Trends in the Use of Preventive Care by Women in the U. S. General Population

Statement of the Problem: The disease burden of breast cancer remains strong as the second leading cause of cancer death among women in the United States. Despite this, the rate of women receiving a mammogram has shown little improvement since 2000. As health insurance will provide expanded coverage for mammography through the Affordable Care Act, it is essential to identify and describe women who have a lower probability of receiving the recommended breast cancer screening. Objective: To describe trends in receipt of biennial mammography in women in the United States, and to further identify independent sociodemographic and clinical predictors of women not receiving a mammography as recommended by national guidelines. Methods: Using data from the 2008-2010 Medical Expenditure Panel Survey (MEPS), I conducted a cross-sectional study and selected women forty years of age or older who had identified the time since their last mammogram as asked in the Preventive Care supplement. Within the final sample of 20,796 women, I assessed trends in mammography use in the previous two years, and differences in sociodemographic, clinical and type of usual source of care (creating 2 groups: no mammography, yes mammography). For analytic purposes, this study was analyzed as a case control study comparing women without a mammogram to women who had received a mammogram in the past 2 years. An unconditional logistic regression model was used to identify predictors of missed mammography and the data were weighted using SUDAAN software to account for the complex survey design and the nationally representative sampling scheme. Results: Overall, in the weighted sample of 210,485,707 women, 26.9% (n=56,532,799) did not receive a mammogram in the previous 2 years. Women in the case group (absence of recommended mammography) were of similar age and race/ethnicity as women in the control group (presence of recommended mammography). Overall, the study population consisted mostly of women aged 40-59 comprised by 60.1% of the cases and 57.9% of the controls. In multivariable modeling, the strongest predictor of missing a recommended mammogram was not having a usual source of care (OR=2.86; 95% CI, 2.52-3.25) and women without insurance during the study period (OR=2.34; 95% CI, 2.07-2.65). Conclusions: More than 1 in 4 women, or approximately 57 million women, did not receive the recommended biennial mammogram screening. Women without a usual source of care were less likely to receive the recommended screening, but variation across type of usual source of care was not apparent in my study. Age, race/ethnicity and other demographic and clinical characteristics were related to lack of receipt of mammography. With the passage of the Affordable Care Act, targeted interventions to reach the population subgroups less likely to receive the recommended mammography screening are essential.

1% per year for women aged fifty and over. 13 The increase in survival rates and decrease in death rates can be attributed to prevention efforts such as early detection and improved screenings. 13 Prevention services and screenings play an important role in providing direct and indirect health benefits and lowering health care costs.

Variability in Prevention
Multi-faceted environmental and socioeconomic factors are known to have a significant impact on variability in chronic disease rates and access to preventive care.
For chronic conditions such as diabetes, obesity, and hypertension, data from the Centers for Disease Control and Prevention (CDC) shows that the black non-Hispanic population has the highest disease rate followed by Hispanic and American Indian/Alaskan native in 2009. 11 More recently, the National Health Interview Survey (NHIS) reported on ten year trends (2000-2010) of adults 45 years of age and older with multiple chronic conditions. The survey suggests racial and ethnic disparities in the prevalence of two or more chronic conditions more non-Hispanic black adults had two or more chronic conditions than adults in other racial and ethnic groups. 14 Disparities not only exist among race/ethnicity but also among income levels. The NHIS found that the prevalence of two or more chronic conditions in the 45-64 age group significantly decreased as family income increased; those below 100% of the poverty line had a prevalence rate four times those 400% above the poverty line. The percentage of adults that did not receive or delayed medical care due to cost in the 45-64 age group increased 36% in the last ten years. In contrast, the percentage of the 65 and older age group that delayed or did not receive medical care due to cost remained  have consistently shown the highest percentage over the time period evaluated, followed by women aged 65-74. 2 By percent of poverty level, the proportion of women having a mammogram was directly related to the percent of poverty, as the percent of poverty increased (or becomes wealthier) the percent of women receiving a mammogram also increased. 2 Similarly, each level of higher education (categorized as no high school, high school, college or more) resulted in a higher percentage of mammography screening. 2 When looking at insurance, individuals with private insurance had the highest usage of mammography while the uninsured had the lowest. 2 Although select disparities have improved over time, the overall rate of women receiving a mammogram has not improved since in the last decade and variability in prevention remains.

