Perceptions of insurance coverage for screening mammography among women in need of screening

Authors

  • Ann Scheck McAlearney Sc.D.,

    Corresponding author
    1. Division of Health Services Management and Policy, School of Public Health, The Ohio State University, Columbus, Ohio
    2. Department of Pediatrics, College of Medicine and Public Health, The Ohio State University, Columbus, Ohio
    • Division of Health Services Management and Policy, The Ohio State University, 1583 Perry Street, Atwell 246, Columbus, OH 43210-1234
    Search for more papers by this author
    • Fax: (614) 438-6859

  • Katherine W. Reeves M.P.H.,

    1. Center for Population Health and Health Disparities, Division of Population Sciences, Comprehensive Cancer Center, The Ohio State University, Columbus, Ohio
    Search for more papers by this author
  • Cathy Tatum M.A.,

    1. Center for Population Health and Health Disparities, Division of Population Sciences, Comprehensive Cancer Center, The Ohio State University, Columbus, Ohio
    Search for more papers by this author
  • Electra D. Paskett Ph.D.

    1. Center for Population Health and Health Disparities, Division of Population Sciences, Comprehensive Cancer Center, The Ohio State University, Columbus, Ohio
    2. Division of Epidemiology, School of Public Health, The Ohio State University, Columbus, Ohio
    Search for more papers by this author

Abstract

BACKGROUND

Breast carcinoma remains a significant health problem in the U.S., especially among underserved populations. Although screening mammography is recommended for early detection, in 2002, approximately 25% of women age > 40 years had not had a mammogram within the past 2 years. The current study examined perceptions of insurance coverage and cost as barriers to screening mammography within an underserved, predominantly low-income population of women in need of a mammogram.

METHODS

Between 1998 and 2002, face-to-face interviews were conducted with 897 women age ≥ 40 years. All women were part of a randomized, controlled study evaluating a health education intervention designed to improve mammography screening. They were asked questions at baseline about cost and insurance coverage as barriers to mammography screening. Women's reports of their level of insurance coverage for mammography were compared with actual coverage by their insurance type to determine the accuracy of their perception of insurance coverage for mammography. The relation between perception of insurance coverage and the barrier of cost was investigated.

RESULTS

Greater than half of the women who needed a mammogram identified cost as a barrier to mammography; however, 40% of these women had an inappropriate perception of their insurance coverage. Underestimating or not knowing the level of mammography coverage was strongly associated with reporting cost-related difficulty (odds ratio [OR] = 4.57, 95% confidence interval [95% CI], 1.95–10.70 for the underestimate category; OR = 4.42, 95% CI, 1.80–10.88 for the don't know category), regardless of true coverage levels.

CONCLUSIONS

Providing women with information regarding their actual coverage for mammograms may reduce the impact of cost as a barrier to screening mammography. Cancer 2005. © 2005 American Cancer Society.

Breast carcinoma remains the second leading cause of cancer death among women in the U.S. The American Cancer Society estimated that 40,580 women died and another 215,990 women were diagnosed with breast carcinoma in 2004 alone.1 Screening mammography is viewed as an important means of detecting breast carcinoma early,2 and mammography utilization has increased over the past decade. Yet in 2002, nearly 25% of women age ≥ 40 years had not received a mammogram within the past 2 years,3 and nearly 40% of low-income women had never had a mammogram.4

As efforts are made to further increase screening, it becomes more important to understand the various barriers to receiving mammograms. Insurance and cost barriers are noticeably salient among concerns reported by women who do not receive mammograms. Studies repeatedly show that health insurance status and type are significantly predictive of mammography use,5–11 although these effects vary.12, 13 Cost has also been shown to be a barrier to mammography screening, and has been widely studied.14–22

However, to our knowledge little attention has been paid to considering the barrier of cost in the context of actual cost and coverage information.23 The actual cost of a mammogram varies considerably with insurance coverage. Women with private insurance or Medicaid are typically fully covered for the cost of a screening mammogram. Women insured by Medicare receive 80% coverage, leaving women responsible to pay approximately $30 out-of-pocket, depending on where they live. Women dually insured by Medicare and Medicaid or Medicare and private insurance typically do not have to pay anything. Uninsured women are responsible for the full cost or must take advantage of reduced cost programs such as the Breast and Cervical Cancer Control Program.

