0096
DATA MINING THE CAPFSA RED CROSS CHILDREN’S
HOSPITAL DATABASE 1991-2000 AS PART OFF INJURY PREVENTION PLANNING AB van As, G Dragosavac#, N du Toit *,AJW Millar, H Rode Trauma Unit, Department of
Paediatric Surgery, Institute of Child Health, Red Cross War Memorial
Children’s Hospital, University of Cape Town, South Africa, *Child Accident Prevention Foundation of Southern
Africa #Mine-Max Limited Introduction Trauma is the leading cause of morbidity
and mortality in children across the globe and a huge public health problem
in South Africa. The aim of the Child Accident Prevention
Foundation of Southern Africa (CAPFSA) is to reduce all intentional and
un-intentional injuries of all severity through research, education,
environmental change and to produce recommendations for legislation. Since
1991 we have been capturing data from all injured children treated at the
Red Cross Children’s Hospital for trauma in our database. There are
currently 88822 patients entered in the database. Aims/Objectives To study patterns and trends from all trauma patients
presenting to Red Cross Children’s Hospital with the ultimate goal of
accident prevention. Materials/Methods We explored our database with novel advanced data-mining
methods (visualisation, decision trees and case-based reasoning, supplied
by Mine-Max Ltd) Results The most common injuries were falls (n = 32766). 21,2%
(n = 6953) fell from the bed, while 11.1% (n = 3.638) fell from playground
equipment. From all seriously injured children 39,4% fell from the bed,
while only 2.8% fell from playground equipment. The second most common cause was traffic accidents (n =
11915). Motor vehicle accidents involving a pedestrian were the most common
with 64.2% (n = 7645). Bicycle accidents were second with 15.3% (n = 1819),
passenger motor vehicle accidents third with 13,7% (n = 1630). Passenger-minibus accidents and motorcycle accidents
were relatively rare (4,45 respectively 0,5%). 7241 patients presented with burns, of which 72,1% were
fluid burns, 11,7% were due to flame burns. From the 54 patients presenting
unconscious, 44,4% were due to flame burns. 3.302 patients presented after assault. 51,7% (n = 1709)
after blunt assault, 22,3% (n = 735) were raped, 13,1% (n = 433) presented
after sharp assault, while 1,2% (n = 40) were attacked with human bites.
Detailed data-mining results will be discussed during the presentation. A brief overview will be given on the various active and
passive prevention strategies, that are used by CAPFSA in an attempt to
prevent childhood injuries. Conclusions Data-mining is a powerful and promising
tool in the third millennium. Close co-operation between information
technologists, data-collecting clinicians and injury prevention workers
could prove to be an exciting pathway in the future of injury control.