Although screening for breast and cervical carcinoma has been widely accepted as beneficial, specific segments of the population are not receiving these services as frequently as recommended. The objective of this study was to describe differences in breast and cervical screening prevalence among those with activity limitations and those residing in rural areas. Disparities in the incidence rates of in situ breast carcinoma and cervical carcinoma, as well as mortality among rural and urban women, are also described.
Data from the Iowa Behavioral Risk Factor Surveillance System (BRFSS) and Iowa's Surveillance, Epidemiology, and End Results (SEER) were used to determine the prevalence of screening and adverse outcomes among rural populations. To describe the rural nature of counties, the authors used the number of residents per square mile for each county and classified the results into five groups. Mulitple logistic regression was used to determine the prevalence of screening among those with activity limitations and rural residents.
Using the BRFSS, those with activity limitations and those residing in rural Iowa were less likely screened for breast or cervical carcinoma. This translated into a lower in situ breast carcinoma incidence rate and a higher invasive cervical carcinoma incidence rate among rural women relative to their urban counterparts. No differences were found for mortality from these cancers.
Screening for breast and cervical carcinoma has been widely accepted as beneficial, although some recommendations differ with respect to frequency and the age at which to begin and end the screening.1 Despite these recommendations, there are still segments of the general population that are not receiving these screening services as frequently as recommended. As a result, cancers may be diagnosed at later stages and mortality may be increased.
A major effort to reduce health disparities is one of the two overarching goals of Healthy People 2010. One critical avenue for reducing health disparities is the identification of subgroups that are underserved and evaluation of the extent of the disparity. One of the emerging issues is that of disease development and risk factor prevalence among persons with disabilities.2 Recently, some studies have shown that women with disabilities are less likely screened for breast and cervical carcinoma,3 whereas other studies have not.4 In addition, minority women, those of lower socioeconomic status, and rural women were less likely screened for breast and cervical carcinoma.5 As a result, populations with these specific characteristics are not likely to achieve national or state screening objectives for the year 2000.
There are at least two segments of the population that are of particular interest in Iowa, namely, those with disabilities and rural populations. Disability is of significance because of the rapidly increasing elderly population in Iowa.6 The number of elderly Iowans, and, as a result, the number of those having a disability, are expected to increase significantly during the next century.6, 7 Currently, Iowa has the highest percentage of residents age 85 years or older in the nation and is fourth in the percentage of persons age 65 or older.8 Although there is an increase in awareness of disability-related issues, little attention has been directed toward preventive care for persons with disabilities.
Iowa's rural population is also of particular interest from a policy perspective. Iowa is a rural state, with 39.4% of the population living on farms, in open rural areas, or in towns with fewer than 2500 residents.9 Unique health care issues are associated with residents' geographic locations in terms of access to primary care, hospital utilization, and cancer treatment.10–12
Delivering preventive care in primary care practices is often more difficult for rural than for urban practices. Rural Americans are demographically different from urban Americans in ways that may affect preventive care, in that the former population tends to be older, poorer, and less educated. Medical practice is also different in rural areas, with fewer physicians and physician contacts, and patients spend less time with their physicians.10
Several studies have identified that rural women were less frequently screened for breast and cervical carcinoma.5, 13 Few studies have discussed lower screening prevalence in conjunction with later stage of disease presentation or increased mortality.
Therefore, the objective of this study was to identify differences in the receipt of breast and cervical carcinoma screening services among subgroups of the Iowa female population, with particular focus on those with activity limitations and those residing in rural areas. As a result of the anticipated lower screening by rural populations, we hypothesized that the age-adjusted in situ breast carcinoma occurrence rate for women residing in rural counties was lower than for women in urban counties. In addition, the age-adjusted incidence of invasive cervical carcinoma is anticipated to be higher among rural women than among their urban counterparts. Finally, mortality from breast and cervical carcinoma is expected to be elevated among rural women compared with urban women.
