This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Calvocoressi, L. Articles by Jones, B. A. Search for Related Content PubMed PubMed Citation Articles by Calvocoressi, L. Articles by Jones, B. A.

A Prospective Study of Perceived Susceptibility to Breast Cancer and Nonadherence to Mammography Screening Guidelines in African American and White Women Ages 40 to 79 Years

¹ Department of Epidemiology and Public Health and ² Department of Diagnostic Radiology, Yale University School of Medicine, New Haven, Connecticut

Requests for reprints: Beth A. Jones, Department of Epidemiology and Public Health, Yale University School of Medicine, 60 College Street, New Haven, CT 06520. Phone: 203-785-2890. E-mail: beth.jones{at}yale.edu

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
This prospective study examined the influence of perceived susceptibility to breast cancer on nonadherence to recommended mammography screening guidelines. The study population included 1,229 African American and White women ages 40 to 79 years who obtained an index mammography screening examination at one of five urban hospitals in Connecticut between October 1996 and January 1998. Information on perceived susceptibility to breast cancer and on multiple covariates was obtained by telephone interview on average 1.5 months after the index screening. Subsequent adherence to mammography screening guidelines was ascertained by follow-up interview on average 29 months after the index exam. Across race, age, and family breast cancer history, women who believed that their susceptibility was high (i.e., "very likely" to develop breast cancer) were less likely to adhere to screening guidelines than women who believed that their susceptibility was moderate [adjusted odds ratio (OR), 2.83; 95% confidence interval (CI), 1.51-5.30], but the effect was stronger in older women. Women ages 40 to 49 years (but not ages 50-79 years) who believed that their susceptibility was low (i.e., "not likely" or "a little likely" to develop breast cancer) were also less likely to adhere to guidelines than those who reported moderate susceptibility (adjusted OR, 3.07; 95% CI, 1.66-5.68, and adjusted OR, 2.78; 95% CI, 1.63-4.73). In contrast to most previous studies that found a positive linear relationship between perceived susceptibility to breast cancer and mammography screening, these findings suggest a more complex relationship that should be considered when developing interventions to improve adherence to mammography screening guidelines.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Women who adhere to recommended mammography screening guidelines may bring smaller and more treatable tumors to medical attention and may experience less mortality from breast cancer than those who do not participate in regular screening (1). However, whereas most women ages 40 years have had a least one mammogram, the proportion who screen at annual or biennial intervals, in accordance with recommended guidelines, is suboptimal (2, 3). Because the efficacy of screening is enhanced by adherence to a regular screening schedule (1), identifying and intervening on factors that specifically promote adherence (rather than one-time use of mammography) will more effectively decrease disease burden.

Perceived susceptibility to disease is a factor central to many theories of health behavior and is used to explain cancer screening behavior and to inform interventions to promote screening (4). In two meta-analytic reviews (5, 6) greater perceived susceptibility to breast cancer was positively associated with several mammography screening outcomes. Accordingly, interventions aimed at increasing women's perceptions of their susceptibility to breast cancer have frequently been used to increase this screening behavior (7, 8).

Among the studies included in the two meta-analytic reviews (5, 6), however, only a minority examined the effect of perceived susceptibility on adherence to screening guidelines. Additional studies that specifically examine this outcome are needed to establish whether perceived susceptibility bears the same relationship to adherence as to one-time mammography use. Furthermore, most previous studies used cross-sectional designs. Given the potential for a woman's screening behavior to influence self-reported perceptions of susceptibility in cross-sectional analyses (a phenomenon often described as "reverse causation"), prospective assessment of perceived susceptibility in relation to adherence is needed to clarify the causal relationship between these variables. Moreover, whereas there is some evidence that the influence of perceived susceptibility on adherence may differ by race/ethnicity (9), age (10), and family breast cancer history (11), the potential modifying effects of these variables have not been examined in prospective studies of community-based samples. Identifying potential differences in the effect of perceived susceptibility on adherence to screening guidelines in these population subgroups may aid in developing more refined and effective intervention strategies.

From the study Race Differences in Screening Mammography Process,³ we report on the effect of perceived susceptibility to breast cancer examined at baseline and analyzed in relation to nonadherence to mammography screening guidelines assessed at 2-year follow-up, thus overcoming some of the methodologic limitations of prior work. The cohort included White and African American women, the two race/ethnic groups in the United States with the highest breast cancer incidence and mortality rates, respectively (12). In addition to women ages 50 to 79years, this study included women ages 40 to 49 years who are often excluded from studies of mammography screening. Women with and without a family breast cancer history were also included. These subgroups were represented in sufficient numbers to conduct analyses of their potential modifying effects on the relationship between perceived susceptibility and nonadherence.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Study Population
African American and White women ages 40 to 79 years, who presented for a screeningmammogram (hereafter referred to as the "index screening") between October 1996 and January 1998 at five hospital-based mammography facilities in the Connecticut cities with the largest African American populations, were eligible for enrollment. Based on the results of a statewide survey (13) conducted in preparation for this study, these facilities were selected because they served a relatively high proportion (over 20%) of African Americans, thus ensuring adequate numbers of this race/ethnic group. To ensure a relatively equal distribution of participants by race, all eligible African Americans who obtained screening mammograms at these sites were selected for inclusion. An equal number of Whites, randomly selected and frequency matched to the African Americans on hospital and date of mammogram, were also invited to participate.

