PREVENTIVE
MEDICINE
Breast Cancer
21, 405-418
(1992)
Screening Attitudes and Behaviors and Urban Women’
HEATHER BRYANT, Alberta
Program
M.D.
PH.D. ,’ AND ZEVA
MAH,
for the Earl-y Detection of Cancer, Alberta Cancer 1040-7 Avenue SW, Calgary, AB, Canada ZIP 3G9
of Rural M.Sc.
Board,
120,
This study was carried out to assess the breast cancer knowledge, attitudes, and awareness of women age 40 to 74 in Alberta, a Canadian province of 2.4 million people. This analysis compares the attributes of 538 rural women, defined as those living between 1 and 3 hr drive from the major cities in Alberta, and 735 urban women who lived in one of these two cities. Rural women were found to have the same basic knowledge of breast cancer or perceptions of barriers to mammography, but had more negative attitudes about breast cancer itself. Despite their similar access to physician care, they were less likely to have had a recent clinical breast examination or mammogram (P < 0.001). These differences remained when adjustment was made for demographic background variables; the adjusted prevalence rate ratio for a screening mammogram in the past 2 years was 0.52 (95% C.I., 0.43,0.64), and for intention to have a mammogram in the next 2 years, 0.75 (0.63.0.90). The results suggest particular program delivery strategies when planning for provision of breast screening information and service to the large subgroup of rural women. o 1992 Academic press, hc.
INTRODUCTION
Breast cancer is one of the leading causes of death for women in developed countries. In the United States and Canada, more than 415,000 deaths from breast cancer are expected to occur over the next year (1, 2). Canadian statistics show breast cancer as the leading cause of potential years of life lost (3) and, as the most important cause of cancer deaths among women, still exceeding even lung cancer in mortality (1). Because of this, and the potential impact mammographic screening could have in altering these statistics, initiatives have been underway in many developed countries to encourage screening or to develop centralized, quality controlled screening programs on a regional basis (4). These vary in their criteria for age and interval of screening, largely due to unresolved questions in the literature regarding appropriate targets and methodologies. However, as noted in the Healthy People 2000 report, “two facts are clear: [I] there is universal agreement on the need for regular mammograms after age 50, and [2] the proportion of women over age 50 following this recommendation is relatively small” (5). Several potential barriers to screening practices have been cited. In jurisdictions where health care insurance does not cover screening mammography costs for all women, the expense is frequently referred to as a barrier either to women’s acceptance of screening or to a physician’s likelihood to recommend it (6-11). ’ This research was carried out as a part of program Health Division, Alberta Health. ’ To whom reprint requests should be addressed.
initiation;
the program
is funded
0091-7435192
$5.00
by the Public
405 Copyright All rights
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
406
BRYANTANDMAH
This has led to initiatives in the United States to provide low cost mammograms to several underscreened populations (12-16). Geographic accessibility of mammographic facilities has also been noted to be a barrier in some areas (17). Other barriers include a lack of awareness of benefit among women and physicians (7, 8, 10, 18, 19), failure of physicians to refer (7, 20-24), lack of a regular physician (18), and fears about perceived risks (18, 19, 22, 25). One major potential sociodemographic determinant of screening which has not been adequately explored is rural versus urban place of residence. This oversight is not a minor one; approximately 24% of the Canadian and U.S. populations live outside of urban areas (27). The consideration of this variable also requires acknowledgement that a geographic barrier to accessibility is not the only distinction to be considered. Because women with higher educational levels and family income tend to live in cities, one must also consider the potential influence of these factors on health knowledge, attitudes, and behaviors. The purpose of this report is to compare the breast cancer awareness and practices of women living in larger cities with those of women living in small towns or rural areas over an hour’s drive from these centers. The overall study was a baseline survey examining mammography utilization practices in Alberta, a Canadian province of 2.4 million people (26), and was carried out as part of the initiation of a province-wide screening program. At the time of the survey, Alberta, like much of North America, was experiencing increased mammography utilization rates. Mammography had been available for several years by physician referral. Cost was not a barrier in this context, as the provincial health care plan generally covered the cost with no expense to individual women. However, there was little knowledge of whether current use of mammography was primarily screening or diagnostic. Further, it was not known whether current patterns of population-based participation rates targeted particular sociodemographic groups of women effectively. The data presented here describe these patterns, with a focus on the contrasts between the urban and rural respondents. METHODS
