Physician Recommendation for Papanicolaou Testing Among U.S. Women, 2000

Physician recommendation to get a cancer screening test is one of the strongest independent predictors of a woman's decision to be screened (1-7). Several studies have examined physician recommendation to get a cancer screening test or repeat screening test (4, 6-12). Most studies have focused on mammography, however, and relatively few studies have examined physician recommendation to get a Papanicolaou (Pap) test (13-19). Although there has been widespread use of the Pap smear as a screening test for cervical cancer during the past 50 years, many women have not shared fully in this progress (19).

To better understand the influence of certain variables on provider recommendation to receive cervical cancer screening, we developed a conceptual framework based on the systems model of clinical preventive care (20) and the behavioral model of health services use (21, 22). The systems model of clinical preventive care focuses on the patient-physician interaction. It details categories of factors that promote or inhibit preventive care (20). The behavioral model of health services utilization was developed to help understand the use of health services and measure equitable access to health care (21, 22). The behavioral model suggests that people's use of health services is a function of their predisposition to use services, factors that enable or impede use, and their need for care.

The variables addressed in the conceptual model included individual predisposing factors associated with the patient (age, race, Hispanic ethnicity, marital status, lower educational attainment, number of children, shorter duration of residence in the U.S., family history of breast, ovarian, cervical, or endometrial cancer, poor general health status, and knowledge, attitudes, and beliefs about cancer screening). Predisposing factors are antecedents to screening behaviors and include knowledge, beliefs, and attitudes (23). Predisposing factors also include demographic and socioeconomic variables such as age, race, and family history of cancer (1). Screening behaviors vary among women of different socioeconomic and ethnic groups. Lower rates of cervical cancer screening are especially prevalent among elderly and medically disadvantaged individuals with limited financial resources (1, 8, 19, 24). Low-income groups tend to have lower than average rates of screening for cervical cancer than other groups. Women's knowledge of screening guidelines is also a strong predictor of whether a Pap smear is obtained or not (25). Knowledge about cervical cancer has been shown to be positively associated with cancer screening practice among populations who have access to health information (23). Information filtered through an individual's health beliefs and attitudes can influence health-related behaviors and, especially when beliefs or misperceptions are based on incomplete or erroneous information, can lead to nonscreening or irregular screening (19, 26).

The conceptual framework also included enabling factors related to access, affordability, and availability (higher family income, employment status, having health insurance coverage, having a usual source of health care, English language, activities limitations, and increased number of physician visits). Enabling factors allow a predisposition to be translated into behavior (7, 19, 23, 27, 28). Women who do not have health insurance often cannot afford screening services (29). Accessibility of routine health care is important for women who reside in geographically isolated areas such as rural areas (30, 31).

Lastly, the conceptual framework considered reinforcing factors from the environment such as physician specialty (obstetrics and gynecology, general internist, or specialist) and geographic region (region of the U.S., residence in a metropolitan statistical area). Reinforcing factors are those related to feedback the learner receives from others, such as physicians, which may encourage or discourage behavior change (23, 32). External factors in the health care environment or organization have been shown to influence adherence to cervical cancer screening. Especially pivotal is the role of a health care provider to counsel, recommend, or perform the Pap screening test. Studies suggest that having a regular provider, continuity of health care, and having a recent physician visit can be facilitators to adopting or adhering to a screening behavior (33-37). Patients of female physicians, younger physicians, and obstetrician-gynecologists have higher breast and cervical cancer screening rates than other physicians (10, 35, 38, 40).

The objective of this study was to determine the proportion of women who had not received a provider recommendation to get a Pap test, according to several characteristics related to socioeconomic status and access to health care. Provider recommendation, a reinforcing factor, was a dependent variable for some analyses. The important research questions examined in this study included whether older women, those without health insurance, and women from racial and ethnic subgroups (African-American, Hispanic, and Asian and Pacific Islander) received a screening recommendation from their physician.