Because the study focused on assessment of screening behavior, women who obtained diagnostic or follow-up mammograms to assess a breast problem were ineligible, as were women who reported prior breast problems and/or histories of breast cancer, cyst aspiration, and breast biopsy. In accordance with age recommendations for regular mammography screening in the general population (14, 15), women younger than age 40 were not included. In addition, those older than age 79 were excluded to minimize potential recall problems that are assumed to be more common among the elderly, and to limit age-related illnesses and complications that might compromise data collection and adversely influence adherence to mammography screening regimens.

Procedures
Approvals of the institutional review boards of Yale University School of Medicine and each participating hospital were maintained throughout the study period. Study staff reviewed intake sheets of women presenting for mammograms at participating facilities to identify potentially eligible subjects who were then mailed letters of invitation that included a study information sheet and a written consent form to obtain mammography records. Trained interviewers administered a baseline telephone interview to women who gave oral consent and met eligibility criteria. Written consent to review mammography records was returned by mail after the baseline interview. Baseline telephone interviews were conducted 1 to 6 months after the index screening mammogram (mean, 1.5 months; SD ± 0.85 month). Women who completed the baseline interview were contacted 26 months after the index screening to arrange a follow-up interview. The time interval between baseline and follow-up interview averaged 29.4 months (SD ± 1.42 months), with a range of 27 to 41 months.

Study Participation
Recruitment, eligibility, and retention of participants in this study are summarized in Fig. 1, along with additional exclusion criteria specific to this analysis. Of 2,359 potentially eligible women identified from mammography facility intake sheets, 1,982 women were contacted and eligible to participate. Of these, 531 declined, leaving 1,451 (73.2%) who completed a baseline telephone interview. Participation differed significantly across race group (African American, 69%; White, 77%; P < 0.0001) and marginally by age (ages 40-49 years, 76%; 50 years, 72%; P < 0.0517).

View larger version (24K): [in this window] [in a new window]   Figure 1. Subject eligibility and retention.

Variables and Measures
Baseline and follow-up interviews were developed from (a) consultation with radiologists and technicians who provide mammograms to ethnically diverse groups, (b) pertinent issues identified in a focus group of African American women; and (c) the published literature (13, 16-22).

The main predictor, perceived susceptibility to breast cancer, was assessed during the baseline interview with the following question: "How likely do you think is it that you will get breast cancer during your lifetime?" Response categories included (a) not at all likely, (b) a little likely, (c) somewhat likely, (d) very likely, and (e) don't know.

The outcome variable, nonadherence to mammography screening guidelines, was assessed at follow-up, based on American Cancer Society screening guidelines in effect in 1996, at the onset of the data collection period of this study (23). Women ages 40 to 49 years were considered nonadherent if they did not obtain at least one screening within 2 years (+ 2 months) of the index exam. Women ages 50 years were considered nonadherent if they did not obtain at least two screenings within 2 years (+ 2 months) of the index exam. The "+ 2 months" allowed for potential inaccuracies in recall of self-reported dates and for reasonable delays in obtaining a screening appointment. Women who obtained diagnostic or follow-up mammograms for a breast abnormality in the interval between screenings were included in this analysis; however, the diagnostic and follow-up mammograms they obtained were excluded from assessment of adherence to screening guidelines.

For 1,126 respondents (92%), the outcome was determined by self-report, in keeping with the original study design. The remaining 103 women (8%) did not provide sufficient self-reported information to ascertain the outcome (i.e., could not recall the month or year of at least one screening) but did consent to review of mammography records. We included these women in this analysis and used records to determine outcome status. These 103 women did not differ from respondents for whom self-reported data were used on recruitment site or family breast cancer history, but they were more likely to be African American (55% versus 38%, P < 0.0005) and to be age 50 (78% versus 63%, P < 0.0025). In a random sample of150 study participants (identified from a single facility), there was substantial agreement (83%; = 0.656) between self-report and medical record outcome data, with no significant difference in agreement between these data sources by race, age, or family history of breast cancer.

