Utility of Administrative Databases and Big Data on Understanding Glioma Treatment—A Systematic Review

Abstract

Abstract Background Gliomas are a heterogeneous group of tumors where large multicenter clinical and genetic studies have become increasingly popular in their understanding. We reviewed and analyzed the findings from large databases in gliomas, seeking to understand clinically relevant information. Methods A systematic review was performed for gliomas studied using large administrative databases up to January 2020 (e.g., National Inpatient Sample [NIS], National Surgical Quality Improvement Program [NSQIP], and Surveillance, Epidemiology, and End Results Program [SEER], National Cancer Database [NCDB], and others). Results Out of 390 screened studies, 122 were analyzed. Studies included a wide range of gliomas including low- and high-grade gliomas. The SEER database (n = 83) was the most used database followed by NCDB (n = 28). The most common pathologies included glioblastoma multiforme (GBM) (n = 67), with the next category including mixes of grades II to IV glioma (n = 31). Common study themes involved evaluation of descriptive epidemiological trends, prognostic factors, comparison of different pathologies, and evaluation of outcome trends over time. Persistent health care disparities in patient outcomes were frequently seen depending on race, marital status, insurance status, hospital volume, and location, which did not change over time. Most studies showed improvement in survival because of advances in surgical and adjuvant treatments. Conclusions This study helps summarize the use of clinical administrative databases in gliomas research, informing on socioeconomic issues, surgical outcomes, and adjuvant treatments over time on a national level. Large databases allow for some study questions that would not be possible with single institution data; however, limitations remain in data curation, analysis, and reporting methods.