Screening Mammography Rates Overestimated for Older Women

SAN FRANCISCO, June 20 ? Women in their Medicare years may claim to have regular screening mammograms but many aren't as good as their word, a study has found.

Overall, older women are significantly less likely to have regular mammograms than self-reported surveys suggest, according to a report in the June American Journal of Preventive Medicine.

Widely cited self-reported data from the Behavioral Risk Factor Surveillance System (BRESS) and the National Health Interview Study (NHIS) suggest levels as high as 70% to 80% for biennial screening of women ages 65 to 69, with no substantial difference in race or ethnicity. However, Medicare data tell a different story, found Rebecca Smith-Bindman, M.D., and colleagues at the University of California here and at Harvard Medical School

In their study, a 5% representative group of women, ages 65 and older, in 11 Surveillance, Epidemiology, and End Results (SEER) areas from 1991 to 2001 was used to construct Medicare data for 146,669 women. Multivariate repeated-measures logistic regression analysis served as predictors of screening usage.

From 1991 to 2001, rates for screening older women once every two years increased from 35.8% to only 47.9%, the researchers said.

During this period, the age- and race-adjusted proportion of women who had biennial screening mammography increased for all racial and ethnic groups, but was significantly higher for non-Hispanic white women.

The biennial screening rate in 2000 and 2001 was 50.6% fornon-Hispanic white women, 40.5% for African-American women, 34.7% for Asian-Americans, 36.3% for Hispanics, and 12.5% for Native American women, Dr. Smith-Bindman's team reported.

After controlling for age, site, physician access, comorbidities, education, and income, non-Hispanic white women were still more likely to have had a mammogram.

African Americans (odds ratio 0.80, 95% CI, 0.78-0.83), Asian Americans (OR 0.53, CI 0.51- 0.55), Hispanics (OR 0.70, CI 0.67-0.74), and Native Americans (OR 0.37, CI 0.29-0.46) were all less likely than non-Hispanic white women to undergo screening, the investigators reported.

From 1991 to 2001, the overall rates increased for all groups (35.8% to 47.9%). However, at the outset in 1991 and 1992, the rates for non-Hispanic whites were 21% higher than those for African- Americans, 49% higher than the rates for Asians, and 45% higher than the rates for Hispanics.

By 2000-2001, the differentials for whites versus other groups were 25% (African-Americans), 46% (Asians), and 39% (Hispanics). Although the magnitude of the earlier 21% gap was smaller for African-American women, "it is of concern that the gap [25%] has grown over time," Dr. Smith-Bindman said.

In 2000-2001, overall screening rates were 61.2% for women 65 to 69 years old and declined with increasing age, especially after age 75. The importance of these trends remains to be seen, awaiting further evidence of the value of mammography for the older elderly, the investigators wrote.

However, they noted that the American Cancer Society and the American Geriatrics Society, for example, recommend continued screening for older women unless they have comorbid conditions that would limit their life expectancy or make them poor candidates for breast cancer treatment.

Women who saw a primary-care provider in a given year were almost 75% more likely to have a mammogram (OR= 1.73, CI 1.71-1.76.). Women who saw an obstetrician-gynecologist were also more likely to be screened.

A limitation of the study, the researchers said, was the use of Medicare claims. Other studies have questioned the sensitivity of Medicare data for identifying African-American women, and especially Hispanics and Asian-Americans.

In addition, they wrote, women enrolled in Medicare HMOs were not included, and there may also have been screening mammograms not captured by Medicare claims. Finally, the 11 SEER areas, although geographically dispersed, do not represent a random sample of the U.S. population, they noted.

All told, elderly women undergo significantly less screening than self-reported surveys suggest, while the size of the screening discrepancies are especially large for Asian-American and Hispanic women, Dr. Smith-Bindman said.

Finally, she said, despite the smaller size of the screening gap for black women, the fact that the gap has grown over time is worrisome.