A set of primary covariates (sociodemographic, access to medical care, and mammography-related factors) were included in all adjusted analyses. Sociodemographic variables included (a) self-identified race/ethnicity (African American versus White), (b) age (40-49 versus 50 years), (c) marital status (married/living together versus other), (d) education (<12 years, 12 years, >12 years), (e)annual family income (<$15,000; $15,000-49,999; $50,000), and (f) number of household occupants(1, 2, 3, 4 or more). Access to care variables, coded yes versus no, included (a) full, annual mammography insurance coverage and (b) having the same (i.e., usual) health care provider over the past year. Mammography-related factors included (a) knowledge of age-specific mammography screening guidelines (correct versus not correct), (b) receipt of a health care provider's recommendation to obtain a mammogram subsequent to the index screening (yes/no), (c) history of breast cancer in a first- or second-degree relative (yes/no), and (d) history of nonadherence to mammography screening guidelines prior to the index screening (adherent versus nonadherent). The latter was calculated based on the respondent's age and available data on the number of lifetime mammography screenings she reported (0, 1, 2,3, 4, or 5+). Women ages 40 to 49 years were considered nonadherent if they did not obtain at least one screening every 2 years. Women ages 50 years were considered nonadherent if they did not obtain at least 5 screenings.

A wide range of potential confounders of the association between perceived susceptibility and nonadherence was also examined (breast cancer risk factors, health status and behaviors, practical barriers to screening, features of the index screening, and psychosocial factors). Breast cancer risk factors included (a) age at menarche (<12, 12-14, >14 years), (b) age at first birth (nulliparous, first birth at age <30, first birth at age 30); (c) menopausal status (completed, not completed, hysterectomy prior to menopause), (d) use of hormone replacement therapy (never, <1, 1-5, >5 years), (e) height (in inches), and (f) body mass index. Health status and behaviors included (a) self-rated health (excellent, good, fair, poor), (b) pack-years of smoking, (c) physical exercise at least once per week (yes/no), (d) consumption of alcoholic beverages in the year prior to the index screening (yes/no), (e) regular use of vitamin supplements (yes/no), and (f) attendance at religious services (never, once/twiceper year, once every few months, 1-3 times per month, once a week, more than weekly). Practical barriers to screening included (a) travel time to index screening appointment (0-15 versus >15 minutes), (b) special arrangements (e.g., childcare) needed to attend the screening appointment (yes/no), and (c) took time off from work to attend the screening appointment (yes with pay, yes without pay, no). Features of the index screening included (a) hospital where index screening was obtained, (b) whether the index screening was obtained in clinic or mobile van, (c)time to receipt of index screening results (notreceived, 0-1, 2-7, 8-14, 15-28, >28 days), (d) result of the index screening (abnormal versus normal), and (e) whether the index examination was the first screening ever obtained (yes/no). Psychosocial factors included (a) perceived benefits of screening (mammograms very useful versus somewhat/a little/not useful/don't know), (b) embarrassment experienced during the index screening (none versus a little/somewhat/a lot), (c) pain experienced during the index screening (none versus a little/somewhat/a lot), (d) pain experienced during the index screening compared to expectations (greater than, less than, equal to, no expectations), (e) anxiety experienced during the index screening (none versus a little/somewhat/a lot), (f) anxiety experienced during the index screening compared to expectations (greater than, less than, equal to, no expectations), (g) worry over the results of the index screening upon leaving the mammography facility (none, a little/somewhat, a lot), (h) effect of the index screening on breast cancer worry (increased, decreased, no effect), (i) confidence in one's ability to obtain a future mammogram (very confident versus somewhat/a little/not confident/don't know), (j) perceived control over remaining healthy (a lot/some versus a little/none), (k) perceived control over developing cancer (a lot/some/a little, none, don't know), (l) perceived control over recovering from cancer (a lot/some/a little, none, don't know), and (m) received a reminder notice to obtain a mammogram in the 2 years since the index screening (yes/no).

Statistical Analysis
Preliminary analyses involved characterizing the study sample using descriptive statistics. Bivariable analyses examined the unadjusted associations between nonadherence to screening guidelines and (a) perceived susceptibility and (b) each of the primary covariates. The association between perceived susceptibility and nonadherence was then examined in multivariable models in the total sample, adjusted by the primary covariates and additional confounders. The primary covariates were added first and retained in subsequent multivariable models. All additional potential confounders were next added to the model as a set. Then, following the procedure described by Kleinbaum and colleagues (24), we selected a reduced model that provided adjusted estimates of the relationship between perceived susceptibility and nonadherence equivalent to that obtained when all additional potential confounders were included, thus providing complete adjustment by all additional confounders in a more parsimonious model. The potential modifying effects of race, age, and family breast cancer history on the adjusted relationship between perceived susceptibility and nonadherence were then assessed by adding two-way interaction terms between each category of perceived susceptibility and each potential effect modifier to the reduced model.

