BOSTON – A widely used screening tool for colon cancer, sigmoidoscopy, misses precancerous growths in almost two-thirds of women – a disturbing failure rate twice as high as the one seen in men, a government-backed study found.

“All our data until now has been based on men. What this data shows is it’s a lot worse in women,” Dr. Philip Schoenfeld of the University of Michigan and U.S. Department of Veterans Affairs, said.

The researchers said colonoscopy – a more reliable but more expensive test than sigmoidoscopy – should now be considered the preferred method for most women.

Even before the study, patients were increasingly choosing colonoscopies, in part because of serious questions about the reliability of sigmoidoscopies. The latest research is likely to accelerate the trend.

Both colonoscopy and sigmoidoscopy involve the use of a long, flexible tube to inspect the twisting colon for precancerous growths, or polyps. But a colonoscope is longer and is used to examine the entire 6-foot length of the colon, while a sigmoidoscope inspects only the lower 2 feet or so.

The new study, led by Dr. Schoenfeld, involved 1,463 women, ages 50 to 79, at four military hospitals. It was published today in the New England Journal of Medicine and funded by the National Cancer Institute.

The researchers first used colonoscopes to find all polyps. They then calculated how many worrisome ones would have been missed by sigmoidoscopy alone.

Sigmoidoscopes were already in question because they find precancerous tumors in only 66% of average-risk men who have them. In this new study, sigmoidoscopes did even worse in average-risk women, detecting precancerous tumors in just 35% of the patients who had them.

Medical guidelines recommend either regular sigmoidoscopies or colonoscopies after age 50. The number of colonoscopies has already been rising in recent years, especially among high-risk patients. A federal survey put the number of colonoscopies at 14 million in 2002, compared to 3 million sigmoidoscopies.