Abstract

Aims: To determine whether abused and non-abused children differ in the extent and pattern of bruising, and whether any differences which exist are sufficiently great to develop a score to assist in the diagnosis of abuse.

Methods: Total length of bruising in 12 areas of the body was determined in 133 physically abused and 189 control children aged 1–14 years.

Results: Our method of recording bruises by site, maximum dimension, and shape was easy to use. There were clear differences between cases and controls in the total length of bruises. These differences were at their greatest in the head and neck and were less notable in the limbs. A scoring system was developed using logistic regression analysis using total lengths of bruising in five regions of the body. Good discrimination between the two sets of children was achieved using this score; by including a variable that indicates whether a bruise had a recognisable shape the discrimination could be made even better. Given a prior probability of abuse the score can be used to give posterior odds of abuse, given a particular bruising pattern.

Conclusions: The scoring system provides a measure that discriminates between abused and non-abused children, which should be straightforward to implement, though the results must be interpreted carefully. We do not see this score as replacing the complex qualitative analysis of the diagnosis of abuse. This clearly includes history as well as examination, but rather as the beginning of the development of an important aid in this process.