Data from a population of low-income and minority women who were patients of a rural health clinic in North Carolina and who had not received a mammogram in at least the past year were used to explore the issues of coverage and cost as possible reasons for nonadherence to mammography guidelines. These data included a number of questions regarding cost and insurance coverage in relation to mammography use. The results reported here have important implications for providers and policymakers attempting to promote the use of screening mammograms among women age > 40 years who are in need of a mammogram.

MATERIALS AND METHODS

Data Source

The data for the current study are from baseline interviews performed from 1998 to 2000 as part of the Robeson County Outreach, Screening and Education (ROSE) Project, a randomized, controlled trial of an intervention designed to improve breast carcinoma screening rates among rural, poor, and minority women in Robeson County, NC. The Robeson Health Care Corporation (RHCC) is one of the major health care providers in the county, with four sites funded through a Community Health Center program from the Bureau of Primary Care. Lumberton Radiological Associates (LRA) was the primary mammography provider in Robeson County, as well as for the Breast and Cervical Cancer Control Program (BCCCP). Participants in this study used RHCC services. Women who had been patients of the participating health clinic for at least the past 2 years were randomly selected on the basis of being age ≥ 40 years and of not receiving a mammogram in the last year. Chart review and follow-up identified 1503 women who were eligible for the study, and 901 women were interviewed at baseline for a participation rate of 81%. Four women were found to have had a recent mammogram and were not included further in the study, leaving 897 women for analyses. Written informed consent was obtained from all women, and the current study was approved by the institutional review boards at Wake Forest University (Winston-Salem, NC) and The Ohio State University (Columbus, OH).

Key Variables

Face-to-face interviews asked women about their perceptions of cost and coverage of screening mammography. Six questions directly addressed the issues of cost and coverage of screening mammography: 1) What are the most important reasons why you have not had a mammogram? 2) Do you agree or disagree with the statement, “I can only afford to go to the doctor when I am sick, not for tests,” 3) About how much does a routine screening mammogram cost in your community? 4) How much does your insurance/Medicare/Medicaid pay towards a screening mammogram? 5) How much would you have to pay out-of-pocket for a mammogram? 6) Does the cost of a mammogram make it hard for you to get one? These questions focused on the monetary cost of the mammogram itself and not on associated costs, such as for transportation or child care. Response rates for all questions were high, with at most four respondents failing to answer any one of the questions studied.

Statistical Analysis

Basic descriptive statistics were computed to describe the demographics of the study population. Chi-square tests of homogeneity were performed to investigate bivariate relations between the independent and dependent variables related to cost. When the chi-square test revealed significant differences (P ≤ 0.05), standardized residuals were also calculated to identify the cells responsible for the difference. Odds ratios (OR) and 95% confidence intervals (95% CI) for multivariable associations were calculated using logistic regression analysis. Performance of the logistic regression models was assessed using the Hosmer–Lemeshow goodness-of-fit test to assess lack of fit and the area under the receiver operating characteristic (ROC) curve to assess model discrimination.24 Dummy variables were created as appropriate for categorical variables. Age was included in all models as a categorical variable, rather than as a continuous variable. Multinomial logistic regression analysis was used when the outcome variable had more than two response levels.24 STATA software (version 8.0; Stata Corporation, College Station, TX) was used to conduct these analyses. All other analyses were performed using SAS version 8.2 software (SAS Institute Inc., Cary, NC).

