A retrospective review of findings on postoperative cranial computed tomography and magnetic resonance imaging scans at Kenyatta National Hospital

Abstract

Background: Post-operative cranial scans are not necessarily obtained in all patients following neurosurgical procedures. The scans may also be obtained after varying periods of time. Appropriate utilization of the scans should lead to increased identification of the critical post-surgical findings that would influence subsequent patient management. This study was set to determine the situations in which postoperative imaging would be most beneficial in neurosurgical patients at Kenyatta National Hospital (KNH) Methods: This was a retrospective review of patients who underwent cranial surgery at KNH between June 2011 and May 2013 and followed-up for a maximum period of 6 months. Medical records were evaluated to reveal the clinical characteristics of the patients, presence and timing of post-operative scans, imaging findings and subsequent surgical management of the patients. The data was manually collected and thereafter analyzed using statistical computer software (SAS). The evaluation involved descriptive statistics, chi square statistics, Kaplan Meir curves and competing risk analysis. Results are presented in the form of frequency distributions and descriptive statistics. Results: The study found that53 (18%) of the 298 patients reviewed had scans within a median time of 20 days after cranial surgery. Patients that were admitted in the hospital for between 30-90 days had the highest chance of getting a postoperative scan. Forty percent of the scans were ordered for routine follow up and the rest (60%) due to neurological derangement in the patient. Majority (36; 68%) of the post operative scans had a positive finding. However, only 20 (38%) of the scans had findings that necessitated a subsequent surgical procedure. Furthermore, the findings that necessitated a subsequent surgical procedure were much more common among scans ordered due to neurological derangement as compared to scans ordered for routine post-operative follow-up (54% vs 14%, p=0.005). The chances of detecting a positive finding did not vary whether the scans were done within or after 30 days following surgery. Similarly, the post-operative time at which the scan was done (either <30, 30-60 or >60 days) did not influence the chances that the findings of the scan would necessitate subsequent surgical intervention. Conclusions: The proportion of neurosurgical patients that gets post-operative scans at KNH is acceptable because it is comparable to the proportion of patients that is expected to have the postoperative complications that we need to detect and manage. Increasing the proportion of patients who get routine scans is unlikely to optimize the role of the scans in managing the postoperative patients. Instead, patients at a higher risk of developing postoperative complications should be identified and targeted for imaging. There is no particular post-operative period during which scans are collectively better at yielding findings of surgical importance, thus the appropriate time for imaging should be guided by the indication for imaging.