Grant: CIHR Clinician Scientist - Phase 2

Meta-analysis is used to combine results from multiple randomized controlled trials to provide an overall measure of a therapy's effectiveness. I recently developed a new method, the ratio of means (RoM), to combine continuous outcome results that may be more easily interpretable by clinicians compared to traditionally-used mean difference methods. Using comprehensive simulation and comparison to actual data, I have shown that this new method gives comparable results to traditionally-used methods. Other biostatisticians have suggested an alternative ratio method, involving geometric means, a measure unfamiliar to most clinicians. Using comprehensive simulation and comparison to actual data, I plan to determine how the statistical properties of RoM compares to the ratio of geometric means. If RoM provides comparable results, then the use of this measure may make it easier for clinicians to interpret how the meta-analytic results of continuous outcomes apply to their particular patients. This research will also use meta-analytic techniques to evaluate the effectiveness of therapies for critically-ill patients requiring life support therapies to treat either kidney failure or severe breathing failure, both of which have very high mortality. Different methods for providing continuous forms of dialysis in kidney failure and different specialized therapies to improve oxygenation in breathing failure are used without clear guidance whether these methods improve survival. I plan to conduct separate systematic reviews and meta-analyses in both these areas to determine the effectiveness of each of the following therapies: high vs standard dose dialysis and type of clearance (hemodialysis vs hemofiltration) in acute kidney failure; and high frequency oscillation and inhaled nitric oxide in patients with the most severe low-oxygen breathing failure. This will provide effectiveness data to clinicians and identify areas where further research is required.