INTRODUCTION: The diagnosis of shunt malfunction can be difficult even for the experienced clinician and may lead to disastrous circumstances when misinterpreted. Less experienced physicians may rely more on radiographic reports as a primary diagnostic modality. In this study, we evaluated the reliability of using these reports without accurate clinical assessment. METHODS: All shunt revisions seen at Children's Hospital (Birmingham, AL) between January 1996 and August 1996 were reviewed, excluding patients with brain tumors, supratentorial extraaxial fluid collections, and infections. Sixty-eight patients underwent 100 operations for shunt malfunction. All patients had evidence of shunt blockage, disconnection, catheter malposition, or valve pressure incompatibility. The prospective radiographic interpretation of preoperative computed tomography and magnetic resonance imaging scans was reviewed in each case. RESULTS: Twenty-four percent of the reports made no mention of shunt malfunction. In this group, the ventricular system was described as "unchanged," "stable," "normal," "unremarkable," "small," "smaller," "slit," "negative," and "no hydrocephalus," with no other comment to support a diagnosis of shunt malfunction. An additional 9% of reports contained the same terms, while also hinting at some other clinical or radiographic data that suggest the possibility of shunt failure (e.g., a shunt disconnection seen on plain radiographs), despite the scan findings. In all patients in this group, symptoms improved after surgery. CONCLUSION: We conclude that as many as one third of patients presenting with shunt malfunction will not have the diagnosis of shunt malfunction supported by a prospective radiologic interpretation of brain imaging. Although the neurosurgical community can assess the clinical situation to determine the need for surgery, other clinicians can be easily reassured by a radiographic report that does not mention or diagnose shunt malfunction. Today, more than ever, nonneurosurgeons are being called on to evaluate complex clinical situations and may rely on radiographic reports.