Abstract

Objective: To present and demonstrate a new simplified method for synthesizing results of multiple clinical trials in resuscitation research. Methods: The mean difference across studies in the proportion of favorable outcomes between experimental and control groups is calculated. This difference is shown to have a t-distribution. Its significance can be ascertained with a simple t-test. The analysis can be implemented in a one-page computer spreadsheet. Results: Simplified meta-analysis provides high sensitivity and can be extended to include weighting of studies according to size or quality, comparison of subgroups of studies, tests for outliers, and calculation of the power of the meta-analysis. Sample analyses are presented for two experimental forms of cardiopulmonary resuscitation: interposed abdominal compression CPR and active compression-decompression CPR. Conclusions: Traditional narrative reviews, taking note of the proportion of individual studies with statistically significant results, can lead to erroneous conclusions and unnecessary delays in the clinical use of research findings. Simplified meta-analysis can provide rapid, quantitative, and accurate estimates of the amount of benefit or harm from an experimental intervention and can further empower physicians to practice evidence-based medicine.

Comments

This is the author accepted manuscript of Charles F. Babbs, Simplified meta-analysis of clinical trials in resuscitation, Resuscitation, 57, 245-255, 2003. Copyright Elsevier, this version is made available here CC-BY-NC-ND, and the version of record can be found at http://dx.doi.org/10.1016/S0300-9572(03)00038-8.

Keywords

Abdomen, Active compression-decompression CPR, Clinical trials, Guidelines, Interposed abdominal compression-CPR, Methodology, Statistical analysis

Date of this Version

2003

Share

COinS