The data used in this analysis were from women who were interviewed as part of the year 2000 National Health Interview Survey cancer control topical module. The National Health Interview Survey is an annual nationwide in-person health survey of approximately 36,000 households of the civilian, noninstitutionalized population of the U.S. (41). The survey, which incorporates a multistage probability sampling design, is conducted by the National Center for Health Statistics and administered by the U.S. Bureau of the Census, which collects the data through household interviews. The National Center for Chronic Disease Prevention and Health Promotion of the Centers for Disease Control and Prevention, along with the National Cancer Institute, developed the questions on cancer screening practices for the Cancer Control Supplement of the year 2000 survey.

Analyses were limited to women aged 18 years or older with no history of hysterectomy (n = 13,636). For the analysis of the reason for not receiving a Pap test, only women who did not receive a recent Pap test were included (n = 2,333); for this reason, the sample size was smaller for that portion of the analysis. Of the 2,333 women who did not receive a recent Pap test, a small number of respondents (n = 23) were excluded whose race was neither White, Black, Hispanic, nor Asian/Pacific Islander, leaving 2,310 women in the analytic sample. Small numbers precluded looking at the 23 excluded women separately. Additionally, women who reported not having a doctor were excluded, along with those who had not visited a doctor in the past 12 months, leaving 1,502 women available for the Pap recommendation analysis.

Interviews included questions about Pap testing including physician recommendation. Each adult female respondent was asked whether she had ever had a Pap test and, if so, how long it had been since her last Pap test. Recent Pap test use was defined as within the past 3 years in accordance with U.S. Preventive Services Task Force guidelines for routine screening (42). Women who had not had a Pap test in the last 3 years were asked for the most important reason they had not had a Pap smear in this time period. Possible responses included “Doctor didn't order it,” “Doctor didn't say I needed it,” and “Don't have doctor.” Women were also asked, “In the past year, has a doctor or other health professional recommended that you have a Pap smear?”

Interviews also included questions about general health status, demographic and socioeconomic characteristics, factors related to access to health care, and other health-related factors. The independent variables of interest included predisposing factors such as age, race and Hispanic ethnicity, marital status, education, number of children in the household, duration of residence in the U.S., family history of cancer (breast, ovarian, cervical, uterine), and general health status. The independent variables also included enabling factors such as family income, employment status, health insurance status (private, public, none or single service plan), usual source of health care, language of interview (English, Spanish, or English and Spanish), activities limitations, and number of physician visits in the past year. Other independent variables included reinforcing factors such as having seen or talked to an ob-gyn physician, specialist, or general doctor in the past year, region of the U.S. (Northeast, Midwest, South, West), and metropolitan statistical area (or non-metropolitan statistical area). No questions about cancer screening knowledge, attitudes, or beliefs were included in the analysis.

Age-specific and age-adjusted percentages were calculated for physician recommendation for a Pap test. The direct method was used for age adjustment using women aged 18 and older with no recent Pap test and no previous hysterectomy in the 2000 National Health Interview Survey sample as the standard population. General linear contrasts were then used to test for differences in age-standardized recommendation percentages across levels of predisposing, enabling, and reinforcing factors. All analyses used SAS and SUDAAN to calculate the 95% confidence intervals (CI) and to allow for weighting of the estimates to account for the complex sampling design (43). The weights used included adjustments for poststratification and nonresponse. Exact binomial confidence limits were calculated (44). Lack of physician recommendation and other reasons for not having a recent Pap test were stratified according to race and ethnicity (African-American, Hispanic, and Asian and Pacific Islander) and according to health insurance status (private, public, none or single service). Statistical testing for differences across strata defined in this way was not done due to issues associated with multiple comparisons. In these exploratory analyses, point estimates of rates and 95% CI were considered when examining differences across groups.

Multivariate analysis of predictors of physician recommendation for a Pap test was carried out using logistic regression techniques and SUDAAN. A backwards variable selection procedure was used. Variables significantly associated with physician recommendation to get a Pap test at the P < 0.10 level were retained in the final model. Indicator variables for race/ethnicity were also forced into the model.