Screening data were collected as part of the Iowa Behavioral Risk Factor Surveillance System (BRFSS) during 1996–1997. The BRFSS is an ongoing monthly telephone survey financially and technically supported by the Centers for Disease Control and Prevention (CDC) conducted in all states. It is designed to collect information on the health behaviors of noninstitutionalized Iowa residents age 18 years or older and to monitor the prevalence of these behaviors over time. The assessed behaviors are major contributors to illness, disability, and premature death. Examples of topics that are addressed include breast, cervical, and colorectal carcinoma screening; disability; smoking; physical activity; nutrition; diabetes; and cardiovascular risk factors. Additional information about the methodology of the BRFSS is described elsewhere.14
In 1996 and 1997, a total of 18,265 calls were made to potential respondents, 7200 of whom were eligible to participate. The call dispositions for the combined 1996–1997 data collection was as follows: 7200 completed interviews (39.4%), 4787 nonworking phone numbers (26.2%), 2431 business telephone numbers (13.3%), 1154 calls with no one answering on repeated tries (6.3%), refusal of 1113 potential respondents to participate (6.1%), and 1580 calls with other disposition codes (8.7%). The percentage of all numbers called that resulted in completed interviews in Iowa was 39.4% for both years combined, which was substantially higher than the median of all states for this time period (25.6%). The CASRO rate, developed by the Council of American Survey Research Organizations, reflects telephone sampling efficiency as well as the degree of cooperation among eligible people contacted, with higher rates reflecting lower potential for bias in the data. In 1996 and 1997 the CASRO rates for Iowa were 74.9% and 69.4%, whereas the median rates for all states were 68.4% and 62.1%, respectively.
Cancer incidence and mortality data were obtained from the Surveillance, Epidemiology, and End Results (SEER) program as part of the State Health Registry of Iowa. The SEER program is a population-based cancer registry covering the entire state of Iowa, with staff who conducts active, extensive annual follow-up for cancer patients. Abstracters regularly visit hospitals, clinics, and medical laboratories in Iowa and neighboring states to collect cancer data. SEER records data on all newly diagnosed cancers, including selected patient demographics, primary site, morphology, diagnostic confirmation, extent of disease, and first course of cancer therapy. Death certificates are routinely obtained and linked to the identified cases. Women diagnosed with primary breast or cervical carcinoma in Iowa from 1991 through 1995 were included in this study.
Classification of Activity Limitation
During 1996–1997, 6 questions were included in the Iowa BRFSS to assess the presence of activity limitation and underlying conditions. Respondents were classified into three populations with varying degrees of limitations. Individuals who answered affirmative to the first question (“Are you limited in any way in any activities because of any impairment or health problem?”) were classified as having an activity limitation for the current study. Individuals were classified as severely limited if they needed help from other persons with their personal care needs, such as eating, bathing, dressing, or getting around the house and/or needed help from other persons in handling their routine needs, such as everyday household chores, doing necessary business, shopping, or getting around for other purposes. Thus, 3 populations are studied: 1) those without any limitations, 2) those with some limitations but not severely limited, and 3) those with severe limitations. The primary impairments or health problems that limited the individual's activities were used to classify the underlying conditions.
Classification of Residence Location
The rural nature of the respondents' residence was obtained by dividing the number of county residents by the number of square miles per county. This gives a measure of the population density of each of Iowa's 99 counties. Next, we classified the population densities into 5 different groups to create an ordinal scale: less than 20 residents per square mile (21 counties), 20–29 residents per square mile (32 counties), 30–39 residents per square mile (23 counties), 40–99 residents per square mile (15 counties), and 100 or more residents per square mile (8 counties). Counties with the fewest number of residents per square mile were considered “rural,” whereas counties with 100 or more residents per square mile were considered “urban” for this study.
There are several organizations and agencies that provide recommendations about the appropriate population and frequency for breast and cervical carcinoma screening, including the U.S. Preventive Services Task Force, the American Cancer Society, and the National Cancer Institute. The 1996–1997 BRFSS collected data about breast and cervical carcinoma screening. Women age 50 years or older who had had a mammogram in the past 1–2 years and a clinical breast examination (CBE) in the past year were considered appropriately screened for breast carcinoma in the current study. Women were required to have had a Papanicolaou (Pap) smear in the past 3 years to adhere to screening guidelines for cervical carcinoma (there was no age limitation).
Due to the sampling complexities, SUDAAN (version 7.5) was used to calculate the prevalence of activity limitations of adult Iowans age 18 years or older. PROC CROSSTABS was used to estimate the prevalence of breast and cervical carcinoma screening among the different populations. The data are weighted to reflect the responses of the Iowa civilian noninstitutionalized population age 18 years or older.
SUDAAN's multiple logistic regression (PROC LOGISTIC) was used to assess the significance of the associations of activity limitation and population density, adjusting for differences between populations, which warrant controling during modeling of the regression equation. The Wald chi-square test was used to test statistically the importance of variables and interactions between variables. A P value of 0.05 or less was considered statistically significant.