RESULTS

Characteristics of Study Population

The study population (n = 897) reflects the composition of Robeson County, NC, which includes the largest concentrated Native American population in the Eastern U.S. and a large African-American population. This tribe is not federally recognized and therefore its members do not receive health benefits from the Bureau of Indian Affairs. Furthermore, it has an overwhelmingly low-income population, with 71% reporting an annual household income < $20,000. Demographic characteristics are shown in Table 1. The majority of women in the study population had some form of health insurance, although 27.8% had none. Medicaid was the sole source of insurance for 12.8% of these women, and another 12.7% had both Medicaid and Medicare. Nearly 6% had Medicare and private insurance, whereas 10.6% had Medicare only. Medicare Health Maintenance Organizations were not available in Robeson County. Moreover, 30.1% of the study population reported having only private health insurance.

Table 1. Demographic Characteristics and Health History of Study Participants, by Group (n = 897)
CharacteristicsNo. of patients (%)
Age (yrs) 
 40–49380 (42.4)
 50–64327 (36.5)
 ≥ 65190 (21.2)
Race 
 African-American295 (32.9)
 Native American371 (41.4)
 White225 (25.2)
 Other5 (0.6)
Income 
 < $20,000638 (71.1)
 $20,000–$30,000131 (14.6)
 $31,000–$40,00057 (6.4)
 $41,000–$50,00025 (2.8)
 > $50,00015 (1.7)
Insurance status 
 Medicaid only115 (12.8)
 Medicare/Medicaid114 (12.7)
 Medicare/private52 (5.8)
 Medicare only95 (10.6)
 Private only270 (30.1)
 No insurance249 (27.8)
Education 
 ≤ 8th grade182 (20.3)
 Some high school213 (23.8)
 High school graduate502 (56.0)
Working status 
 Full/part time377 (42.0)
 Other520 (58.0)
Marital status 
 Married/living together400 (44.6)
 Other497 (55.4)
Health history 
 Family history of breast carcinoma 
 Yes239 (26.6)
 No658 (73.4)
Regular checkup in past 12 mos 
 Yes622 (69.3)
 No275 (30.7)

Perceived Coverage of Screening Mammography

Perceptions of insurance coverage were defined as “appropriate,” an “underestimate,” or an “overestimate” for women who reported their mammography coverage to be at, below, or above the level of actual coverage for their type of insurance (100% for Medicaid, Medicare/Medicaid, Medicare/private, and private only; 80% for Medicare only; and 0% for no insurance). For example, a woman whose health insurance was provided through Medicare alone and who reported that her insurer covered 100% of the cost of a mammogram would be classified as an overestimate, because Medicare actually pays 80% of the cost of a mammogram. Women could give either the percentage or dollar amount they believed their insurance would pay for a mammogram. Dollar amounts were converted to percentages using the actual cost of a mammogram at Lumberton Radiological Associates at the time of the study ($60 in 1998) for comparison to actual coverage levels. These estimates were focused on coverage for the actual test itself and did not factor in annual deductibles. To be considered appropriate, women's estimates were required to exactly equal their actual level of coverage. The underestimate and overestimate categories were defined as being less than and greater than these coverage levels, respectively. All women in the overestimate category had overestimated their insurance coverage by ≥ 10 percentage points, whereas 99.5% of women in the underestimate category had underestimated their coverage level by ≥ 10 percentage points.

Figure 1 shows the appropriateness of women's perception of insurance coverage classified by insurance category. Forty percent of women asked about their level of coverage for a screening mammogram were incorrect, having underestimated, overestimated, or reported that they did not know. Women who had an inappropriate perception of their level of coverage were more likely to have private insurance or to have Medicare alone. A significant proportion of women with Medicare alone (45%) overestimated their level of coverage. Women with an appropriate perception of their coverage were more likely to be covered only by Medicaid or to have no insurance. In addition, 14% of women reported that they did not know the level of coverage their health insurance provided for a screening mammogram (data not shown).

Figure 1.

Distribution of women's perception of insurance coverage for a screening mammogram, by insurance status (n = 892). Chi-square test of homogeneity: χ2(5) = 313.3; P < 0.0001. Inappropriate includes the categories of underestimate, overestimate, and does not know. Striped bars: inappropriate; dark bars: appropriate.

