Breast and Cervical Cancer Screening among Appalachian Women

Appalachia historically has been underserved by the health care system (1, 2). In addition, rural residence, geographic isolation, poverty, unemployment, lack of education, lack of child care services, and attitudinal and cultural factors may pose barriers to cancer screening among Appalachian women (2, 3, 4, 5, 6).

The United States Department of Health and Human Services (7) sets nationwide target goals for breast and cervical cancer screening. In addition, a primary goal is to achieve equity in health to eliminate disparities. Race, ethnicity, education, income, and rural location have been found to be associated with less use of preventive screening services and lack of health insurance (7). There is little published information about screening rates among women in the Appalachian region. Similar to studies among the United States population, small studies within Appalachia found lower rates of breast and cervical cancer screening among women who were older, less educated, or unemployed; lived in rural settings; had lower incomes; lacked health insurance; or were covered by public health insurance (4, 5).

We used data from the Centers for Disease Control and Prevention BRFSS² to determine the prevalence and to characterize the correlates of breast and cervical cancer screening among Appalachian women. The Appalachian region comprises all 55 counties of West Virginia and a total of 351 counties in Alabama, Georgia, Kentucky, Maryland, Mississippi, New York, North Carolina, Ohio, Pennsylvania, South Carolina, Tennessee, and Virginia.³

The BRFSS collects data on behavioral risk factors for the adult population ≥18 years of age and living in households.⁴ State health departments identify a probability sample of all households with telephones, administer the interviews, and transmit the information to Centers for Disease Control and Prevention for editing, processing, weighting, and additional analysis. Telephone coverage ranges from 87 to 98% across states and varies by subgroups but was not available specifically for the Appalachian region.⁵ The estimated median response rate for 1998 was 59.2% but ranged from 32.5 to 76.7% across states.⁶

To determine whether women ≥40 years of age received breast cancer screening in the recommended time frame, we determined the percentages of women who had received a mammogram or a CBE within the past 2 years. For cervical cancer screening, we determined whether a Pap test had been received within the past 3 years among women ≥18 years of age who had not had a hysterectomy. Other information obtained from BRFSS included age, race, marital and employment status, education, household income, number of persons living in the household, number of dependent children in the household who were <18 years of age, health insurance, health status, height, weight, smoking, doctor visit within the past year, and diabetes. BMI [weight (in kilograms) divided by height (in meters squared)] was categorized according to the International Obesity Task Force classification (BMI <25, 25–29, and ≥30; Ref. 8).

We combined data for 1996–1998 to obtain adequate sample sizes for subgroup analyses. We calculated the percentage of Appalachian women receiving the screening tests and 95% CIs overall and by demographic and other potential correlates (univariate analyses with χ² statistic; P ≤ 0.05). To compare screening rates between Appalachian and other United States women, we calculated age-adjusted percentages, using the age distribution for women from intercensal estimates for 1996, 1997, and 1998 (9). The age-adjusted rates were compared by calculating a Z statistic (P ≤ 0.05). All analyses were weighted to adjust for differences in probability of selection, nonresponse, and noncoverage.

Multivariate logistic regression analyses were conducted to identify factors associated with screening among Appalachian women. Variables that were significant at P ≤ 0.2 in the univariate analyses were included in the stepwise backward elimination procedure, and variables were retained at P ≤ 0.05. We also conducted multiple logistic regression analyses including Appalachian and other United States women to determine whether Appalachian residence was associated with screening rates after adjustment for other factors. All analyses were weighted.

A total of 20,785 Appalachian women ≥18 years of age were interviewed for the BRFSS during 1996–1998. Of these, 28% (5,836) were excluded from the analyses for Pap tests because they had had a hysterectomy or their hysterectomy status was unknown. Information was available for 14,520 women on whether they had had a Pap test within the past 3 years.

At the time of interview, 13,460 women were ≥40 years of age. Information was provided by 13,223 Appalachian women on whether they had had a mammogram within the past 2 years and by 13,124 women on whether they had had a CBE within the past 2 years.

Overall, the weighted estimate of having had a recent mammogram among Appalachian women was 68.8% (95% CI, 67.8–69.9). The rate among Appalachian women was 3.2% lower than that among other United States women (the age-adjusted rates were 68.6% for Appalachian women and 71.8% for other United States women; P ≤ 0.01).

Within Appalachia, the percentages of women screened were higher among those 50–59 and 60–69 years of age than among younger or older women (Table 1). A higher prevalence of mammograms was also found among women who were married; had a higher education level, higher household income, better health status, or fewer children; or lived in two-person households. Women who were employed, retired, or homemakers had mammograms more frequently than did unemployed women. Women who were insured, had had a doctor’s visit in the past year, or had diabetes were also more likely to have been screened. Smokers were less likely than nonsmokers to have had a mammogram. Women with a BMI <25 were somewhat less frequently screened than were women with BMI ≥30 or BMI 25–29.

