Three studies were included in the review: two studies with a total of 73 children and 86 bruises, and one study of 32 bruises in an unspecified number of babies.
Relationship between colour and age of bruise.
One study of 50 children, recruited from a children's hospital emergency department (with children excluded when abuse was suspected or there was an underlying medical condition), reported an association between colour and bruise age (P<0.001): red/blue and purple colours were more commonly seen in bruises less than 48 hours old, and yellow, brown and green colours in bruises over 7 days old. However, red/blue and purple colours were also seen in 4 of 13 bruises over 7 days old, while yellow/brown or green colours were also seen in 9 of 39 bruises less than 48 hours old.
In a second photographic study (36 bruises in 23 white children in an orthopaedic ward), a single observer reported that blue, brown, grey and purple colours were not discriminatory, red colours occurred only in injuries 1 week old or less, yellow occurred only after the first day, and green was seen in 9 of 32 bruises more than 2 days old.
The third study (32 bruises), conducted in babies attending a health visitor clinic, found that yellow colours appeared only in bruises over 48 hours old.
Accuracy and inter-observer variability.
The authors of the emergency department study reported that inter-observer reliability for the assessment of bruising age, colour, tenderness, swelling, or abrasions was poor (kappa -0.03), and that there was poor (40%) overall accuracy in determining the exact age of bruises to within 24 hours. Accuracy was improved (76.2%) for broad categorisation (fresh, 48 hours or less; intermediate, 48 hours to 7 days; old, more than 7 days) by emergency physicians (n=16); the accuracy of trainees (n=39) was lower (52.7%).
The photographic study, using a single observer, produced a correct classification of fresh, intermediate or old bruises (using the same classification system as the emergency department study)in 24 of 44 cases.