Cost and Perceived Coverage as a Barrier to Screening Mammography

Greater than half of the women studied (53%) identified cost as a barrier to receiving a screening mammogram, regardless of insurance coverage. Although women were able to name multiple reasons for not having had a mammogram, the most frequently noted reason was cost. Of women who had never had a mammogram, 20.1% named cost as a reason, whereas 25.8% of women with a previous mammogram but none within the past 2 years noted cost as a reason (Table 2).

Table 2. Reasons for Not Having a Mammogram within the Past 2 Years (n = 298) or Ever (n = 209)
 None within past 2 yrsNever
ReasonNo. of patients (%)No. of patients (%)
Cost77 (25.8)42 (20.1)
Other48 (16.1)28 (13.4)
Not necessary, no problems39 (13.1)36 (17.2)
“No reason”34 (11.4)27 (12.9)
Examination can be painful26 (8.7)5 (2.4)
My doctor never recommended it19 (6.4)22 (10.5)
Never thought about it16 (5.5)30 (14.4)
Forgot/missed appointment13 (4.3)6 (2.9)
I don't know where to go/transportation problems12 (4.0)6 (2.9)
Embarrassed/fear11 (3.7)28 (13.4)
Did not know about it/did not know I needed another one10 (3.4)5 (2.4)
Don't know/missing7 (2.3)2 (1.0)
Time constraints/too much bother or inconvenience5 (1.7)6 (2.9)
Insurance doesn't cover it/no health insurance5 (1.7)4 (1.9)
Lazy3 (1.0)2 (1.0)
Age2 (0.7)2 (1.0)
Someone recommended not having one1 (0.3)0 (0.0)
Examination causes cancer0 (0.0)1 (0.5)

Table 3 shows the ORs between factors associated with women stating that cost makes it difficult to obtain a mammogram. Women who either underestimated or did not know their level of insurance coverage for a mammogram had 4.5 times increased odds of stating that cost makes it hard to get a mammogram compared with women who had an appropriate perception of their coverage (OR = 4.57, 95%CI, 1.95–10.70 for the underestimate category; OR = 4.42, 95% CI, 1.80–10.88 for the don't know category). In contrast, women who overestimated their level of insurance coverage had significantly lower odds of stating that cost made it hard to get a mammogram than women who had an appropriate perception of their coverage (OR = 0.18, 95% CI, 0.06–0.56). The odds of women receiving Medicare noting cost as a barrier to screening mammography were increased 5-fold (OR = 5.34, 95% CI, 1.76–16.20), compared with privately insured women. In addition, women with no insurance had a nearly 19 times greater odds of stating that cost made it hard for them to get a mammogram compared with women with private insurance (OR = 18.69, 95% CI, 7.32–47.69).

Table 3. Associations between Independent Variables and Stating Cost Makes it Difficult to Obtain a Mammogram (n = 866)
 OR (95% CI)
  • OR: odds ratio: 95% CI: 95% confidence interval.

  • a

    P < 0.05.

  • b

    For $0 to >$50 categories Ptrend < 0.001.

