Food Safety and Quality:
Limitations of FDA's Bottled Water Survey and Options for Better Oversight
RCED-92-87, Feb 10, 1992
Pursuant to a congressional request, GAO reviewed the adequacy of the Food and Drug Administration's (FDA) 1990 Bottled Water Survey, focusing on whether the FDA: (1) survey results were representative of the nation's bottled water supply; (2) tests covered all regulated contaminants; and (3) testing for unregulated contaminants was warranted.
GAO found that: (1) FDA did not base the 1990 bottled water survey on a probability sample or use any formal design that considered or estimated the cost of such a survey; (2) the survey did not provide an adequate basis for the FDA statement that bottled water was safe, since FDA only inspected 49 domestic bottled water plants and tested 112 domestic and imported water samples; (3) to control costs, FDA instructed its districts to test the selected samples for only 9 of 31 regulated contaminants; (4) FDA also tested for 11 unregulated contaminants to determine the level of bacteria present and to detect and quantify the level of volatile organic chemicals; and (5) the decision to test for unregulated contaminants was warranted because FDA was considering new standards for those contaminants, which included benzene, which received much media attention, and the cost of performing those tests was only about 4 percent of the total survey cost of $850,000.