SEER*Stat15 was used to calculate cancer incidence and mortality; both were age-adjusted according to the 1970 U.S. standard population. In situ breast carcinoma rates for women age 50 years or older were calculated as a measure of mammography utilization, with higher rates indicating higher rates of utilization. Invasive cervical carcinoma rates were used as an outcome measure of Pap smear screening, because higher rates indicate a lower rate of progression of premalignant to invasive cancer and therefore lower use of Pap smears.
Prevalence of Limitations
A total of 6.4% of females age 18 years or older reported needing help with personal care and/or routine needs (severe limitations), 11.7% reported some limitation (but not personal care or routine needs), and 81.9% reported that they were not limited in any activity. Among all females, the most prevalent health problem or condition among those with any activity limitation was arthritis (18.9%), followed by back/neck problems (18.7%) and “other” conditions (18.4%) (see Fig. 1). The most prevalent condition among those with severe limitations were “other” conditions (19.7%), back/neck problems (17.3%), and arthritis (16.7%). Among those with some limitation, the most prevalent conditions were arthritis (20.0%), back/neck problems (19.5%), and “other” conditions (17.8%%).
Among females age 50 years or older, 10.1% reported being limited in their personal care or routine needs; thus, they were classified as having a severe activity limitation. An additional 6% of females age 50 years or older reported a limitation but were not limited in their personal care or routine needs (some limitation). The remaining 73.8% of females in this age group did not report an activity limitation.
Among this population age 50 years or older, the most prevalent condition was arthritis (25.3%), followed by fractures, bone/joint injury (15.6%), and “other” conditions (13.9%) (Fig. 1). The top three underlying conditions among those with severe and some limitations were the same.
Cervical Carcinoma Screening
Figure 2 displays the use of cervical carcinoma screening services by level of activity limitation and respondent residence (county population density). A lower percentage of women who reported needing help with their personal care or routine needs (severely limited) had a Pap smear during the past 3 years (65.5%), versus those who were classified as limited but not severely (72.8%), and individuals without any limitation (78.9%). Only 68.5% of residents of the most rural counties received Pap smears in the preceding 3 years, compared with 81.5% of residents of the most urban counties.
When controlling for having health insurance, the woman's educational level, age, and household income, statistical differences were observed according to the number of residents per square mile and the level of activity limitation (Wald chi-square = 3.5274; degrees of freedom [df] = 8; P = 0.0005) for cervical carcinoma screening. Among those with severe activity limitations and among those without limitations, the association between county population density and being screened for cervical carcinoma was statistically significant (Table 1). Especially among women with severe limitations, a substantially lower percentage of residents of counties with fewer than 100 residents per square mile received Pap smears during the 3 years preceding the interview, compared with residents of the most urban counties. Differences among those without activity limitation did exist but were not as pronounced.
Table 1. Percentage of Women Screened by Level of Activity Limitation and Residence Location, 1996–1997
Multivariate model controlled for income and having health insurance.
Multivariate model controlled for education, age, income, and having health insurance.
Population density (per square mile)
Among specific levels of population density, the presence of an activity limitation affected cervical carcinoma screening only for those residing in the two groups of the most urban counties. Among counties with 40–99 residents per square mile, a lower percentage of those with severe limitations were screened, but among residents of the most urban counties a slightly higher percentage were screened compared with those without limitations.
Figure 3 shows that the age-adjusted invasive incidence of cervical carcinoma decreased with increasing population density, from 9.3 among residents of the most rural counties to 7.3 per 100,000 among those residing in the most urban counties. The age-adjusted mortality rate was similar across the different population densities, at about 2 per 100,000, although the number of cervical carcinoma deaths was typically low (an average of 44 per year for the state as a whole).
Breast Carcinoma Screening
Only 1 of every 3 women age 50 years or older had had a mammogram during the past 1–2 years and a CBE during the past year (Fig. 4). The percentages of women with severe, some, and no limitations who were screened for breast carcinoma were relatively similar for all 3 populations—36.7%, 39.6%, and 32.3%, respectively. The percentage of women screened for breast carcinoma in the two groups of most rural counties was about seven percentage points below that of women residing in other counties.
