P17.18.B FAILURE OF PERSUASION: WHEN PRACTICE GUIDELINES DON’T CHANGE PRACTICE IN NEURO-ONCOLOGY

Abstract BACKGROUND Five randomized controlled trials (RCTs) of anticonvulsant prophylaxis, conducted over the last 40 years, have each failed to demonstrate a reduction in the frequency of first seizures in patients with primary or metastatic brain tumors, and practice guidelines published by the major neurology, neurosurgery, and oncology professional organizations recommend against this practice. Nevertheless, physician prescribing preferences and actual treatment practice have remained largely unchanged, subjecting patients to unwarranted potential toxicity. We explore this phenomenon, its explanations, its wider implications for neuro-oncologists, and potential remedies. MATERIAL AND METHODS A PRISMA-compliant literature search was conducted to identify all studies comparing anticonvulsant prophylaxis to no prophylaxis for seizure occurrence and survival in patients with brain tumors. Studies describing physician beliefs and practices were also identified. Relevant data was extracted from all studies fulfilling a priori inclusion and exclusion criteria. A level of evidence was assigned to each study using AAN/AANS/CNS criteria, and meta-analytic summary statistics were calculated using a random effects model and inverse variance technique. Further analyses were conducted using linear regression and ANOVA. RESULTS The pooled outcome from the five identified RCTs (441 patients) did not support the use of prophylactic anticonvulsant medication for primary prevention (RR 0.95 [0.58-1.55], p=0.85). Risk of bias was strongly associated with study outcome: lower quality RCTs were more likely to have outcomes suggesting a benefit (R2=0.83, p=0.032). Neither physician attitude toward prophylaxis (mean difference 0.2% [-43.6, 44.03], p=0.99, 4 studies) nor actual practice (mean difference 15.5% [-33.8 to 2.78], p=0.06, 6 studies) changed significantly between pre- vs. post-guideline publication. CONCLUSION Despite aggressive broadcasting of RCT data and corresponding practice guidelines, and a strong commitment to evidence-based practice by EANO and SNO, anticonvulsant prophylaxis remains the prevalent standard of care for seizure-naive brain tumor patients. Additional research to establish whether this disturbing paradox persists for other critical practices is essential. Individualized strategies to facilitate guideline adherence using established techniques for behavior change - reducing friction, piggybacking, and strengthening cues - should be published together with all future practice guidelines.