Meta-analysis: Methods, strengths, weaknesses, and political uses

A critical examination of meta-analysis studies by John H. Noble, Jr, Ph.D, AHRP Treasurer, who recently retired from his position as an endowed chair at Catholic University. The article is published in the Journal of Laboratory and Clinical Medicine.

While analyzing the methods of meta-analysis, the article is primarily an indictment of the state of science in the hands of biomedical researchers who are not free of conflicts of interest.
Dr. Noble cites recent examples shrouded in controversy to make his case–not only against industry’s tainted science–as demonstrated in the published reports about the COX 2  painkillers and the SSRI antidepressants, but also against the tainted science produced and disseminated by the National Institutes of Health.  Case in point is the ARDS Network lethal multi-site experiment in which critically ill patients are being exposed to increased risk of death.
Dr. Noble makes several recommendations for restoring confidence in the scientific enterprise, including Congressional action in its oversight capacity:
  “Congress will have to pass legislation that either bans or severely limits ‘trade secret’ status to the results of clinical trials in support of new drug of device applications for FDA approval.”
“Close linkage of US PhRMA with the Office of the U.S. Trade Representative is considered a threat to affordable medicines and equitable health care throughout the world by inserting preferential property provisions in bilateral and multilateral trade agreements. Protection of frivolous patent claims through these agreements promotes the practice of ‘evergreening’ of brand-name drugs to the detriment of generic-drug development and serves to distort biomedical research priorities in the interest of monopolistic market share while encouraging biomedical researchers to behave badly.”

There is simply no way to preserve scientific integrity in a sea of corporate compromise and scientists with vested interests of their own.

Contact: Vera Hassner Sharav
The general methodology, strengths and weaknesses, and political uses of metaanalysis are examined. As a systematic study of all studies that have been conducted to answer a specific question or hypothesis, meta-analysis is strong in revealing structural flaws and sources of bias in primary research and in posing promising research questions for future study. It cannot exceed, however, the limits of what is reported by primary researchers.

Meta-analysis is particularly challenged to quantify the size of a common effect of treatment across reported trials because of (1) the clinical diversity of the trials and (2) the myriad of potential differences among patients with varying characteristics within the trials. Without access to the original data of reported trials, meta-analysis cannot overcome the bias of underpowered trials toward overstatement of the size of main treatment effects, nor the tendency for such trials to falsely conclude there were no statistically significant adverse events. Although severely compromised by ghost-written or honoraryauthored reports of primary research, meta-analysis can make use of its methods to focus on the conflicts of interest and likely sources of bias of such research and make known what precautions should be taken by would-be consumers. Examples show how meta-analysis has clarified thinking about the off-label use of selective serotonin reuptake inhibitors for treating child and adolescent depression, use of low-tidal volume respirator assistance for acute lung injury and acute respiratory distress syndrome patients, and the long-term use of COX-2 inhibitors for relieving
arthritic pain. Recommendations are made for Congressional action.

J Lab Clin Med 2006;147:7–20

 Contact Mr JOHN NOBLE JR []