A scoring system for use in the diagnosis of acute abdominal pain in childhood

ABSTRACT

A statistical analysis was undertaken of 1158 children admitted to a surgical ward for the management of acute abdominal pain. Over two-thirds (40 percent) of the children had non-specific abdominal pain while 29.7 percent had appendicitis. The remainder were found to have had urinary tract infections (11.7 percent), constipation (7.5 percent), gastroenteritis (5.8 percent) or intussusception (5.3 percent). A stepwise discriminant analysis of the data collected during their evaluation was performed, using the BMDP statistical software package. Demographic and clinical features, as well as the results of ancillary investigations, were included in the data. The programme generated a classification function of a sub-set of 18 variables which best discriminated among the diagnostic groups. The coefficients of the classification functions were then combined with the rank order of selection of the variables to derive a scoring method for predicting the diagnosis. The results of urine culture were excluded since these would be unavailable during early clinical assessment. The scores for the diagnostic groups fell within the following ranges - 1 - 23 non-specific abdominal pain; 20 - 48 appendicitis; 35 - 84 gastroenteritis; 75 - 88 constipation and 89 - 140 intussusception. It is suggested that this scoring method be evaluated by a prospective study to test its validity (AU)