March 3, 2016

Conclusions from Multiple Studies Can Be Flawed

The gold standard of medicine has been to find one treatment that makes everyone healthy.  To find this "holy grail" authors compare the results from a large number of studies (meta-analyses, systematic reviews, Cochran Reports).  Yet conclusions from multiple studies can be flawed when a disease varies widely as stroke does.  Here is two examples of what I am talking about.

Researchers are doing new clinical trials on drugs that did not help large numbers of people with cancer.  Now that researchers know how to identify different subtypes of cancer, they want to see if old studies had a poor match between the drug and the subtype of cancer in the sample.

Researchers in the Efficacy of Nitric Oxide Study (ENOS) studied the effects of nitric oxide (NO) on stroke outcomes (1).  ENOS looked at the records of over 4,000 stroke survivors in 23 countries. NO had no beneficial effect. The ENOS authors did a 2nd analysis with a subset of people with strokes that affected the front half of the brain (1).  Ninety days later the subjects who received NO had significantly better outcomes on the Mini-Mental State Exam (cognition), Barthel Index (ADLs), EuroQol Scale (quality of life), and Zung Depression Scale.

Bottom Line. Authors of meta-analyses, systematic reviews, and Cochran Reports need to take a more nuanced approach.  When deciding what helps people with a disease that varies widely, conclusions that ignore how samples differ across studies can dismiss beneficial treatment effects.

Example of a Nuanced Approach. The CAREX exoskeleton helped stroke survivors draw circles by providing two kinds of assistance (2).  High functioning stroke survivors who had strength but lacked control improved when they were given path assistance.  Low functioning stroke survivors who lacked strength improved when CAREX supported the arm so they had less weight to move.  

1. Woodhouse L, Scutt P, Krishnan K, Berge E, Gommans J, Ntaios G, Wardlaw J, Sprigg N, Bath
     P. Effects of hyperacute administration (within 6 hours) of transdermal glyceryl trinitrate, a nitric
     oxide donor, on outcome after stroke. Stroke. 2015;46:3194-3201.

1 comment: