Power meta-analysis

Cochrane Library. The Cochrane Library [44] includes The Cochrane Database of Systematic Reviews, a collection of regularly updated, systematic reviews of the effects of health care. It is maintained by contributors to the Cochrane Collaboration. Cochrane reviews are reviews mainly of randomized controlled trials. To minimize bias, evidence is included or excluded on the basis of explicit quality criteria. Data are often combined statistically, with meta-analysis, to increase the power of the findings of numerous studies, each too small to produce reliable results individually. Database of Abstracts of Reviews of Effectiveness is also included. It consists of critical assessments and structured abstracts of good systematic reviews published elsewhere. The Cochrane Controlled Trials Register with bibliographic information on controlled trials and other sources of information on the science of reviewing research and evidence-based health care are part of the Cochrane Library. It is commercially available on CD-ROM or the Internet. 

In 2001 two Danish researchers, Asbjorn Hrobjartsson and Peter Gotzsche, published an influential meta-analysis in which they estimated the difference between the effects of getting a placebo versus doing nothing at all.14 Although they found a significant placebo effect, especially in the treatment of pain, the overall effect seemed very small - much smaller than would have been expected of a powerful treatment. On the basis of these data, the researchers asked Is the placebo powerless and answered their own question by concluding that there was little evidence that placebos have powerful clinical effects. 

Meta-analysis aims to increase the statistical power of the available evidence by combining the results of smaller trials together using specific statistical methods. The validity of the meta-analysis will depend on the quality of the evidence on which it is based and how homogeneous or comparable the samples are. Combining very heterogeneous study populations can lead to bias. 

Combining studies can also very effectively increase power for primary or secondary endpoints or for particular subgroups. Individual studies are unlikely to be powered for secondary endpoints and subgroups, and meta-analysis can be... 

Dolovich, L., Addis, A., Vaillancourt, J.M.R., Power, J.D.B., Koren, G., and Einarson, T.R. (1998) Benzodiazepine use in pregnancy and malformations ot oral clefts meta-analysis of cohott and case-control studies. BMJ 317 839—843. 

Coenzyme QIO is a powerful antioxidant naturally occurring in the mitochondria of myocardium, and it is an electron carrier in the mitochondrial synthesis of ATP. Patients with heart failure have lower myocardial levels of coenzyme QIO, but supplementation has been demonstrated to have variable benefits in randomized controlled trials. One meta-analysis on the use in congestive heart failure showed improvements in stroke volume, ejection fraction, cardiac output, cardiac index, and end diastolic volume index. " Another antioxidant associated with beneficial effects in cardiac patients is lycopene, a natural constituent of tomatoes. Lycopene is the major carotenoid found in human serum, and epidemiological studies have indicated an effect of dietary supplementation in reducing heart disease. Few dietary interventions have been reported one study showed a mild but significant hypocholesterolemic effect, and another showed a significant reduction in LDL oxidation. " Animal studies show an antiatherogenic effect of DHEA, and a review of the clinical trials and studies on DHEA in males with coronary heart disease reported a favorable or neutral effect. Plasma levels of DHEA are decreased in patients with chronic heart failure in proportion to its severity. ... 

An excellent comprehensive review of the design, analysis and interpretation of a meta-analysis has been published by Petitti. If careful attention is paid to its potential problems, a meta-analysis can increase statistical power, resolve apparently conflicting results, and give more precise estimates of effect size. 

Pavesi L, Subburaj S et al (2001) Apphcation and validation of a computerized cough acquisition system for objective monitoring of acute cough A meta-analysis. Chest 120 1121-1128 Power JT, Stewart IC et al (1984) Nocturnal cough in patients with chronic-bronchitis and emphysema. Am Rev Respir Dis 130 999-1001... 

---