Cost and coverage 
 Appropriateness of perception of insurance coverage 
  Appropriate1.00 (—)
  Underestimate4.57 (1.95–10.70)a
  Overestimate0.18 (0.06–0.56)a
  Don't know4.42 (1.80–10.88)a
 Insurance status 
  Private only1.00 (—)
  Medicaid only1.01 (0.44–2.29)
  Medicare/Medicaid1.01 (0.38–2.69)
  Medicare/private0.31 (0.08–1.20)
  Medicare only5.34 (1.76–16.20)a
  No insurance18.69 (7.32–47.69)a
 Out-of-pocket costb 
  $01.00 (—)
  $1–501.28 (0.60–2.75)
  >$508.22 (3.64–18.56)a
  Don't know2.35 (0.92–6.00)
Demographic characteristics 
 Age (yrs) 
  40–491.00 (—)
  50–640.61 (0.38–0.97)c
  ≥ 650.57 (0.25–1.30)
 Race 
  White1.00 (—)
  African-American0.41 (0.24–0.70a)
  Native American0.54 (0.33–0.90)c
  Other3.28 (0.40–27.14)
 Income 
  ≥$20,0001.00 (—)
  <$20,0001.73 (1.02–2.96)c
 Education 
  ≤8th grade1.00 (—)
  Some high school1.17 (0.65–2.13)
  High school graduate1.00 (0.59–1.68)
 Working status 
  Other1.00 (—)
  Full/part time0.90 (0.53–1.53)
 Marital status 
  Other1.00 (—)
  Married/living together0.81 (0.52–1.26)
Health history 
 Regular checkup in past 12 mos 
  No1.00 (—)
  Yes0.61 (0.40–0.94)c
 Ever had a mammogram 
  No1.00 (—)
  Yes0.81 (0.50–1.31)
 Doctor has recommended mammogram 
  No1.00 (—)
  Yes1.64 (1.08–2.47)c
 Level of worry about breast carcinoma 
  Not at all/slightly1.00 (—)
  Moderately0.84 (0.47–1.50)
  Quite a bit/extremely2.01 (1.20–3.37)a
 Perceived health 
  Fair/poor1.00 (—)
  Good0.75 (0.45–1.26)
  Very good/excellent0.78 (0.46–1.31)

Women's perceptions of their out-of-pocket costs for a mammogram were also significantly related to their assessments of cost as a barrier to screening mammography. Regardless of whether their perceptions were accurate, the more women reported they would have to pay, the more likely they were to say that cost made it hard for them to get a mammogram. Women who reported they would have to pay > $50 out-of-pocket for a mammogram had approximately 8 times the odds, compared with women who reported no out-of-pocket costs, of noting cost as a barrier (OR = 8.22, 95% CI, 3.64–18.56). At the time of the study, the charge for a mammogram was $60 at Lumberton Radiological Associates where the women received their mammograms. Therefore, reported out-of-pocket costs > $50 represented nearly the full cost of the mammogram.

Perceptions of Coverage and the Barrier of Cost

Appropriateness of perception of insurance coverage was significantly related to identifying cost as a barrier to mammography. Multinomial logistic regression was used to determine factors that were associated with appropriateness of perception of insurance coverage, using appropriate as the referent category (Table 4). Excluding subjects with missing data resulted in a sample of 861 women for the current analysis. The ORs for each significant independent variable that compared the odds of having one of three levels of appropriateness of perception of insurance coverage (underestimate, overestimate, or don't know) with the odds of having an appropriate perception of coverage, controlling all other independent variables included in the regression, are each presented in Table 4. Variables which were not significant in any of the levels of the outcome variable are not presented, and the significance of independent variables varied by the level of the outcome variable.

Table 4. Associations between Independent Variables and Appropriateness of Perception of Insurance Coverage for Mammography (n = 861)a
CharacteristicsAppropriate ORUnderestimate OR (95% CI)Overestimate OR (95% CI)Don't know OR (95% CI)
  • OR: odds ratio; 95% CI: 95% confidence interval.

  • a

    Only odds ratio for significant variables are shown. The variables included in all models included age (40–49 years, 50–64 years, and ≥ 65 years), race (white, African American, Native American), marital status (married/living together vs. other), educational level (≤ 8th grade, some high school, high school graduate), employment status (full/part time vs. other), income (≥ $20,000 vs. < $20,000), insurance status (private only, Medicaid only, Medicare/Medicaid, Medicare/private, Medicare only, no insurance), perceived health (fair/poor, good, very good/excellent), regular checkup in past 12 months (no vs. yes), ever had a mammogram (no vs. yes), number of times identified cost as a barrier to mammography (0, 1, ≥ 2).

  • b

    P ≤ 0.01.

  • c

    Cell count too small for estimate.

  • d

    P ≤ 0.05.