These bivariable and multivariable analyses used logistic regression to generate maximum likelihood estimates of odds ratios (OR) with 95% confidence intervals (95% CI) and two-sided P values. For ease of interpretation, we selected the category of perceived susceptibility with the lowest nonadherence (the "somewhat likely" category) as the reference group. For significant two-way interactions of perceived susceptibility by race, age, andfamily breast cancer history, ORs and 95% CIs for each subgroup were calculated by summing the relevant estimates and calculating the respective variances of these sums. Multicollinearity was ruled out with Spearman andPearson's correlation coefficients and with multivariable multicollinearity diagnostics (25). In multivariable models, values for missing data were multiply imputed using a data augmentation algorithm (26). A comparison of multiply imputed and complete case models yielded substantively similar results.

To assess the robustness of the findings, we conducted a time-to-event analysis to determine if an alternative mammography screening outcome, time to first screening since the index exam, would bear the same relationship to perceived susceptibility as nonadherence, adjusted by the same covariates used in the reduced model above. Hazard ratios and 95% CIs were generated for this analysis using Cox regression. All analyses were carried out with SAS software, version 8.2 (SAS System for Windows, SAS Institute Inc., Cary, NC).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Characteristics of the Study Population
As shown in Table 1, sizeable proportions of African Americans and women ages 40 and 49 years were included in the study population; the remaining respondents were White and were ages 50 to 79 years. Most respondents were married or living with a partner and had at least a high school education. Less than half had annual family incomes of $50,000 or more. The majority retained the same health care provider over the year prior to the baseline interview, received a provider's recommendation to obtain a mammogram during the 2 years subsequent to the index screening, reported full, annual insurance coverage for screening mammograms, and correctly identified the mammography screening guidelines for their age group. Nearly a third reported a first- or second-degree relative with breast cancer. Less than 20% of respondents reported a history of nonadherence to mammography screening guidelines prior to the index screening (measured at baseline). In contrast, 48% did not adhere to mammography screening guidelines subsequent to the index screening (i.e., the outcome variable, measured at follow-up). However, these measures are not directly comparable; the outcome variable was based on detailed assessment of all mammograms obtained subsequent to the index screening, whereas measurement of history of adherence was constrained by available baseline data (i.e., 0, 1, 2,3, 4, and 5+ lifetime mammograms) and likely underestimates nonadherence.

Multivariable Logistic Regression Analysis
Table 3 displays the effect of perceived susceptibility on nonadherence in the total sample, adjusted by the primary covariates and additional potential confounders. Compared with those who reported moderate susceptibility, the odds of nonadherence were no longer significantly elevated among those who indicated that they did not know how susceptible they were, once this association was adjusted by the primary covariates. However, the odds of nonadherence among women who reported that their susceptibility to breast cancer was low or high did remain significantly elevated after adjustment by these covariates. Including numerous additional potential confounders in the model did not substantially impact the association between any category of perceived susceptibility and nonadherence. In a reduced model that provided estimates nearly identical to the model that included all primary covariates and additional potential confounders, the largest OR was observed for women who reported that they were very likely (versus somewhat likely) to develop breast cancer (adjusted OR, 2.83; 95% CI, 1.51-5.30). The adjusted odds of nonadherence among women who reported that they were not likely or a little likely to develop breast cancer were also elevated compared with the somewhat likely category (OR, 1.51; 95% CI, 1.07-2.12, and OR, 1.64; 95% CI, 1.20-2.23, respectively).

View larger version (19K): [in this window] [in a new window]   Figure 2. Adjusteda ORs and 95% CIs for nonadherence to mammography screening guidelines by perceived susceptibility to breast cancer,b Connecticut, 1996-2000. aAdjusted by race, marital status, education, income, household size, mammography insurance, usual health care provider, knowledge of screening guidelines, received a recommendation from a health care provider to obtain a mammogram, family breast cancer history, history of adherence to screening guidelines, menopausal status, self-rated health, pack-years of smoking, pain experienced at index screening compared to expectations. bThe "don’t know" category of perceived susceptibility is excluded from this figure. cOR. d95% CI.

Similar to the nonadherence analysis, in the total sample believing that one was "very" compared with "somewhat" likely to develop breast cancer had an adverse effect on screening (adjusted hazard ratio, 0.67; 95% CI, 0.48-0.92) as did believing that one was "not" or "a little," compared with "somewhat" likely to develop this disease (adjusted hazard ratio, 0.75; 95% CI, 0.63-0.89, and adjusted hazard ratio, 0.77; 95% CI, 0.66-0.90, respectively). Also similar to the nonadherence analysis, two-way interactions between age and the (a) very likely; and (b) don't know categories were not significant. The interaction of age and the "a little likely" category was significant (P < 0.0346) and the interaction of age and the "not likely" category was P < 0.0819. Although not quite as robust as the comparable interaction terms in the nonadherence analysis, the effects were similar; these categories of perceived susceptibility were more strongly associated with less frequent screening in women ages 40to 49 years than in women ages 50 years.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

 
In this study, women who reported moderate susceptibility to breast cancer were most likely to adhere to mammography screening guidelines. High perceived susceptibility adversely influenced adherence in the total sample, but its influence was stronger among women ages 50 years. In contrast, the adverse effect of low perceived susceptibility was found only in women ages 40 to 49 years. These effects of perceived susceptibility were found in relation to time to first screening as well as to nonadherence.

