INNV-33. IMPLEMENTATION OF ELECTRONIC DATA CAPTURE FOR USE IN NATURAL HISTORY STUDIES: UTILITY OF CENTER FOR CANCER RESEARCHS (CCR) LABMATRIX AND SCRIBE SYSTEMS FOR THE NEURO-ONCOLOGY BRANCH NATURAL HISTORY STUDY (NOB-NHS)

摘要

Accurate and prompt data collection is vital to clinical research. Traditionally, patient reported outcomes (PROs) have been captured through paper/pencil format and clinical data by retrospective clinical document review. Limitations include issues of errors, data omissions and monetary cost. We report an innovative tool for data collection. The primary objective of the NOB-NHS is to provide longitudinal collection of data including clinical status and PROs. Data was initially recorded by data analysts from review of the medical record in nine forms housed in CCRs Labmatrix system. All data fields were reviewed by the clinical team for relevance to the research question, redundancy, and timing of data collection. Free text fields were replaced by pre-determined, standardized pick lists of possible responses. Data fields were then converted into 2 electronic data collection forms (new patient and follow-up) in Scribe, a CCR-developed system for PROs/electronic form data. Feasibility and concordance testing was completed. Lastly, the clinical team was trained to enter data using tablets/laptops in the clinical setting. To date, 400 patients have been registered onto the NOB-NHS, with 222 collected through electronic systems. Links are sent via email to clinicians (nurse practitioners/clinical fellows) and patients that open a survey-style data collection form with drop-down menus and pre-formatted fields. Upon completion, the form automatically syncs to Labmatrix for data deposition and storage. On average, the Study Entry Form is completed in 5 minutes; the Clinic Visit Form in 2 minutes. PROs completion rate is > 95%. Data collected in a systematic and timely manner permits researchers to build stronger queries, reduces time spent on data cleaning, and facilitates statistical analysis. Standardization of the data fields and training of both clinical staff and data managers allows research to be reliable, especially for longitudinal studies such as natural history studies.