Among women aged ≥18 years who had no history of hysterectomy, 83.3% (95% CI, 82.4-84.1) of 13,636 women in the sample had had a Pap test in the last 3 years. Among 2,310 women who had not had a recent Pap test, reported reasons for not receiving a Pap test included: “No reason/never thought about it” (48.0%; 95% CI, 45.5-50.7), “Doctor didn't order it” (10.3%; 95% CI, 8.7-12.0), “Didn't need it/didn't know I needed this type of test” (8.1%; 95% CI, 6.7-9.6), “Haven't had any problems” (9.0%; 95% CI, 7.6-10.5), “Put it off” (7.4%; 95% CI, 6.2-8.7), “Too expensive/no insurance” (8.7%; 95% CI, 7.3-10.2), “Too painful, unpleasant, embarrassing” (3.5%; 95% CI, 2.5-4.6), and “Don't have doctor” (1.7%; 95% CI, 1.2-2.4).

In the smaller sample of women who had had a doctor visit in the last year but who had not had a recent Pap test (n = 1,502), about 86.7% (95% CI, 84.5-88.6) of the women reported that their doctor had not recommended a Pap test in the last year.

Similar results were obtained after stratification by race and Hispanic ethnicity, and health insurance status with a few exceptions. Hispanic women were more likely to report “never thought about it” and less likely to report “put it off” compared with White women or Black women (Table 1). Asian/Pacific Islanders were less likely to report “too expensive/no insurance” compared with the other groups. Those with no insurance were much more likely to report “too expensive/no insurance” as the reason for no Pap test compared with those with health insurance (Table 2).

Among women who had a doctor visit in the last year but who had not had a recent Pap test (n = 1,502), factors associated with not receiving a doctor recommendation to get a Pap test included age <30 years or ≥65 years, never having been married, shorter duration of residence in the U.S., and not having been seen by a general physician (Table 3). African-American women were as likely to have received a doctor recommendation to get a Pap test as White women. Hispanic women were as likely to have received a doctor recommendation to get a Pap test as non-Hispanic women.

In multivariate analysis, factors positively associated with doctor recommendation to get a Pap test included being ages 30 to 64 years, having been born in the U.S., and having seen a specialist or general doctor in the past year (Table 4). Women who had lived in the U.S. for <10 years were less likely to have had a doctor recommendation to get a Pap test as compared with those who had been born in the country. No associations were observed with race or Hispanic ethnicity in multivariate analysis.

We carried out a further analysis to compare women in our analytic sample with those who were excluded from the analysis because they did not have a recent doctor visit and therefore had no opportunity to have a doctor recommendation to get a Pap test. Among women aged 18 or older with no history of hysterectomy, 22.5% of women who reported no doctor visit in the past year were Hispanic as compared with 10.7% of women who did have a doctor visit in the past year (results not shown). About 58.6% of women with no doctor visit in the past year were White as compared with 74.5% of women who did have a doctor visit in the past year. About 8.8% of women with no doctor visit in the past year were Asian/Pacific Islander as compared with 5.5% of women who did have a doctor visit in the past year (P = 0.041). Women who reported having had a doctor visit in the past year were also much more likely to have a usual source of care (88.5% versus 52.7%).

Physician recommendation to obtain a Pap test can be understood as an interplay of predisposing factors associated with the individual, only some of which are modifiable, enabling factors associated with the individual that are modifiable, and external reinforcing factors that are associated with the physician and the health care or geographic environment. The effects of these three categories of determinants are mediated by a complex set of social factors that influence physicians to recommend the adoption of a particular screening modality.