Miller (27) proposed that health beliefs, such as perceived disease susceptibility, have less influence on a health behavior once that behavior is habitually practiced. One reason for the stronger effect of low perceived susceptibility on women ages 40 to 49 years than on women ages 50 years may be that mammography screening was a relatively new behavior and may not yet have become a habit for many of the younger women. In addition, during the data collection period of this study, controversy was rife over the efficacy of mammography screening in women ages <50 years. In January 1997, the majority report of a divided National Institute of Health Consensus Development panel concluded that universal screening recommendations for women ages 40 to 49 years were unwarranted and that each woman, with available information on risks and benefits, should decide for herself whether to obtain screening mammograms (28). Although most medical and preventive health organizations (e.g., American Cancer Society and National Cancer Institute) implemented annual screening recommendations for all women ages 40 years soon thereafter (29), the panel'shighly publicized report may have discouraged screening among women with low perceived susceptibility in the 40 to 49 years age group.

As noted, although its effect was stronger among older women, high compared with moderate perceived susceptibility was associated with greater nonadherence in the total sample. This is a departure from most previous findings and, in accordance with previous work, a recent meta-analytic review of 13 studies (6) found that across screening outcomes (e.g., intention to screen, one-time screening, and adherence to guidelines) perceived susceptibility was positively associated with screening. However, three of six studies in that review that specifically measured adherence found a negative association with perceived susceptibility (30-32), consistent with our finding. It may be that the adverse effect of high perceived susceptibility is a more salient factor in studies that examine regular screening behavior.

However, several characteristics of previous studies may have decreased the likelihood of detecting an adverse effect of high perceived susceptibility. For example, most studies used cross-sectional designs. In those studies, some women who did not screen according to guidelines may have minimized their perceived susceptibility to breast cancer as a rationale for nonadherence. Our prospective analysis eliminated the potential for such reverse causation. In addition, in a recent study (33) that found a curvilinear relationship between adherence and breast cancer worry (a construct associated with perceived susceptibility), the investigators attributed their unique finding to the large sample (n= 6,512) of the study. These investigators speculated that prior studies of average-risk samples did not include sufficient numbers of severely worried women to discern a negative effect of worry on adherence. In our analysis (N = 1,229), only a small minority of women (n = 60) considered themselves very susceptible to breast cancer, suggesting that a similar phenomenon may be operative in studies of adherence and perceived susceptibility. Moreover, most prior studies included samples that predominantly consisted of White women. The substantial number of African Americans in our analysis increased the size of the very susceptible category, which, in turn, increased the likelihood of detecting a significant association in a sample of this size. Of the 60 women who composed the very susceptible group, 35 were African American (7.3% of the African American sample) and 25 were White (3.3% of the White sample). More important, though, the relationship of high perceived susceptibility to nonadherence held for both race groups.

If our results are replicated in large, prospective studies, the implications for intervention are substantial. If the optimal level of perceived susceptibility to breast cancer needed to maintain adherence is moderate rather than high, interventions that increase perceived susceptibility beyond this optimal level may result in less, rather than more, screening. Population-level interventions (e.g., public health media campaigns, community education programs, and medical education literature) that provide a uniform message aimed at increasing awareness of breast cancer risk may not, therefore, be effective in promoting adherence in all women. Avenues such as telephone counseling and interactions with health care providers may be more efficacious; differences in the effect of perceived susceptibility (e.g., by age) can be considered and individual perceptions of susceptibility can be elicited and addressed in the context of other interrelated and important motivators of health behavior (e.g., beliefs about breast cancer, individual resources, and the utility of prevention strategies; ref. 34).

Furthermore, the relationship between perceived susceptibility and "actual" risk of breast cancer should be considered. Information on actual risk, derived from several risk assessment models, is increasingly available to consumers. A recent Cochrane review of 13 studies found that providing actual risk information increased utilization of screening tests, including mammography (35). But it is possible that the impact of actual risk information on adherence may vary, depending on how susceptible a woman believes she is and on whether her perceived risk is consistent with or differs from her actual risk.⁴ Adding to this complexity, current risk models do not account for many yet to be identified risk factors for breast cancer (36) and, owing to error or uncertainty, cannot provide precise individual risk estimates (34, 37). Moreover, these models may be of particularly limited value to African American women who are at high risk of developing aggressive disease and who tend to be diagnosed with breast cancer at an earlier age than White women, on whom the models were validated (36). Better models of breast cancer risk that are applicable to population subgroups are needed, as is further study of the role of perceived susceptibility in comprehensive strategies to increase adherence to screening guidelines.