No. of times cost identified as a barrier to mammography    
 01.001.001.001.00
 11.002.52 (1.38–4.62)b0.42 (0.18–1.00)b1.46 (0.76–2.81)
 ≥ 21.008.89 (4.02–19.66)b0.06 (0.01–0.29)b12.43 (5.84–26.48)b
Insurance status    
 Private only1.001.001.001.00
 Medicaid only1.000.04 (0.01–0.10)bc0.10 (0.04–0.28)b
 Medicare/Medicaid1.000.07 (0.02–0.24)bc0.12 (0.04–0.39)b
 Medicare/private1.000.17 (0.05–0.60)bc0.31 (0.09–1.07)
 Medicare only1.000.03 (0.01–0.13)bc0.10 (0.03–0.37)b
 No insurance1.00cc0.00078 (0.00015–0.0042)b
Age    
 40–491.001.001.001.00
 50–641.000.82 (0.47–1.44)2.98 (0.81–10.95)1.13 (0.61–2.09)
 ≥ 651.000.55 (0.20–1.54)4.59 (1.03–20.46)d1.00 (0.39–2.56)
Race    
 White1.001.001.001.00
 African American1.000.51 (0.27–0.98)d1.19 (0.40–3.59)0.99 (0.51–1.92)
 Native American1.001.16 (0.64–2.11)0.95 (0.32–2.77)1.23 (0.65–2.35)
 Other1.000.18 (0.01–2.99)c0.91 (0.06–14.49)
Income    
 ≥$20,0001.001.001.001.00
 <$20,0001.001.43 (0.75–2.76)1.51 (0.34–6.74)2.45 (1.16–5.15)d
Ever had a mammogram    
 No1.001.001.001.00
 Yes 0.87 (0.47–1.60)0.76 (0.26–2.28)0.32 (0.19–0.56)b

Women who identified cost as a barrier to screening mammography at any point during the interview had increased odds of underestimating their level of insurance coverage (OR = 2.52, 95% CI, 1.38–4.62) compared with women who did not identify this barrier. Women who reported cost as a barrier multiple times had increased odds of underestimating (OR = 8.89, 95% CI, 4.02–19.66) or reporting not knowing (OR = 12.43, 95% CI, 5.85–26.48) their level of insurance coverage. Similarly, women who reported that cost was a barrier to mammography had significantly lower odds of overestimating their level of insurance coverage (OR = 0.42, 95% CI, 0.18–1.00 for once; and OR = 0.06, 95% CI, 0.01–0.29 for at least twice).

Insurance coverage was also significantly related to appropriateness of women's perceptions of the cost and coverage for a mammogram. Women with any type of public insurance had significantly lower odds of underestimating their level of insurance coverage than did women with private insurance (ORs ranged from 0.03 for Medicare only to 0.17 for Medicare/private). With the exception of women covered by both Medicare and private insurance, these publicly insured women also had significantly lower odds of reporting that they did not know their level of insurance coverage (ORs of approximately 0.10).

A number of other variables were additionally related to perceptions of insurance coverage. Women age ≥ 65 years had nearly 5 times the odds of overestimating their level of insurance coverage compared with women ages 40–49 years (OR = 4.59, 95% CI, 1.03–20.46). African-American women had significantly lower odds of underestimating their level of insurance coverage compared with whites (OR = 0.51, 95% CI, 0.27–0.98), but no differences were observed with other racial groups. Women having an annual income < $20,000 had greater odds (OR = 2.45, 95% CI, 1.16–5.15) of not knowing their level of insurance coverage. In contrast, women who had a mammogram in the past had decreased odds of reporting not knowing their level of insurance coverage (OR = 0.32, 95% CI, 0.19–0.56).

DISCUSSION

As efforts to increase the utilization of screening mammography among vulnerable women continue, it is essential to improve our understanding of the various barriers to receiving mammograms so that providers and policymakers can work to overcome them. The current study explores the issue of perception of coverage as it relates to cost as a barrier among women nonadherent to screening guidelines, and the results show that the cost barrier itself may be as much a knowledge barrier as a financial barrier.

