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

Background: Despite new advances in science, there is often a considerable lag between the generation of important research findings and their use by health care professionals. Objective: The purpose of LIFE CYCLE is to investigate the relative role of evidence and other factors in determining clinical recommendations and practice in cancer control. Metastatic breast and non small cell lung (NSCLC) cancers were chosen as case studies. All novel chemotherapy agents (NCA) approved for use in Canada between 1992 and 2002 were considered (n=9). Methods: We conducted cumulative meta-analyses of efficacy vinorelbine and taxanes (paclitaxel and docetaxel) in NSCLC from 1992-2002 replicating Antman et al. For each included randomized trial (RCT) we considered all types of publications, including abstracts from conference proceedings and subsequent journal publication(s). Cumulative evidence has been used in meta-analyses over sequential time windows (Folmann method). Additionally, bibliometric analyses have been performed to look at the payback (e.g. editorial, narrative review, textbooks and guidelines), and identify key milestones in the life cycle of evidence. Finally, we will plot the full body of research against the 1992-2002 use of the vinorelbine and taxanes in Ontario Regional Cancer Centers. Results: Vinorelbine: Out of 58 potentially eligible abstracts or full publications of RCTs, 7 RCTs were identified that provided meaningful comparisons and sufficient information to be included in cumulative meta-analysis. Of these, Le Chevalier (1994) was the first RCT that significantly supported the use of vinorelbine. Combining Le Chevalier's results with the other RCTs provided a pooled summary point that over time consistently favoured a survival advantage but confidence intervals crossed the null value at all times. Nevertheless, Le Chevalier had a strong literature payback, dominating the other trials. Taxanes: Out of 101 potentially eligible reports, only 13 provided meaningful comparisons and sufficient information. The pooled data analysis reached significance in 2002, when 8 RCTs results were available. Conclusions: This study demonstrates the relatively small pool of RCT evidence for NCAs. These two commonly used NCAs demonstrate different patterns of evolution of evidence and it is also reflected in their different citation patterns.