally accepted method for skeletal age estimation is the Atlas method of Greulich and Pyle (GP) (1959). Extensive research has been conducted regarding the flexibility and applicability of this method to other race and socio-economic groups, with most research indicating that the GP Atlas is not a suitable method for skeletal age estimation for other world populations. As little research in this regard has been done using South African samples, this project aims to determine whether the GP Atlas is still an applicable method of skeletal age estimation for South African children between the ages of 0 and 13 years. Skeletal age estimates using the GP Atlas were performed on 1356 (821 males, 535 females) pre-existing digital radiographs from Red Cross War Memorial Children's Hospital, Salt River Mortuary and Tygerberg Mortuary. The target age range for this sample was birth (0) to 13 years, in order to incorporate the appearance of the ossification centres of the hand and wrist, but exclude epiphyseal fusion in this region. Although there was a strong positive correlation between skeletal age (SA) and chronological age (CA) [Males: R = 0.96; Females: R = 0.97], there was a high degree of variation in the sample, with overestimation and underestimation of skeletal age across both sexes and all age groups. This variation increased with an increase in age. Regression analysis, using a best-fit fixed-effect model, random-intercept mixed-effect model and a random intercept-random slope mixed-effect model, attempted to account for the variation in the sample by using skeletal age as a predictor for chronological age. The random intercept-random slope mixed effect model was shown to be the most appropriate model for this sample. • vii • Variables for sex, race and socio-economic status were included in the regression analysis to determine their effect on predicted chronological age. The regression model indicated that sex (and the interaction between sex and skeletal age) influenced predicted chronological age by slightly reducing the sample variation, but the sample was too variable for the effects of race and socio-economic status to be determined. The model also indicated that the majority of the variation within the sample was due to random variation. Possible reasons for this include the differences in biological origin between the South African and GP samples, the GP methodology, and the timing of growth and development of South African children on the whole. Due to the high degree of random variation in this sample, which cannot be accounted for completely by sex, race or socio-economic status, the GP Atlas is not reliable for skeletal age estimation of South African children. Other methods, such as the Phillips (2008) method for dental maturation, should be sought as an alternative when the chronological age of a child comes into question.

Grantor

Department of Human Biology

Year dc:date.issued

2012

Author dc:creator

• Speed, Belinda

Advisor dc:contributor.advisor

• Morris, Alan.G.

• Anatomy

Handle dc:identifier.uri

http://hdl.handle.net/11427/39150

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/39150