UCL Great Ormond Street Institute of Child Health


Great Ormond Street Institute of Child Health

Prof Tim Cole

Prof Tim Cole

UCL GOS Institute of Child Health

Faculty of Pop Health Sciences

Joined UCL
1st Apr 2002

Research summary

Tim Cole is currently working a several projects relating to the statistics of growth and development.  Here is a recent article summarising his work over the past 25 years.

His SITAR method of growth curve analysis (Int J Epidemiol 2010;39:1558-66) summarises cohort growth curves in terms of an average curve, defined as a natural cubic spline, plus a set of three random effects per subject that define how each subject's growth curve differs from the average. The three random effects define the size (mean centile of the growth curve), tempo (timing of a developmental marker such as age at peak height velocity) and velocity (the relative growth rate) of each subject's curve. The SITAR analysis typically explains 99% of the variance in height during puberty, 95% of weight in infancy, 80% of BMI in childhood and smaller percentages for hormones and other outcomes. This means that the three SITAR parameters provide a comprehensive yet parsimonious summary of how individual children grow. Tim is currently funded half-time by the MRC for 3 years to work on the SITAR method, developing the methodology, applying it to several birth cohorts, and writing an R package library for open access dissemination of the method.

Tim Cole is also involved in designing a new set of growth charts for the UK, working with a growth chart expert group sponsored by the Royal College of Paediatrics and Child Health chaired by Professor Charlotte Wright (Glasgow University). The aim is to extend the 0-4 years UK-WHO charts published in 2009 to include puberty, and introducing novel features for the assessment of pubertal growth and development. 

A third strand is improving methodology for child obesity intervention trials, where there is uncertainty about the best choice of outcome measure, i.e. change in raw BMI or change in BMI z-score. Work with Professor Russell Viner and Professor Catherine Law suggests that BMI z-score is a valid outcome measure, which reflects a change of view since Cole's recommendation in 2005 that raw BMI change is the best outcome measure to use (Eur J Clin Nutr 2005;59:419-25).

Forensic age assessment is another area in which Tim Cole has recently become involved. In 2011 he provided expert witness evidence to more than 10 age assessment court cases in Australia involving Indonesian fishermen accused of people smuggling. The men said they were under 18 years old yet they had a mature hand-wrist x-ray, which the State argued was evidence that there were over 18. (Being under 18 they would be repatriated, but over 18 there was a mandatory 5-year jail sentence for people smuggling.) Partly as a result of Cole's evidence nearly all the cases were dismissed and the men repatriated, often after many months in custody. Australia has now agreed not to use radiographic bone age assessment in such cases. 

However radiographic dental age assessment is another process which is still used in many parts of the world, and the statistical basis for it is even more flawed than for bone age. Cole is working with others in the UK and elsewhere to stop dental age assessment universally. 


University of Cambridge
DSC, Medical Statistics | 2002
University of Cambridge
PhD, Clinical Medicine | 1983
University of Cambridge
MA Cantab, Mechanical Engineering | 1975
University of Oxford
BPhil, Biomathematics | 1970
University of Cambridge
BA, Mechanical Engineering | 1968


Tim Cole has been professor of medical statistics at UCL Institute of Child Health since 1999. He originally studied engineering at Cambridge and statistics at Oxford, and has been employed by the Medical Research Council for his entire career since 1970, working first at the MRC Pneumoconiosis Research Unit in South Wales and then from 1975 at the MRC Dunn Nutrition Unit in Cambridge. Since 1999 he has been an external scientific staff member with the MRC. His research interests cover many aspects of child growth assessment, including growth chart construction (the LMS method), growth curve analysis (the SITAR method), factors relating early growth to later outcome, the IOTF child obesity cut-offs (cited over 5000 times to date), body size scaling and forensic age assessment. He has published over 450 peer-reviewed papers and his h-index stands at 99. In 2006/07 he was appointed Honorary Fellow of the Royal College of Paediatrics and Child Health, and also Fellow of the Academy of Medical Sciences. In 2016 he was awarded the Rank Prize for Nutrition, and also the Royal Statistical Society's Bradford Hill Medal.

He has also been heavily involved in peer review, both as a statistical and subject reviewer, acting as referee for over 50 journals and a statistical reviewer (latterly statistics editor) for the BMJ since 1991. He has also been a regular member of the BMJ primary care hanging committee and more recently the BMJ Manuscript Meeting. He is in charge of statistical review for the Archives of Disease in Childhood.