Since the introduction of the Papanicolaou (Pap) test in the 1940s, the incidence of and mortality from cervical cancer have decreased dramatically in countries with organized screening programs. Although cervical cancer screening has never been subject to randomized clinical trials, epidemiological data from countries that have instituted systematic screening programs strongly support the role of screening in reducing cervical cancer incidence and mortality. Data from case-control studies and large-scale screening programs conducted in the United States, Europe, and Australia suggest that Pap smear screening can reduce the incidence of invasive cancer by 60%–90%.^3–8

Data from the 2000 National Health Interview Survey (NHIS) and 2004 Behavioral Risk Factor Surveillance System (BRFSS) indicate that 83%–86% of U.S. women aged ≥18 have had a Pap smear within the past 3 years, which is the guideline proposed by the U.S. Preventive Services Task Force (USPSTF).^9,10 Although this appears to be a relatively high screening rate, there are certain groups of women in the United States who are less likely than others to be screened. A number of factors have been associated with lower rates of cervical cancer screening, including lack of a usual source of healthcare,^10–13 lack of health insurance,^10,12,14 low income,^10,12 low educational attainment,^10–12,14 obesity,¹⁵ smoking,^12,13 immigrant status,¹⁶ foreign birth,¹⁰ and not being married or living with a partner.^10,11,14 Recent data from the American Cancer Society (ACS) indicate that Asian and Hispanic women are less likely to be screened than non-Hispanic white and black women.¹⁷ Lower rates of cervical cancer screening have also been reported for Mexican American women,¹⁸ Vietnamese American women,¹⁹ and women living in Appalachia.²⁰ Although one might expect that women with a personal or family history of cancer might be more likely to undergo regular screening, data from the Women's Health Initiative (WHI) Observational Cohort Study suggest otherwise.¹²

Given the undisputed value of cervical cancer screening in reducing morbidity and mortality, this study sought to understand the factors associated with maintaining regular screening in a nationally representative sample of 25–64-year-old women. Our data source was the National Cancer Institute's (NCI) 2005 Health Information National Trends Survey (HINTS), a cross-sectional health communication survey that captures not only sociodemographic and general health information but also cancer-specific items not included in other national health surveys, notably NHIS, BRFSS, and the National Health and Nutrition Examination Survey (NHANES). Specifically, HINTS 2005 allowed us to examine whether knowledge of the role of human papillomavirus (HPV) in disease and awareness of cervical cancer screening guidelines predicted screening adherence. Also unique to HINTS 2005 was an item that allowed us to query cancer information-seeking behavior.

The prevalence estimate of recent cervical cancer screening (i.e., Pap test within the past 3 years) from HINTS 2005 was somewhat higher than the 2005 estimate from the Centers for Disease Control (CDC) and National Center for Health Statistics (NCHS).²⁹ Whereas 90% of HINTS respondents reported having a recent Pap test, 87% of 25–44-year-old women and 81% of 45–64-year-old women surveyed by the NCHS reported having a Pap test within 3 years. The only sociodemographic variable that was strongly associated with screening maintenance was access to healthcare coverage. This is consistent with 2005 NCHS data, which found that insured women were more likely than uninsured women to have had a Pap test within the past 3 years.²⁹ Others have found that compared with uninsured women, women who have health insurance are more likely to have not only timely Pap tests^12,30–34 but other cancer screening tests as well.^32,35–37 This is not surprising, as health insurance plans usually cover or defray the cost of cancer screening. Unfortunately, HINTS 2005 did not query usual source of healthcare, a variable that has been found to be a critical predictor of cervical cancer screening.^13,28,37 The combined effect of insurance and a usual source of care as a marker of access to care has been strongly associated with receipt of cancer screening.^12,30,38 In one national study, 35% of women who lacked health insurance and a usual source of care received a Pap test in the previous year compared with 67% of women who had both health insurance and a usual source of care.³⁹