Our findings should be evaluated in the context of the limitations and strengths of this study. Among its limitations, the study examined perceived susceptibility with a single item, so there is potential for measurement error. Furthermore, we included history of nonadherence to mammography screening guidelines among the covariates to enable us to examine the influence of perceived susceptibility on nonadherence subsequent to the index examination independent of the effect of past screening behavior; but, given the limitations of available data, this variable likely under controls for screening history. In addition, the study relied on self-reported mammography data that are vulnerable to inaccurate recall. Substantial agreement between self-reported adherence and mammography records in a random sample of study participants was reassuring, but these methods of outcome ascertainment were not in complete agreement in that subsample. Furthermore, because subject recruitment was restricted to urban, hospital-based facilities, the study findings may not generalize to women who receive screenings in rural areas or from private radiology/medical offices or freestanding mammography clinics. Nonparticipation and loss to follow-up may also have compromised the generalizability of the study findings, especially among African American women who were less inclined to participate or to provide follow-up data than their White counterparts. In addition, women lost to follow-up were more likely than those who completed the study to report high perceived susceptibility to breast cancer and to report a history of nonadherence during the baseline interview. This suggests that loss to follow-up led to an underestimation of the prevalence of high perceived susceptibility and might have attenuated the effect of high perceived susceptibility on nonadherence assessed at follow-up.

The strengths of this study include its prospective design, relatively large sample size, diversity of the study population, and adjustment of the primary association by multiple covariates that have been associated with mammography screening in prior studies (19, 38-48). In addition, we showed that the effect of perceived susceptibility that we observed held for time to first screening subsequent to the index exam, as well as for nonadherence. The diverse study population enabled us to determine that the negative effect on adherence of high perceived susceptibility applied to African American and White women with and without a family breast cancer history across a broad age range, whereas the adverse effect of low perceived susceptibility was limited to younger women. These findings thereby contribute to clarifying the causal relationship between perceived susceptibility to breast cancer and nonadherence and to refining our understanding of the impact of perceived susceptibility on specific population subgroups for which screening is recommended. These results also suggest that efforts to increase adherence to mammography screening guidelines should account for individual perceptions of susceptibility to breast cancer, reflecting the potentially complex nature of the relationship between these variables.

	Acknowledgments

 
We thank the following hospitals in Connecticut that allowed access to their patients and medical records: Bridgeport Hospital, Lawrence and Memorial Hospital, St. Francis Hospital and Medical Center, Waterbury Hospital, and Yale-New Haven Hospital, and Lisa Schlenk for her invaluable assistance with data collection and management.

	Footnotes

 
Grant support: National Cancer Institute grant RO1-CA-CA70731 (B.A. Jones), Agency for Healthcare Research and Quality grant RO3 HS11603 (L. Calvocoressi), and National Institute of Mental Health grant 5T32-MH-14235 (L. Calvocoressi).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

³ B.A. Jones, M.P.H., Ph.D., principal investigator.

⁴ In this study, 40% of women who reported high perceived susceptibility had a first-degree relative with breast cancer and may have been at actual higher-than-average risk of developing breast cancer, whereas the remaining women who reported high perceived susceptibility did not report breast cancer in a first-degree relative and may have been at lower actual risk.

