Cumulative meta-analysis to determine key milestones in the Life Cycle of Evidence in Cancer Care

Background: Science progresses through a series of paradigms that are held to be true until they are replaced by a better approximation of reality. Despite new advances in science, there is often a considerable lag between the generation of important research findings and the use of this knowledge by health care professionals. Objective: The purpose of LIFE CYCLE-ECC is to investigate the relative role of evidence and other factors (e.g. funding approval, drug industry promotion, consumer advocacy) in determining clinical recommendations and practice in cancer control. Metastatic breast cancer and non small cell lung cancer were chosen as case studies on which to base the study. All novel chemotherapy agents (NCA) approved for use in Canada between 1992 and 2002 were considered. Methods: We are conducting cumulative meta-analyses of efficacy of each identified NCA (n=13) following the model of Antman et al.[1] For each included trial we will consider all types of related publications, starting with abstracts from conference proceedings, to subsequent journal publication(s). Cumulative evidence will be used in meta-analyses over sequential time windows. In addition, bibliometric analyses are being performed to look at the payback (e.g. editorial, narrative review), in order to identify key milestones in the life cycle of evidence, which result in translation into recommendations. Finally, we will plot the full body of research against the actual use of the NCAs in Ontario Regional Cancer Centers from 1992-2002. Follow-up qualitative analyses using focus groups with key stakeholders will be performed to better understand influences for practice decisions in Ontario. Results: Currently we are screening all eligible papers. With the large number of documents involved, this study outlines the necessity to establish the methodology to manage all the quantitative and qualitative information in order to guarantee feasibility. Potential pitfalls include: identifying a coherent starting point for the analysis; tracing the research findings forwards versus backwards; coping with and storing the enormous number of papers involved; develop new meta-analytic approaches to deal with a body of research including multiple drugs and outcomes at discrete times of research; develop ways to fully utilize bibliometric citation analysis and the cumulative rates analysis; presenting the findings in a multi-dimension document. Conclusions: This research will make a major contribution to our understanding of the way results of clinical research eventually feed through into application and which key milestones are most influential. However before embarking on such a large-scale, multi-level study, many ideas and methodological issues must be explored and resolved.