We compared screening rates of Appalachian women and other United States women according to sociodemographic factors by calculating the difference (D) between the two percentages (Table 1). For many sociodemographic subgroups, screening was reported by a somewhat smaller percentage of Appalachian women than by other United States women. For example, mammograms were received by a smaller percentage of Appalachian women who were ≥60 years of age, had less than a high school education, or who had not visited a doctor in the past year than by other United States women.

Multivariate logistic regression analyses confirmed that, among Appalachian women, a doctor visit within the past year, education, and household income were strong predictors for having had a mammogram (data not shown). Other predictors of mammogram screening were age, marital and health status, insurance coverage, smoking, diabetes, and BMI.

Multivariate logistic regression analyses also confirmed a significantly lower rate of screening in Appalachian women than other United States women after adjusting for other predictors of screening (adjusted OR, 0.89; 95% CI, 0.84–0.94).

Overall, the weighted estimate of Appalachian women who had had a recent CBE was 75.1% (95% CI, 74.1–76.1). The rate among Appalachian women was 2.5% lower than that among other United States women (the age-adjusted rates were 75.0% for Appalachian women and 77.5% for other United States women, P ≤ 0.01).

Within Appalachia, CBE screening was more common among women who were ≤60 years of age, married, or employed; or had higher education or income, better health status, health insurance; or who had visited a doctor within the past year (Table 1). Smokers were less commonly screened than nonsmokers.

For many socioeconomic subgroups, a somewhat smaller percentage of Appalachian women reported screening than other United States women. For example, CBE screening rates among women who were ≥60 years of age, never married, widowed, retired, or living in a three-person household or had not visited a doctor in the past year were lower among Appalachian women than among other United States women (Table 1).

The multivariate analysis confirmed that, among Appalachian women, CBE screening was associated with having had a doctor visit within the past year, age, marital status, education, household income, number of persons in the household, and BMI (data not shown). Multivariate logistic regression analyses also confirmed a significantly lower rate of screening in Appalachian women than other United States women after adjusting for other predictors of screening (adjusted OR, 0.91; 95% CI, 0.85–0.97).

Overall, the weighted estimate of Pap screening among Appalachian women was 82.4% (95% CI, 81.5–83.3). The screening rate among Appalachian women was 2.5% lower than that among other United States women (the age-adjusted rates were 81.1% for Appalachian women and 83.6% for other United States women, P ≤ 0.01).

Within Appalachia, Pap screening was less common among women who were ≥65 years of age, were widowed or retired, had less than a high school education or no health insurance, or had not had a doctor’s visit within the past 2 years (Table 2).

Within some sociodemographic subgroups, a somewhat smaller percentage of Appalachian women reported having had Pap screening than did other United States women (Table 2). For example, among women who were ≥40 years of age, were never married, were homemakers, had low income, or had not visited a doctor in the past year, Pap tests were received less frequently by Appalachian women than by other United States women.

Within Appalachia, the multivariate analyses confirmed the associations between Pap screening and age, race, marital status, education, income, number of persons in the household, health status, doctor visit within the past year, and BMI (data not shown). Multivariate logistic regression analyses also confirmed a significantly lower rate of screening in Appalachian women than other United States women after adjustment for other predictors of screening (adjusted OR, 0.85; 95% CI, 0.79–0.92).

Our results suggest that a high percentage of Appalachian women are being screened for breast and cervical cancer. However, screening rates were lower than the national target screening rates of Healthy People 2010 and among older women; women of lower socioeconomic status, measured by education, income, or health insurance; and women who did not visit a doctor in the past year. These women may face significant barriers to receiving screening services.

Our results confirm findings of other United States (5) and Appalachian (4, 5) studies that older women generally have lower rates of screening. We also noted lower rates of mammography screening among women 40–49 years of age. Mammography screening among women in this age group has been controversial because evidence is unclear as to whether it yields a significant reduction in mortality (10). Specific age groups may be targeted for screening interventions to achieve the national health objectives.

Special efforts to reach underserved women for breast and cervical cancer screening in states of the Appalachian region as well as elsewhere in the United States began in the early 1990s (11, 12, 13). Nevertheless, similar to earlier observations (5, 14) we found that socioeconomic indicators, such as lower education or income or lack of health insurance, were associated with lower screening rates. We also found that women who had visited a doctor within the past year were screened at much higher rates than were those who had not visited a doctor. A study among indigent, rural Appalachians found that lack of knowledge and cost were frequent reasons given for not having received screening services (15). Interventions for these subgroups need to be based on effective strategies that have been found to reach underserved women. Mobile mammography van services and integration of screening at primary health care sites are successful approaches for screening older women (16). In addition to mobile units, community-based education interventions and establishment of local screening centers have been found to be effective approaches in Appalachian and rural settings (13, 17, 18, 19).

A possible limitation of this study is response bias, because ∼27% of contacted United States households did not respond to the surveys. In addition, the telephone survey excluded women living in households without a telephone. Another limitation is that self-reported information about cancer screening practices may differ from information obtained from the records of healthcare providers. Persons tend to overreport their use of screening and to underreport the time since their last screen (20). Studies of the reliability of cancer screening information collected as part of BRFSS, however, have shown that the reliability of self-reported information about breast and cervical screening is reasonably good (21).