The reliability of digital images when used to assess burn wounds.

Sixty burn wounds were assessed in person. The same observer later assessed them using digital images of different sizes. The file sizes tested were 2.25, 5.5 and 9 MByte per image. There was good agreement between the diagnoses of burn depth made using the digital images and those made in person, with kappa scores of 0.53-0.60. There were no major differences between the three file sizes. The assessments made of the partial-thickness burns showed a lower rate of agreement between the in-person and the digital image assessments and for these burns the 2.25 MByte images were apparently as good or better than the larger images. There was little difference between the three file sizes in terms of observer confidence, usefulness of the location shot, or perceived image quality. There was no significant advantage in using larger file sizes to assess burn wounds.