Statistical researcher discusses reliability of literature claims
Stanley Young, assistant director for bioinformatics at the National Institute of Statistical Sciences in Research Triangle Park, N.C., was on campus Tuesday, April 29 to discuss the reliability of claims coming from human medical observational studies. Many of these studies only replicate 10-20 percent of the time, whereas randomized clinical trials replicate more than 80% of the time. Young discussed the systematic problems with how observational studies are conducted.
According to Young, vitamin E was widely reported to protect against heart attacks. However, large, well-conducted randomized clinical trials failed to replicate the claims. Likewise, similar claims were made concerning individuals with a so-called type-A personality having a greater predisposition for heart attacks. The claim failed to replicate in two separate studies.
“If the system in place is not working very well, you can’t count on the workers to change the system,” said Young. The managers at the top are responsible to change the system.”
Young discussed how technical changes, managerial changes, funding agencies and journal editors can repair this broken process and lead to more reliable claims along with fairer ways to judge their reliability.