Computed tomography (CT) scans of the brain commonly are obtained in older patients to search for causes of syncope and altered mental status and to identify stroke. They often are used to protect against litigation (e.g., when patients request them for evaluation of apparently benign conditions). However, the diagnostic yield of CT scans is low, and the cost is high. Obtaining unnecessary CT scans is likely to increase as the population ages. In the older population, concerns about adverse reactions caused by contrast materials, transport accidents, delirium, and agitation may further tip the cost-benefit analysis toward forgoing CT scans. Hirano and colleagues conducted a prospective study to determine the frequency of CT scanning in an older cohort, the number of abnormal findings detected, and the impact these findings had on patient care.

The cohort was drawn from a previous study of hospitalized patients 70 years and older. Those who had a CT scan at admission were included, and those who were unable to communicate because of debilitating medical conditions or language barrier were excluded. Indications for the CT scans of 117 patients were abstracted, and abnormal findings were recorded. Medical records were examined to identify the course of treatment and to determine whether neurologic symptoms had been present at the time the CT scans were obtained.

Complications occurring as the result of CT scans also were compiled, as were the use of sedatives and restraints and the presence of anxiety or agitation during or around the scanning period. Study outcomes included positive CT scan results and positive results leading to a change in management strategy. The authors also looked at demographic and clinical factors that predicted a positive CT scan finding.

The average age of the 117 patients in this study was 80 years. Of the 96 patients (82.1 percent) who underwent a neurologic examination before scanning, 15 (15.6 percent) were found to have new focal neurologic deficits. Only seven of 29 CT scans with evidence of significant abnormalities (6 percent of all scans) were determined to be true positives. Of the total true-positive and incidental scans, 10 scans led to management changes, which means slightly more than one third of all positive scans and less than one tenth of all scans resulted in a change of treatment.

New focal neurologic deficits were the only significant predictor of treatment changes (odds ratio = 13.2). Mental status change did not predict treatment change, but 13 scans (11.1 percent of all scans) were ordered specifically because of mental status changes, and 35 scans (29.9 percent) were ordered in the presence of mental status changes noted in the chart. There were no adverse reactions to contrast materials, but there were several instances of sedation or restraint requirements and agitation and positioning problems, including three unreadable CT scans because of motion artifact.

In this prospective cohort study, 6.0 percent of scans had true-positive findings, and 8.5 percent led to treatment changes (i.e., the number needed to scan to alter treatment was 12). Only new focal neurologic deficits predicted management changes. These findings suggest that although older patients have conditions for which brain imaging is ordered, the yield of brain CT is low.

The authors conclude that better clinical criteria, such as a focus on neurologic deficits, should be defined to guide decisions about which patients really need CT scans.