Approximately 40% of women were incorrect or unsure of how much their insurance covered for a screening mammogram. Perceptions of insurance coverage varied by type of insurance arrangement, however. For example, nearly 80% of privately insured women had an inappropriate perception of their coverage. Furthermore, 53% of the women in the study population identified cost as a barrier to mammography. Therefore, the question of whether misperception of insurance coverage was related to the existence of a cost barrier to mammography was considered. Indeed, women's understanding of their coverage—even when inaccurate or incomplete—appears to affect their perceptions of cost as a barrier. Among our respondents, women who underestimated or did not know their level of coverage had 4.5 times increased odds of stating cost-related difficulty in getting a mammogram. Further, the more women reported they would have to pay out-of-pocket for a screening mammogram, the more likely they were to say that cost made it hard for them to get a mammogram, whether or not their perceptions were accurate.

Given that inaccurate perceptions of insurance coverage are related to reporting that cost makes it hard to get a mammogram, it is important to understand the characteristics that make a woman more likely to have such misperceptions. The current study found that women who identified cost as a barrier, women with very low incomes (< $20,000), and women age ≥ 65 years, had significantly increased odds of having misperceptions of their insurance coverage. Therefore, providing these women with information on actual levels of insurance coverage could serve to correct these misperceptions and thereby reduce the perceived cost barrier for these women.

Also important to policymakers is our finding that women insured by government programs had consistently lower odds of underestimating or not knowing their level of insurance coverage for mammography compared with privately insured women. Therefore, it appears that although holders of public health insurance are well informed of their coverage for mammography, private insurers may need to do a better job of communicating the levels of such benefits to their enrollees. In our study, one indication of the success of government health insurance programs is the finding that the majority of women who were insured by Medicaid alone had an appropriate understanding of their complete coverage of screening mammograms. Further, these women were no more likely to report that cost made it hard for them to get a mammogram than were women with private insurance, despite the finding that, by definition, Medicaid recipients would have lower incomes compared with women with private insurance. This provides evidence that understanding of insurance coverage for mammography can reduce the barrier of cost, even among women for whom cost would be expected to present a significant barrier.

Several study limitations should be noted. First, our cross-sectional study is limited to a population of largely low-income women from a rural area of the U.S. Although these results may be generalized across other low-income populations of women, certain demographic characteristics of this area such as a high proportion of Native Americans distinguish this study population. However, controlling for factors such as race and income in analyses has improved the robustness of the findings. Other limitations stem from reliance on self-reported insurance status. It was not possible to classify private insurance by type of plan, and thus it was not possible to obtain exact coverage rates. As private insurers with service in North Carolina generally provide 100% coverage for screening mammography, this was assumed to be the level of coverage for women with private insurance, although variability is possible. Finally, the study population consisted only of women who were currently noncompliant with mammography screening guidelines, as one goal of the ROSE project was to understand reasons for nonadherence. Thus, it cannot be stated that women who receive mammograms have a more accurate perception of cost and coverage for mammography. It can be concluded, however, that among women in need of a mammogram, cost presents an important barrier and misperceptions of insurance coverage may contribute to the existence of this barrier.

Because greater than half of a population of low-income women in need of a mammogram reported that cost had made it hard for them to obtain a mammogram as recommended, uncovering ways to reduce this barrier is critical. Misinformation regarding coverage and cost appears prevalent, however, regardless of insurance type. Women who identified cost as a barrier were much more likely to believe they had less coverage than they actually did, or to report that they did not know how much of the cost of the screening test was covered.

Successful interventions among both rural and urban populations have invested in options to reduce or remove cost as a barrier to screening mammography.16, 18, 19, 25–27 These results, though, suggest that improving women's knowledge about the actual out-of-pocket costs and insurance coverage for screening mammograms may reduce the overall impact of cost as a barrier. Although one cannot conclude, on the basis of these results, that providing women with accurate cost information will increase compliance with mammography screening guidelines, it does seem likely that appropriate information can help. Efforts to disseminate accurate information regarding screening costs and coverage may help providers and policymakers reduce the financial barrier to screening mammograms—especially when the barrier itself is related to inaccurate knowledge rather than to actual costs.

Ancillary