Received 1/26/04; revised 6/14/04; accepted 7/ 2/04.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Michaelson JS, Satija S, Kopans D, et al. Gauging the impact of breast carcinoma screening in terms of tumor size and death rate. Cancer 2003;98:2114–24.[CrossRef][Medline]
2. Jones BA, Patterson EA, Calvocoressi L. Mammography screening in African American women: evaluating the research. Cancer 2003;97(1 Suppl):258–72.
3. Carney PA, Harwood BG, Weiss JE, Eliassen MS, Goodrich ME. Factors associated with interval adherence to mammography screening in a population-based sample of New Hampshire women. Cancer 2002;95:219–27.[CrossRef][Medline]
4. Vernon S. Risk perception and risk communication for cancer screening behaviors: a review. J Natl Cancer Inst Monogr 1999; 25:101–19.
5. McCaul KD, Branstetter AD, Schroeder DM, Glasgow RE. What is the relationship between breast cancer risk and mammography screening? A meta-analytic review. Health Psychol 1996; 15:423–9.
6. Katapodi M, Lee K, Facione N, Dodd M. Predictors of breast cancer risk and the relation between perceived risk and breast cancer screening: a meta-analysis. Prev Med 2004;38:388–402.[CrossRef][Medline]
7. Champion VL, Skinner CS, Foster JL. The effects of standard care counseling or telephone/in-person counseling on beliefs, knowledge, and behavior related to mammography screening. Oncol Nurs Forum 2000;27:1565–71.[Medline]
8. Aiken LS, West SG, Woodward CK, Reno RR, Reynolds KD. Increasing screening mammography in asymptomatic women: evaluation of a second-generation, theory-based program. Health Psychol 1994;13:526–38.[CrossRef][Medline]
9. Glanz K, Resch N, Lerman C, Rimer BK. Black-white differences in factors influencing mammography use among employed female health maintenance organization members. Ethn Health 1996;1:207–20.[Medline]
10. Champion V. Relationship of age to mammography compliance. Cancer 1994;74(1 Suppl):329–35.
11. Aiken L, Gerend M, Jackson K. Subjective risk and health protective behavior: cancer screening and cancer prevention. In: Baum A, Revenson T, Singer JE, editors. Handbook of health psychology. Mahwah (NJ): Lawrence Erlbaum Associates, Inc.; 2001. p. 727–46.
12. Ries LAG, Eisner MP, Kosary CL, et al. SEER cancer statistics review, 1975-2000. Bethesda (MD): National Cancer Institute; 2003.
13. Jones BA, Culler CS, Kasl SV, Calvocoressi L. Is variation in quality of mammographic services race linked? J Health Care Poor Underserved 2001;12:113–26.
14. Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines for the early detection of cancer, 2003. CA Cancer J Clin 2003;53:27–43.[Abstract/Free Full Text]
15. Leitch AM, Dodd GD, Costanza M, et al. American Cancer Society guidelines for the early detection of breast cancer: update 1997. CA Cancer J Clin 1997;47:150–3.[Abstract]
16. Stein JA, Fox SA, Murata PJ, Morisky DE. Mammography usage and the health belief model. Health Educ Q 1992;19:447–62.[Medline]
17. Bastani R, Marcus AC, Hollatz-Brown A. Screening mammography rates and barriers to use: a Los Angeles County survey. Prev Med 1991;20:350–63.[CrossRef][Medline]
18. Calnan M. The health belief model and participation in programmes for the early detection of breast cancer: a comparative analysis. Soc Sci Med 1984;19:823–30.
19. Fajardo LL, Saint-Germain M, Meakem TJ III, Rose C, Hillman BJ. Factors influencing women to undergo screening mammography. Radiology 1992;184:59–63.[Abstract/Free Full Text]
20. Fine MK, Rimer BK, Watts P. Women's responses to the mammography experience. J Am Board Fam Pract 1993;6:546–55.
21. Lerman C, Rimer B, Trock B, Balshem A, Engstrom PF. Factors associated with repeat adherence to breast cancer screening. Prev Med 1990;19:279–90.[CrossRef][Medline]
22. Zapka JG, Harris DR, Stoddard AM, Costanza ME. Validity and reliability of psychosocial factors related to breast cancer screening. Eval Health Prof 1991;14:356–67.[Abstract/Free Full Text]
23. Leitch AM. Controversies in breast cancer screening. Cancer 1995;76(10 Suppl):2064–9.
24. Kleinbaum DG, Kupper LL, Muller KE, Azhar N. Applied regression analysis and other multivariable methods. Pacific Grove (CA): Brooks/Cole Publishing Company; 1998.
25. Allison P. Logistic regression using the SAS system. Theory and application. Cary NC: SAS Institute, Inc.; 1999.
26. Schafer JL. Analysis of incomplete multivariate data. New York: Chapman and Hall; 1997.
27. Miller S. Applying cognitive-social theory to health protective behavior: breast self-examination in cancer screening. Psychol Bull 1996;119:70–94.[CrossRef][Medline]
28. NIH Consensus Development Conference Statement: breast cancer screening for women ages 40-49, January 21-23, 1997. NIH Consensus Development Panel. J Natl Cancer Inst 1997;89:1015–26.[Abstract/Free Full Text]
29. Leitch AM. Breast cancer screening: success amid conflict. Surg Oncol Clin N Am 1999;8:657–72, vi.
30. Cole SR, Bryant CA, McDermott RJ, Sorrell C, Flynn M. Beliefs and mammography screening. Am J Prev Med 1997;13:439–43.[Medline]
31. Facione NC. Perceived risk of breast cancer: influence of heuristic thinking. Cancer Pract 2002;10:256–62.[CrossRef][Medline]
32. Lindberg NM, Wellisch D. Anxiety and compliance among women athigh risk for breast cancer. Ann Behav Med 2001;23:298–303.[CrossRef][Medline]
33. Andersen MR, Smith R, Meischke H, Bowen D, Urban N. Breast cancer worry and mammography use by women with and without a family history in a population-based sample. Cancer Epidemiol Biomarkers Prev 2003;12:314–20.[Abstract/Free Full Text]
34. Leventhal H, Kelly K, Leventhal EA. Population risk, actual risk, perceived risk, and cancer control: a discussion. J Natl Cancer Inst Monogr 1999;25:81–5.
35. Edwards A, Unigwe S, Elwyn G, Hood K. Effects of communicating individual risks in screening programmes: Cochrane systematic review. BMJ 2003;327:703–9.[Abstract/Free Full Text]
36. Bondy ML, Newman LA. Breast cancer risk assessment models: applicability to African-American women. Cancer 2003;97(1 Suppl):230–5.
37. Claus EB. Risk models used to counsel women for breast and ovarian cancer: a guide for clinicians. Fam Cancer 2001;1:197–206.[CrossRef][Medline]
38. Hitchcock JL, Steckevicz MJ, Thompson WD. Screening mammography: factors associated with adherence to recommended age/frequency guidelines. Women Health 1995;1:221–35.
39. Halabi S, Skinner CS, Samsa GP, Strigo TS, Crawford YS, Rimer BK. Factors associated with repeat mammography screening. J Fam Pract 2000;49:1104–12.[Medline]
40. Harper AP. Mammography utilization in the poor and medically underserved. Cancer 1993;72(4 Suppl):1478–82.
41. Lee JR, Vogel VG. Who uses screening mammography regularly? Cancer Epidemiol Biomarkers Prev 1995;4:901–6.
42. Pearlman DN, Rakowski W, Ehrich B, Clark MA. Breast cancer screening practices among black, Hispanic, and white women: reassessing differences. Am J Prev Med 1996;12:327–37.[Medline]
43. Stein JA, Fox SA, Murata PJ. The influence of ethnicity, socioeconomic status, and psychological barriers on use of mammography. J Health Soc Behav 1991;32:101–13.[CrossRef][Medline]
44. Song L, Fletcher R. Breast cancer rescreening in low-income women. Am J Prev Med 1998;15:128–33.[CrossRef][Medline]
45. Taylor VM, Thompson B, Montano DE, Mahloch J, Johnson K, Li S. Mammography use among women attending an inner-city clinic. JCancer Educ 1998;13:96–101.
46. Thomas LR, Fox SA, Leake BG, Roetzheim RG. The effects of health beliefs on screening mammography utilization among a diverse sample of older women. Women Health 1996;24:77–94.[CrossRef][Medline]
47. Ulcickas Yood M, McCarthy BD, Lee NC, Jacobsen G, Johnson CC. Patterns and characteristics of repeat mammography among women 50 years and older. Cancer Epidemiol Biomarkers Prev 1999; 8:595–9.
48. Zapka JG, Stoddard A, Maul L, Costanza ME. Interval adherence to mammography screening guidelines. Med Care 1991;29:697–707.[Medline]