All behavior models have shortcomings, as do the models we have adapted. For instance, a determinant of screening is the presence or absence of a dialogue between the physician and the patient (35). This framework does not consider the presence of a dialogue between the physician and patient, such as physicians who support, promote, or recommend the screening test. The model also does not consider other possible spheres of physician influence, including their knowledge, beliefs, priorities, and attitudes (33). To better understand the health care environment, including organizational factors within the practice setting, future models would be more revealing if contextual factors associated with patient visits were linked to physician recommendations (32, 45). Contextual factors include the setting in which patient visits occur, as contrasted with individual-level patient or provider characteristics.

The observed decrease in doctor recommendation for a Pap test in the oldest age categories may be due to screening guidelines indicating that routine cervical cancer screening may not be necessary for all older women (42). Guidelines published prior to this survey also suggested that Pap testing could be discontinued after age 65 in women who had had regular previous screenings in which the smears had been consistently normal (46, 47). Previous studies have examined barriers to obtaining a physician recommendation for Pap testing among younger women (4, 7, 19, 24, 28, 48, 49).

With respect to other research questions addressed by the present study, race, Hispanic ethnicity, and health insurance status were not found to be important determinants of whether women received a screening recommendation from their physician. Nevertheless, physician-patient communication about cancer screening is an important part of culturally appropriate health care for minority patients who face severe language and cultural barriers to using adequate health services (1, 50, 51). In prior studies, physician recommendation to get a Pap test or mammogram has been found to be an important predictor of cancer screening in studies of Hispanic, Asian, African-American, and White women (51-53). The results of the present study suggest that special efforts are needed to ensure that immigrant women receive a recommendation from their physician to get a Pap test. For many foreign-born women, particularly Asian-American women, there is a lack of familiarity with Western preventive concepts (1). There may also be diminished seeking of appropriate health care, with or without symptoms. This may be particularly true among “hard-to-reach” medically underserved women. Frequently, these women only seek episodic care and are easily missed when providers lack support personnel, reminders, or themselves are unaware of low-cost screening and treatment options for immigrant women (4).

Studies have found consistently that individuals who report provider recommendation or encouragement are more likely to complete screening (32, 54). In the present study, the lack of statistical significance among the factors that have been previously shown to be associated with recommendation to get a screening test could be due to the small sample size.

With respect to other limitations, self-reported information about provider recommendation may differ from information obtained from records of health care providers or from what is reported in physician surveys. In addition, the doctor may have not recommended the Pap test be done based on the new screening guidelines which suggest longer intervals for screening ranging from 1 to 3 years, depending on factors such as age, screening history, type of Pap test, and history of immunosuppression (42, 54). Guidelines other than the U.S. Preventive Services Task Force for cervical cancer have also suggested screening less frequently than annually after three consecutive normal annual Pap tests (55, 56).

A further limitation is that information about the reasons for physician visits was not asked for in the survey, and neither was the context of the visit. Physicians must prioritize delivery of multiple preventive services during office visits, and pelvic exams and Pap tests may be more likely to be done during regular visits or visits for a complete physical examination rather than with visits for acute or chronic illnesses (11, 14, 57, 58). Patients may be especially likely to undergo cancer screening in conjunction with a health maintenance visit, offering visit, or physical examination (11, 33). Cancer screening has been related to other factors not examined in the present study such as health beliefs, for example, beliefs about the benefits of having a cancer screening test (7, 28). A further issue is that the 2000 National Health Interview Survey asked for the most important reason for not having a Pap test. It is possible that women would give more than one reason if it were allowed. However, it is not possible to examine the relative importance of combinations of reasons.

In summary, these findings suggest that the lack of a physician recommendation contributes to underuse of Pap screening by many eligible women. Given research that shows the effectiveness of physician recommendations in improving use and increasing cancer prevention, increased physician recommendations could contribute significantly to increased Pap screening use in the U.S. The results of this study suggest that physicians do not always appropriately recommend a Pap test to women. Encouraging other researchers to use the context of the clinic visit, as well as theoretical models that address these issues in future research, will lead to a better understanding of why physician recommendation does not always lead to a Pap test.