RMTD-02 an Analysis of the Evidence Underpinning the Neurosurgical Oncology Clinical Practice Guidelines

To standardize treatment and improve patient outcomes for neurological conditions, the Congress of Neurological Surgeons (CNS) created evidence-based clinical practice guidelines (CPGs). CPGs rely on systematic reviews (SR), considered the gold standard for evidence, along with meta-analyses. However, previous studies illustrated SRs and meta-analyses informing CPGs in various specialties may be of sub-optimal quality when judged by rigorously by PRISMA (Preferred Reporting Instrument for Systematic Reviews and Meta-Analyses). Given such analysis has not been performed for neurosurgical oncology, this investigation evaluated the quality of the SRs and meta-analyses involved in neurosurgical oncology CPGs. The CNS website was screened for active neurosurgical oncology CPGs. All SR and meta-analyses cited under each CPG were extracted. SRs whose methods did not meet the PRISMA-P definition of a SR were excluded. We evaluated the eligible SRs and meta-analyses using the PRISMA checklist. Multiple linear regression analysis was performed using year of publication, intervention type, conflict of interest (COI) statement, and study design as covariates. Statistics performed with Python version 3.11.9. We identified seven neurosurgical oncology CPGs, of which six included eligible SRs/meta-analyses. Out of 2,763 citations, 20 were SRs (0.72%). The median [IQR] % SRs for each CPG was 0.95 [0.22, 1.12]. PRISMA scores ranged from 37 to 79.6, with a median of 54.6 [43.5, 75.9]. Multiple regression analysis revealed stated conflict of interest (COI), publication after 2010, and non-intervention studies were associated with increased PRISMA by 14 (p=0.016), 13.3 (p=0.028), and 12.9 (p=0.049) points respectively. SRs/meta-analyses are infrequently included in neurosurgical oncology CPGs, and their quality as measured by PRISMA is variable. Detailing COI is paramount for increased SR PRISMA score, as SR/meta-analysis authors can control this factor. Furthermore, results suggest recent SR/meta-analyses improvement and opportunity for intervention studies to glean insight from the SR/meta-analyses quality of non-intervention studies.