Figure 2.

Virtuous cycle feedback loops. Using existing results and materials to refine hypotheses and develop new insights can produce “virtuous cycles,” where the research efforts of today feed tomorrow’s discoveries. Figure 2 shows two kinds of such cycles that are implied by the ancillary study workflow described here. An ASMS can facilitate both types of cycles by smoothing the flow of information, enabling collaboration, simplifying workflows and allowing researchers to make the most of existing materials and information.

(i) Full study cycle. The nine steps in the ancillary study workflow form a virtuous cycle that spans the full life of a study, from the first glimmer of an idea through publication. For simplicity, Figure 2 breaks these steps into three phases (study initiation, study execution, and results sharing). These steps are roughly equivalent to those that form the “inner,” study-based loop in Kahn and Weng’s conceptual model for clinical research informatics [96]. In such cycles, published hypotheses and shared data from completed studies are used to generate future discovery cycles by providing inspiration and ingredients for follow-up studies.

(2) Incremental review cycles. An ASMS can also facilitate smaller-scale virtuous cycles during all phases of an ancillary study. First and foremost, during the study initiation phase, the information and tools made available by an ASMS allow incremental refinement of hypotheses and study plans according to existing data, specimen availability, and consent limitations. During later phases of a study, an ASMS can make it easier to share and review new information as it is collected, allowing feedback of new insights into study investigations, operations, analyses, and conclusions. Of course, in-progress studies governed by clinical trial regulations will provide less scope for immediate use of this type of feedback than the kinds of pre-clinical, exploratory studies common among our collaborators

Nelson et al. BMC Medical Informatics and Decision Making 2013 13:5   doi:10.1186/1472-6947-13-5
Download authors' original image