When including the variables education, income, and health insurance, the interaction between the level of activity limitation and county population density was statistically significant (Wald chi-square = 2.9563; df = 8; P = 0.0028) for use of breast carcinoma screening (Table 1). In addition, the relation between place of residence (population density) and the use of breast carcinoma screening was statistically significant among those with severe, some, and no activity limitations (see Table 1). Thus, county residence does influence the use of breast carcinoma screening at all three levels of activity limitation.
In particular, breast carcinoma screening declined with increasing population density in all three groups of activity limitation. However, when determining whether activity limitation affected breast screening at particular levels of population density, only statistically significant differences were found for 2 of the 5 levels of rurality, among those with 20–29 and those with 30–39 residents per square mile. Thus, those with activity limitations residing in the most rural counties were as likely to have been screened for breast carcinoma as those without limitations.
Figure 3 shows that the age-adjusted in situ breast carcinoma rate increased with increasing population density, from 37.3 per 100,000 for the most rural counties to 47.7 per 100,000 for residents of the most urban counties. This is an indication that mammography is underutilized in counties with a lower population density, i.e., rural counties. The unadjusted mortality rate showed a dramatic decline from 45.2 per 100,000 for residents of the most rural counties to 31.8 per 100,000 for the most urban counties, a 30% difference (not shown). However, after controlling for age differences between the different population density categories, this difference disappeared, indicating a higher percentage of elderly women, those at greatest risk of breast carcinoma, in the most rural counties (Fig. 3).
To achieve the national health objectives for the year 2000, and to reduce disparities between populations as proposed by the objectives for the year 2010, the use of breast and cervical carcinoma screening by all segments of the population needs to be assessed. Based on the current study, differences in the receipt of breast carcinoma screening services were found, particularly among Iowa's rural population and, to a lesser degree, among those with activity limitations, who were less likely screened. This lower rate of screening may likely have resulted in a lower rate of in situ breast carcinoma detection as well as a higher unadjusted rate of breast carcinoma mortality among rural populations, although the age-adjusted mortality rate showed little difference across population densities.
There has been some controversy about the use of breast carcinoma screening among women with and without disabilities. Ives et al.4 found differences between the populations, as did Nosek and Gill;3 but women with physical disabilities were equally likely to receive mammograms within the preceding 2 years, according to Nosek and Howland.16 The latter study also determined that one of the main reasons that women with physical disabilities did not receive mammograms was their inability to get into the physical position required. Some impairments, such as severe kyphosis or being bed-ridden,17 can be accommodated by mammography equipment; but not all mammography facilities, especially mobile vans, are expected to have the necessary equipment. Additional studies are warranted that will focus on patient, physician, equipment, system, and environmental barriers to breast carcinoma screening.
There may be several reasons for the differences between the few studies that have contrasted the receipt of services by persons with and without disabilities. First, the definition of disability or limitation and the prevalence of underlying conditions varies among studies, which may impact the receipt of breast carcinoma screening. For example, Nosek and Howland16 studied women with physical disabilities, with 56% of women reporting spinal cord injuries, polio, and neuromuscular disorders as underlying conditions. Nosek and Gill3 and Ives et al.4 defined disability as limitation of the instrumental activities of daily living. Unfortunately, it is unclear what the prevalence and type of underlying conditions were in these two studies. In the current study, the definition of activity limitation was relatively broad, because respondents were asked about any limitations in any way because of any underlying condition. The predominant conditions underlying the activity limitation were musculoskeletal in nature, such as back/neck problems, arthritis/rheumatism, and fractures. Although screening for breast carcinoma is important for all women with (and without) limitations, there may be screening differences among women with certain types of underlying conditions leading to certain limitations, which may have affected the results found by different studies.
Second, there may be several barriers associated with receipt of breast carcinoma screening that may also be related to the functional status of women. For example, lower socioeconomic status is adversely associated with breast carcinoma screening and is also related to functional status. For example, the variable of activity limitation became statistically nonsignificant when income was included as a covariate in the model. Similar arguments can be made for other barriers. The decline in mammography use with age and number of functional limitations in the study by Nosek and Gill3 may have been the result of not adjusting for the women's educational level; this declines with increasing age and is strongly associated with the receipt of breast carcinoma screening. In our study too, a lower percentage of women with severe limitations were screened for breast carcinoma (Fig. 2).
Third, although mammography is an important tool for detecting breast carcinoma at an early stage, the concurrent practice of CBE is a necessary component of an effective screening practice, because some cancers may be missed with mammography alone.18 None of the other three studies used mammography and CBE as a single outcome measure; rather, these studies only included mammography. In the current study, we included receipt of both mammograms and CBEs as a comprehensive tool for detecting breast carcinoma early.
