We found that in the context of a reminder program to encourage repeat mammograms among insured patients, important barriers to examination completion remain. In particular, younger age, more recent health plan membership, lower family income, and obesity significantly reduced the odds of completing a mammogram. Our findings are consistent with findings from other studies,10,13,15,33
including the finding that family income (and associated factors) appears to be a stronger predictor than race. However, this study serves to highlight that access to mammograms through health insurance and reminding patients that mammograms are due does not completely alleviate patient barriers. Disparities cannot be addressed solely by providing access to clinician advice about the procedure. Thus, to improve mammogram screening rates among the insured, it will be important for health plans to expand beyond reminding women to obtain mammograms.
In another study at this site that used an alternative method of assessing compliance with mammography guidelines, the Prevention Index (PI), or the proportion of time a patient is observed during which the patient has had the prevention services needed or is appropriately “covered according to guidelines,” those aged 50–59 (compared to those 60–69) and those with shorter periods of health plan membership had lower mammography PIs.33
The current study expands on these findings by exploring patient-reported factors that could explain the lower likelihood of mammography completion among younger women. Younger targeted women were significantly more likely to report being “too busy” to have mammograms. Although we did not explore patient perceptions of how this barrier could be alleviated, it may be useful to evaluate such interventions as worksite and mobile mammography facilities or the availability of after-work appointments for younger women. Younger targeted women also had more worries about mammograms than older women. The worries largely related to concerns about the accuracy of mammograms that could cause either missed cancers or unnecessary surgery. Younger women's perceptions in this regard have some basis in fact because although the USPSTF recommends mammograms every 1–2 years among women aged 50–69, the balance of benefits and harms grows more favorable as women age.4
This highlights the need for assistance with informed decision making in younger women and the continued need for improved technologies for breast cancer screening.
Obesity has been described previously as a barrier to mammogram completion.16,34
This study helps confirm the strength of this association by using objective measures, such as height and weight and mammography examination data. We also elucidated several issues related to the obesity barrier. Most importantly, we found that the prevalence of this barrier was very high; nearly half (47%) of this community-based sample of insured women who had already had a prior mammogram and had been weighed were obese. The findings in our sample are not dissimilar to those of the National Health and Nutrition Examination Survey (NHANES) for 2005–2006. There, 42% of women aged 40–59 in the United States were found to be obese.35
The somewhat higher prevalence of obesity found here may be explained by the fact that the state of Oregon has a higher prevalence of obesity among adults than the national average.36
Given the obesity epidemic,37
the higher incidence and mortality from breast cancer among the obese,34
and the need for patients to participate in regular screening to achieve desired reductions in mortality,8
obesity is an exceedingly common and important barrier to mammography. Therefore, it is crucial to understand what, specifically, about obesity is causally linked to lower rates of mammography screening completion among these women.
Obese women were significantly more likely to report “too much pain” from mammograms when compared to the nonobese. Other studies have shown that up to 35% of all women receiving mammograms complain of pain (24.7% did so in this study), and that pain serves as a deterrent to repeating mammograms among all women.38
We could only identify one small study that evaluated the relationship of obesity to pain from mammograms, and this study did not find that obese women reported more pain with mammograms.39
The reason obese women may more frequently report pain remains uncertain. Mammogram-related pain in the Sharp et al. study39
was unrelated to breast size or reported sensitivity to pain in general. At least one study suggests that obesity may be associated with a lowered pain threshold.40
The null findings in our study about obesity and mammography may be as important as the positive ones. Other studies have hypothesized that the adverse effect of obesity on mammogram screening relates to a reduced focus on preventive healthcare among obese women, discriminatory attitudes or counseling by healthcare providers, and embarrassment.16
Consistent with the findings of Ferrante et al,34
we did not find differences among the obese and nonobese in the reporting of these factors. In our survey, obese respondents were as likely as nonobese respondents to report that their PCPs had recommended a mammogram, and they were as likely to have received various counseling components. Thus, clinician-oriented interventions likely cannot, by themselves, fully address the complexity of patient barriers to mammography completion among the obese. Patients in general and obese patients in particular may benefit from interventions to reduce pain, such as patient-controlled compression38
or, ultimately, the use of alternative technologies for breast cancer screening.
This study has several limitations. It was conducted at a single HMO, so the findings may not be generalizable to other populations. Because members of the HMO where the study was conducted closely resemble the general community, however, our findings are likely to be relevant to larger populations. In the cohort analyses, individual race data were not complete, and missing race data and all family income data were from neighborhood estimates (geocoded). These data may not be as accurate as those derived from self-report. The phase 2 (survey group) results need to be interpreted with caution. This group was not selected to be a representative sample of the population; rather, they were selected in a manner to oversample those with barriers in order to further elucidate them. Thus, the specific frequency of barriers found in this subgroup would not likely be the same as that found in the general population. Conclusions comparing those with and without barriers (e.g., pain in the obese vs. nonobese) should be generalizable, however. We did not have the power in phase 2 to examine patient-reported barriers among income and membership subgroups. Also, we could not assess obesity subgroups by race in either phase of the study because of limited sample size. Finally, only 50% of patients responded to the survey, which could lead to response bias.