All the behavioral and psychosocial variables were strongly associated with screening adherence. Overall, women who maintained on-schedule Pap tests appeared to be generally healthier than women who did not obtain regular Pap smears. Compared with women who never smoked, current smokers were significantly more likely not to have regular Pap tests, a finding that is consistent with data from the 1998 NHIS¹³ and the Women's Health Initiative.¹² This is particularly concerning because smoking is a recognized risk factor for cervical cancer.^40–42 Unfortunately, although smokers often acknowledge the health risks posed by smoking, they tend to minimize or deny their personal risk while attributing a greater degree of risk to other smokers.^43–48 This so-called optimistic bias,^48–51 or sense of personal immunity to the health risks of smoking,^52–54 may operate with respect to other health behaviors as well. For example, compared with nonsmokers, current smokers are less likely to have on-schedule mammograms and colorectal cancer screenings^12,38,55,56 and less likely to adhere to medical treatment.^57–59 A number of personality characteristics associated with smoking may also predispose smokers to neglect regular screenings. For example, compared with nonsmokers, smokers have been found to evidence greater depression, negative affect, anxiety sensitivity, and hostility^60–65; a tendency toward sensation seeking and impulsivity^66–69; deficient coping skills⁷⁰; a present time perspective⁷¹; and a tendency to rationalize risk.⁵³

Because unhealthy lifestyle behaviors have been noted to cluster together, it was not surprising that obesity was also strongly associated with not maintaining on-schedule screening. This too is concerning, as overweight women have significantly higher mortality rates from cervical cancer than normal weight women.^72,73 With one exception,⁷⁴ studies have consistently found an inverse relationship between BMI and cervical cancer screening. Obese women are more likely to delay medical care and less likely to have up-to-date Pap tests than normal weight women.^15,75–80

Mood was another significant predictor of screening status. Women who experienced psychological distress to such an extent that it substantially interfered with activities of daily living were significantly less likely to adhere to screening guidelines. The few studies that have examined the relationship between depressed mood and cervical cancer screening have yielded inconclusive results. One family practice-based study found that depressed women were slightly more likely to be screened than women with a chronic physical health condition,⁸¹ whereas other studies have reported no significant relationship between depression and cervical cancer screening.^82,83 A Canadian population-based study reported an interaction effect of age: compared with their nondepressed counterparts, younger depressed women were more likely to report a recent Pap test and middle-aged depressed women were less likely to report a recent Pap test.⁸⁴ Leiferman and Pheley⁸⁵ studied the effect of mental distress on preventive health behaviors in a community-based sample. Women who reported high levels of distress were significantly less likely to have had on-schedule Pap tests than women who were not distressed. Significantly lower rates of on-time Pap tests have been observed in women with psychiatric and substance use disorders.⁸⁶ Because we employed a broad definition of psychological distress that would include women with low-level dysphoria as well as women with major depressive disorder, our data suggest that even low levels of psychological distress can interfere with cervical cancer screening. One could speculate that women with a mood disorder might have less energy and fewer cognitive resources to plan for and pursue preventive healthcare services, such as cancer screening.

A strength of this study was the use of a nationally representative dataset to compare screening maintainers with nonmaintainers. To the best of our knowledge, this is one of the only studies to address this question. There are, however, a number of potential limitations to this study. First, because HINTS is a cross-sectional survey, causation cannot be inferred. Furthermore, because screening status was based on self-report, prevalence estimates may be biased due to several factors: the telescoping phenomenon (which is the tendency to report events as having occurred more recently than they actually occurred),^87,88 social desirability,⁸⁹ and the possibility that women may have confused nonscreening gynecological procedures with Pap tests.⁹⁰ Numerous studies have found a lack of concordance between self-reported Pap smears and medical record data. In general, women tend to overreport Pap screening in a given time period and underestimate the time from the previous screening.^91–94 Consequently, Pap screening rates that are based on self-report likely overestimate screening prevalence,^95–98 and our prevalence data may reflect this tendency. Another potential limitation involves the use of single items to measure key constructs, such as proficiency in English and cancer information seeking, which may have lowered the reliability of these constructs and thereby diminished the chances of identifying significant relationships. Finally, the overall response rate for the survey was relatively low, although comparable to other national telephone surveys, which reflects the decreasing trend in response rates for these types of surveys.⁹⁹ It should be noted that HINTS is a random-digit-dial landline survey that does not capture cell phone users. Although an estimated 7.8% of U.S. households in 2005 were cell phone only subscribers,¹⁰⁰ noncoverage of these households in traditional landline surveys is thought to have only a slight impact on outcome.^100,101 There are, however, notable differences between landline and cell phone users. Compared with landline users, cell phone only users tend to be younger, less affluent, unmarried, and Hispanic.^100,102 With respect to health behaviors, cell phone only users appear more likely to consume alcohol, smoke, be psychologically distressed, and be uninsured than landline users.¹⁰¹ Overall, rapidly declining telephone survey response rates coupled with the cost and logistics involved in conducting cell phone surveys pose a serious challenge to future telephone-based survey research.^102–104