Development, Implementation, and Use of a Local and Global Clinical Registry for Neurosurgery

Physicians are being challenged to obtain data for outcomes research and measures of quality practice in medicine. We developed a prospective data collection system (registry) that provides data points across all elements of a neurosurgical stereotactic radiosurgery practice. The registry architecture is scalable and suitable for any aspect of neurosurgical practice. Our purpose was to outline the challenges in creating systems for high quality data acquisition and describe experiences in initial testing and use. Over a two year period, a multicenter team working with software engineers developed a comprehensive radiosurgery registry based on a MS-Sequel® server platform. Three neurosurgeons at one center were responsible for final editing. Alpha testing began in September 2012 and server-based beta testing began in February 2013. The major elements included demographics, disease-based items (47 categories for different brain tumors, vascular malformations, and functional disorders) with relevant clinical grading systems, treatment-based items (imaging, physics, clinical), and follow-up data (clinical, imaging, subsequent therapeutics). Nine hundred patients were entered into the registry at one test center, with new entries and follow-up data entered daily at the point of contact. With experience, the mean time for one new entry was 6 minutes. Mean time for one follow-up entry was 45 seconds. The system was made secure for individual use and amenable for both data entry and research. Analytics used different filters to create customized outcomes charts as selected by the user (e.g., survival, neurologic function, complications). A local or multicenter prospective data collection registry was created for use across 47 clinical indications for stereotactic cranial radiosurgery. Further refinement of fields and logic is ongoing. The system is reliable, robust, and allows use of rapid analytical tools. Large medical registries will become widely used for collection and analysis of large data sets and should have broad applicability to many other elements of neurosurgical and medical practice.