Abstract BACKGROUND Modern-day clinical trials are exorbitantly expensive and time intensive but remain single use and single user. While conducting a prospective survey study investigating sociodemographic, psychological, and disease-related factors associated with uncertainty in patients with central nervous system and other malignancies, we developed a novel strategy for real-time data display and data sharing that could change the paradigm for publication of neuro-oncology clinical trials. MATERIAL AND METHODS Adult patients seen for routine surveillance and imaging visits after completing a planned course of cancer therapy were given detailed questionnaires to assess perceptions of their disease and the emotions attached to those perceptions. All survey and demographic results were stored in REDCap. During the phase-1 study, we took advantage of the REDCap API to create a tool for publishing and analyzing the data as it was being collected allowing investigators from different institutions to review and discuss the evolving dataset in real time. We’ve made the bespoke data-exploration tool available as part of this publication permitting readers to download the data for themselves to conduct their own analysis. RESULTS Two hundred twenty-two patients were enrolled (mean age, 61 years, 64% male). Most common histologic subtypes were primary brain tumor (22%), CNS lymphoma or leukemia (14%), and lung cancer (13%). The versatility of the data display interface led to exploration of several hypotheses not considered at the design stage of the trial, including the associations of fear of recurrence and hope (OR -0.40 -0.52, -0.28, p 0.001), fear of recurrence and belief in cure (OR -0.39 -0.51, -0.26, p 0.0001}, and ECOG score and fear of recurrence (0.24 0.024-0.45, p = 0.029). Identification of these putative associations prompted in-course changes in type of data collected, the number of patients accrued, and the type of hypotheses examined. CONCLUSION Our study disclosed unexpected associations by using a novel real-time data display tool build into the study data collection system. These unanticipated findings allowed us to make adaptive changes in study design during the course of the trial. This tool will also allow universal sharing of all deidentified study data at the time of study publication through a simple link embedded in the manuscript, as well as access to the data display interface. This innovation will provide readers with the ability to explore their own novel hypotheses, without the need for programming, statistical expertise or special software, thus enlisting all readers as co-investigators. This crowd sourcing approach is easily adaptable to any trial, can dramatically enhance the yield of individual published studies, can act as a guardrail to ensure academic integrity, and can overcome the abject failure of data sharing mandates of top-tier journals.
Franklin et al. (Wed,) studied this question.