Sustained effect of an intervention to limit ordering of emergency department lumbosacral spine films

Attempts at modification of test-ordering practices among housestaff demonstrate reversion to prior behavior after relaxation of surveillance. To test the hypothesis that use of pre-specified criteria for obtaining emergency department (ED) lumbosacral spine (LSS) films would reduce LSS utilization, we designed a cross-sectional observational study, with 1-year follow-up. The primary endpoint was estimate of efficacy, expressed as a proportionate change in LSS films from 1982 to 1992, adjusted for ED volume. The secondary endpoint was estimate of safety of this protocol. We found a 28% proportionate decrease in LSS films from 1982 to 1992. Among 520 patients with back complaints not meeting criteria for films, 4 had serious back pathology on follow-up that may have been detected on the index ED visit if LSS films had been ordered. All four were HIV(+), three of whom initially denied, then subsequently admitted, to active intravenous drug use. We conclude that use of a simple requisition form is associated with a quantitatively and statistically significant sustained reduction in utilization of ED LSS radiography. This protocol appears to provide a safe, cost-efficient means of limiting LSS film ordering by housestaff, with the possible exception of HIV(+) individuals. Because the association between HIV and radiographic abnormalities may be confounded by intravenous drug use, we cannot determine, on the basis of our data, whether HIV(+) status should be considered an independent criterion for obtaining LSS films.