Effect of health care system factors on test ordering.

Abstract

To determine the effect of the emergency department (ED) environment and other health care system factors on test ordering for children with acute abdominal pain. We reviewed the encounter records of 1140 consecutive children seen in either the pediatric clinic or ED of an inner-city teaching hospital with a complaint of acute abdominal pain (< 72 hours). In the ED and the clinic, patients were seen by medical students, pediatric residents, and general pediatric faculty members. Measured data on test ordering included the number of tests ordered and the type of tests ordered; specifically examined were the throat culture, urinalysis or urine culture, and chest radiograph. Measured health care system factors included (1) encounter location; (2) resident involvement and level of training; (3) student involvement; and (4) faculty member's years of experience and sex. Of the 1140 children, 117 (10.2%) were seen in the ED, 531 (47.1%) were seen by a resident, 344 (30.2%) were seen by a medical student, and 195 (17.1%) were seen by a faculty member with more than 10 years of clinical pediatric experience. After controlling for initial signs and symptoms in multiple logistic regression, a child treated in the ED was no more likely to have had tests ordered than one who was treated in the clinic. Neither resident involvement nor resident training level affected test ordering. Except for decreasing the likelihood of having a urinalysis or urine culture ordered (odds ratio [OR] = 0.30; 95% confidence interval [CI], 0.15-0.63), student involvement did not affect test ordering. Also, except for decreasing the likelihood of having a throat culture ordered (OR = 0.45; 95% CI, 0.25-0.83), being seen by a pediatrician with more than 10 years of experience did not affect test ordering. Children seen by female physicians were more likely (OR = 2.41; 95% CI, 1.57-3.70) to have at least 1 test ordered. For children seen for a complaint of acute abdominal pain, we found little evidence that test ordering is affected by encounter location, resident involvement, student involvement, or faculty member experience.