This article has been cited by other articles:

				M. E. Ford, S. L. Havstad, M. E. Fields, B. Manigo, B. McClary, and L. Lamerato Effects of Baseline Comorbidities on Cancer Screening Trial Adherence among Older African American Men Cancer Epidemiol. Biomarkers Prev., May 1, 2008; 17(5): 1234 - 1239. [Abstract] [Full Text] [PDF]

				S. E. Kim, E. J. Perez-Stable, S. Wong, S. Gregorich, G. F. Sawaya, J. M. E. Walsh, and C. P. Kaplan Association Between Cancer Risk Perception and Screening Behavior Among Diverse Women Arch Intern Med, April 14, 2008; 168(7): 728 - 734. [Abstract] [Full Text] [PDF]

				A. B. Dailey, S. V. Kasl, T. R. Holford, L. Calvocoressi, and B. A. Jones Neighborhood-Level Socioeconomic Predictors of Nonadherence to Mammography Screening Guidelines Cancer Epidemiol. Biomarkers Prev., November 1, 2007; 16(11): 2293 - 2303. [Abstract] [Full Text] [PDF]

				A. B. Dailey, S. V. Kasl, T. R. Holford, and B. A. Jones Perceived Racial Discrimination and Nonadherence to Screening Mammography Guidelines: Results from the Race Differences in the Screening Mammography Process Study Am. J. Epidemiol., June 1, 2007; 165(11): 1287 - 1295. [Abstract] [Full Text] [PDF]

				B. A. Jones, K. Reams, L. Calvocoressi, A. Dailey, S. V. Kasl, and N. M. Liston Adequacy of Communicating Results From Screening Mammograms to African American and White Women Am J Public Health, March 1, 2007; 97(3): 531 - 538. [Abstract] [Full Text] [PDF]

				L. Calvocoressi, M. Stolar, S. V. Kasl, E. B. Claus, and B. A. Jones Applying Recursive Partitioning to a Prospective Study of Factors Associated with Adherence to Mammography Screening Guidelines Am. J. Epidemiol., December 15, 2005; 162(12): 1215 - 1224. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Calvocoressi, L. Articles by Jones, B. A. Search for Related Content PubMed PubMed Citation Articles by Calvocoressi, L. Articles by Jones, B. A.