At least as important as screening by women with activity limitations is the underutilization of breast carcinoma screening services by rural women, which translated into a lower age-adjusted rate of in situ breast carcinoma for this population. A lower rate of cancer screening has been noted for rural women; however, few studies have also identified the impact on health outcomes. In fact, the use of breast carcinoma screening services among the general female population in Iowa is lower than in many other states.19 Thus, population-based efforts, with particular focus on women in rural counties and those with activity limitations, are important to undertake.
Similar to breast carcinoma screening, we found that the percentage of women residing in more rural counties who received a Pap smear during the past 3 years was significantly lower than that of their urban counterparts. Although our study only found sporadic differences, women with limitations were less likely screened than those without limitations in the National Health Interview Survey data.3 However, Ives et al.4 did not show differences between populations.
The current study found that the incidence rate of invasive cervical carcinoma decreased with increasing population density. This apparent decrease is likely associated with decreased frequency with which Pap smears are received among rural women. A higher percentage of premalignant cancers are becoming invasive cervical carcinomas among Iowa's rural women. Sung et al.20 found that the rate of cervical carcinoma was higher among Georgia's Medicaid recipients than among their rural counterparts. Racial differences may have contributed to this difference. Although racial and ethnic minorities are more likely to reside in Iowa's urban counties, the percentage of minorities in Iowa is less than 10% and is unlikely to have affected the cervical carcinoma rates in our study. In fact, only the incidence rate among counties with at least 100 residents per square mile decreased when only white women were included, from 7.2 to 6.9 per 100,000. Because the current study was descriptive in nature, reasons for underutilization of cervical carcinoma screening among rural women were not assessed.
Although there are many potential barriers to breast and cervical carcinoma screening, such as cost, one of the most significant barriers is lack of recommendation by a physician.21 Therefore, increasing physician compliance with screening recommendations may be an important focus. Successful strategies to increase physician compliance include chart checklists, educational programs, and reminder systems (handwritten or computer-generated) attached to the chart of the patient's visit.22, 23 Unfortunately, most interventions have been evaluated using university-based clinics and have not included rural clinics. Rural medical practice differs from urban practice in several ways: 1) there are more rural areas with no primary care physician; 2) those rural areas with physicians have fewer of them per population; 3) rural primary care physicians are organized in smaller groups, with fewer office personnel and smaller offices; and 4) rural physicians tend to spend less time with each patient.10 All of these factors impact on the utilization of breast and cervical carcinoma screening. Thus, implementation of office reminder systems may not be as successful among rural practices as has been shown in more urban practices, although studies are needed in this area.
Strengths and limitations of the current study can also be identified. First, a strength of the study was its ability to compare Iowans who had activity limitations and those residing in rural areas in the same analysis. Most studies have included one or the other of those factors, but not both. Second, several covariates were included in the multivariate analysis controlling for differences between persons with activity limitations and residents of counties with varying population densities. Third, specific target populations were identified with implications for health-planning purposes. Fourth, screening, incidence, and mortality data were included in the same study, resulting in a more complete description of the issues. Fifth, the data used were population-based, resulting in broader applications.
A limitation of the study was that the BRFSS includes only noninstitutionalized Iowans. Thus, nursing home residents, who have high levels of activity limitations, were not included. Second, only respondents with telephones are included in the BRFSS. The percentage of Iowans without telephones is very small and not expected to have biased the study results to any significant extent. Third, because the BRFSS is a telephone interview, in some cases communication devices such as TTYs may have been mistaken for fax lines. In addition, those with mental disabilities that made them unable to respond to telephone questions were not included in the data. It is expected that a number of noninstitutionalized Iowans with activity limitations were precluded from participation in the telephone interview. However, it is unclear how many persons this affected and to what extent this biased the study's findings. Fourth, the use of a county's population density assumes a uniform measure of a county's rurality. However, because of the relatively large number of Iowa counties (99) and their geographic size, it is expected that a county's population density is a relatively accurate indication of rurality.
In conclusion, despite the low overall prevalence of breast carcinoma screening in Iowa (33.4%), there are segments of the Iowa population who are even less likely screened, i.e., rural women and those with activity limitations, and this resulted in adverse health outcomes. Lower utilization of cervical screening has likely resulted in higher invasive cervical carcinoma detection rates among Iowa's rural residents. Increasing focus should be placed on rural women and those with activity limitations.