The usual data massaging wasn't enough, like ignoring 19 of the 30 existing studies as of then, so they changed the confidence level required from the standard

The usual data massaging wasn't enough, like ignoring 19 of the 30 existing studies as of then, so they changed the confidence level required from the standard 95% to 90%. Look up commentary on their 1993 "Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders" for more details.