Usual source of care for Prevention
Previous data demonstrated a statistically significant benefit of having both a usual source of care and insurance on the likelihood of receiving preventive care. This trend was also true for receipt of a mammogram among women aged 40-69 years old. 9 Having either usual source of care or insurance, but not both, gave inconclusive results in the receipt of mammogram. The study also determined that the uninsured group without a usual source of care was more predominant in Hispanic and non-white subgroups and in households without a high school education. 9 According to a study using the NHIS database in 1999, having a 'usual place' or 'usual place and provider' was associated with increased likelihood of having received preventive service or screening. 15 A similar 1996 NHIS study found that women having a usual source of care were 4 times as likely to receive Pap smears, 2 times as likely to receive breast exam, and 3 times as likely to receive mammogram. 10

Affordable Care Act and Increased Coverage of Mammography
The Affordable Care Act, which was signed into law in March 2010, will reach full implementation by 2014. As part of this act, preventive health care services including mammography, are covered with no cost sharing (such as copayments, co-insurance, and deductibles) when offered by a provider in network. Mammography, among other preventive services deemed eligible, is also covered under Medicare with no out of pocket costs for women at least forty years of age, as long as the provider accepts assignment. 16

Summary
As full implementation of the Affordable Care Act approaches, the rate of women receiving a mammography screening as recommended according to national guidelines remains suboptimal. In 2010, only about 67% of women aged 40 or older received a mammogram in the last two years. 2  To increase the sample size for analyses, the use of 3 years of data was necessary.

Study Definitions
Use  Figure 1. Amongst the eligible sample, 26.9% (n=56,532,800) did not report receiving a mammogram in the previous 2 years. Several demographic and clinical differences were identified between women not receiving a mammogram and those women receiving a mammogram in the previous 2 years. As presented in Table   1, women in the case group (absence of recommended mammography) were of similar age and race/ethnicity as women in the control group (presence of recommended mammography). Overall, the study population consisted mostly of women aged 40-59 comprised by 60.1% of the cases and 57.9% of the controls. The prevalence of women aged 80 years or older was higher in the cases, 13.9% compared to 6.8% in the controls. Cases were more likely to have less than a high school education (20.4% in cases and 11.0% in controls), be widowed, divorced, separated or never married (52.1% in cases and 39.0% in controls), and have a negative/poor income (16.2% in cases and 8.7% in controls) or have a low income (25.6% in cases and 15.6% in controls). As presented in Table 2, controls had a slightly elevated history of any cancer (17.5% in controls compared to 13.4% in cases) as well as a history of breast cancer (5.1% in controls compared to 3.2% in cases). The prevalence of women that did not engage in moderate to vigorous physical activity was higher in the cases (53.1% in cases and 45.8% in controls). Cases were also more likely to have a fair/poor perception of their own health status (22.7% in cases and 16.0% in controls).
Patterns of mammography utilization by usual source of care are detailed in Table 3.
Overall, cases were less likely to have a usual source of care when compared to the controls (23.7% in cases and 8% in controls). Of women who did have a usual source of care, the cases were less likely to have a primary care provider (38.5% in cases and 49.8% in controls). However, women citing their usual source of care as a specialist or non-medical doctor were similar when it came to receipt of mammography. Provider type presented a significant amount of missing data in both groups (34.9% in cases and 39.8% in controls) and therefore further analyses were not possible. Overall, the majority of women in the study population (> 75% in both groups) answered yes to being insured any time during the survey period. However, cases were less likely to have insurance then controls (22.6% in cases and 7.7% in controls).
The results of the logistic regression model designed to identify independent predictors of missed mammography are presented in Act was implemented to cover screening mammography with no cost-sharing for eligible health plans on or after September 23, 2010. As full implementation of health care reform is underway, and the cost barrier for screening mammography has been minimized, the findings of this study emphasize important areas for health care intervention for women at risk of missed screening. With breast cancer as the second leading cause of cancer death among women 13 , the results of the study also serve as a reminder for disease and preventive care awareness, especially when predispositions exist.
One of the strongest findings from my study was the increased odds among women without a usual source of care or insurance. Women without a usual source of care were ~3 times more likely than those with a usual source of care to miss a recommended mammogram. Similarly, women without insurance during the study period were ~2 times more likely than those with insurance to miss a recommended mammogram. In my study, 1 out of 4 did not have a usual source of care. Of women who did have a usual source of care, those who did not receive a mammogram were also less likely to have a primary care provider. These results are consistent with previous data that has demonstrated a statistically significant benefit of having both a usual source of care and insurance on the likelihood of receiving preventive care (e.g. Within the study, I assessed additional predictors based on previous research, [24][25][26][27][28] including education, marital status, poverty status, and physical activity level, all of which may contribute to use (or lack of use) of mammogram. Women with at least a high school degree were 27% more likely to report having a recent mammogram than women with less than a high school education. Women who were widowed, divorced, separated, or never married had 34% higher odds of missed mammogram than women who were married (OR=1.34; 95% CI, 1. 21-1.49