Medical research is repetitive, often poorly designed, fails to take account of earlier evidence and produces results not fit for purpose, report Sir Iain Chalmers of the James Lind Library in Oxford and Professor Paul Glasziou in The Lancet online (June 15). 

 Many authors fail to read the literature before starting their studies, so duplicate what has already been done. And few are eager to see negative results published: under-reporting of research is usually the result of decisions by sponsors and researchers, not journal editors. Tens of billions of dollars a year are wasted because of correctable problems, they conclude.