The Alberta Knowledge, Attitude, and Behavior Study was a telephone interview survey carried out in the month of March 1991. Interviewers, all of whom had some prior health care or research background, were selected and trained by the study coordinator (Z.M.) and were monitored on a weekly basis throughout the survey for consistency with the designated recruitment and interview methods. The survey used a standardized questionnaire which took approximately 20 min to administer. Households in the two major urban centers (Edmonton and Calgary, with metropolitan populations of 790,000 and 680,000, respectively (26)) were selected through random digit dialing (RDD) (28). For our rural sample, we selected names using a randomized population weighted formula from all listed household numbers in two rural health units in Alberta. All of the rural locations selected were between 1 and 3 hr drive from a mammographic facility, but the women could have lived in small towns, in villages, or on farms within the area we designated as “rural.” The utilization of directory lists in these rural areas was necessary as
BREAST
CANCER
SCREENING
PRACTICES
407
lack of saturation of the exchange codes made RDD too inefficient; however, it is estimated that 97% of all households in rural Alberta have telephones, and of these at least 85% have listed numbers (29). Thus, this method would have a negligible effect in biasing the sample. The protocol required that numbers be called up to five times before they were eliminated. Once a household was reached, interviewers ascertained whether it included any females age 40 to 74. If so, we attempted to enroll her in the survey, calling back at a later time if necessary to reach the respondent. If more than one woman of the requisite age lived in the household, a standard procedure was used to decide which of the women could be enrolled as the study participant. The other criterion for eligibility, besides age and gender, was lack of a personal history of breast cancer. Respondents were asked about their own breast health history, their knowledge of breast cancer and breast cancer screening, and their personal breast cancer screening practices. Women were asked questions on their perceptions about mammograms, their perceived susceptibility to breast cancer, and their intentions to have future mammograms. Some information was also collected on their demographic characteristics and their usual sources of health care. Questionnaires were checked throughout the data collection period for completeness and accuracy. They were then entered using SAS-FSP, and verification checks were run on the resultant data set. For this study, the knowledge, attitudes, and behaviors of rural and urban women were compared using the frequency procedures in SAS (30). Mantel-Haenszel estimates were used for comparisons of proportions and the calculation of adjusted prevalence rate ratios and confidence intervals; homogeneity of stratified rates was addressed using Breslow-Day testing for homogeneity. As noted above, there is some discrepancy in the optimal age recommendations for screening behaviors, which may affect the likelihood of screening activities taking place among women in the controversial age ranges (below 50 and over 69). Thus, analyses of breast cancer and health variables were analyzed first using the entire sample and then focusing only on the women for whom screening recommendations are consistent, i.e., women age 50 to 69. RESULTS
The interviewers reached women in 1,741 age-eligible households, of whom 1,350 participated in the study, for a response rate of 78%. Response rate was not significantly different in rural or urban areas. Of the respondents, 38 were ineligible due to a personal history of breast cancer, and 39 either declined to give their age or date of birth at the end of the interview or revealed then that they were not actually in the target age range. Thus, the total sample remaining is 1,273 women, of whom 538 were rural and 735 were urban. Although the rural and urban sample had roughly the same age distribution, the rural women were of lower income and education and were more frequently married (Table 1). Employment status was marginally different between the two groups as well. These data reflect the actual differences between the urban and the
408
BRYANT
AND MAH
TABLE DESCRIPTIVE
40-49 50-59 60-69 70-74 Education Less than high school High school graduate Post-high school University or college degree Marital status % married Employment status % employed out of home Incomeb
MEDICINE
Breast Cancer
21, 405-418
(1992)
Screening Attitudes and Behaviors and Urban Women’
HEATHER BRYANT, Alberta
Program
M.D.
PH.D. ,’ AND ZEVA
MAH,
for the Earl-y Detection of Cancer, Alberta Cancer 1040-7 Avenue SW, Calgary, AB, Canada ZIP 3G9
of Rural M.Sc.
Board,
120,
This study was carried out to assess the breast cancer knowledge, attitudes, and awareness of women age 40 to 74 in Alberta, a Canadian province of 2.4 million people. This analysis compares the attributes of 538 rural women, defined as those living between 1 and 3 hr drive from the major cities in Alberta, and 735 urban women who lived in one of these two cities. Rural women were found to have the same basic knowledge of breast cancer or perceptions of barriers to mammography, but had more negative attitudes about breast cancer itself. Despite their similar access to physician care, they were less likely to have had a recent clinical breast examination or mammogram (P < 0.001). These differences remained when adjustment was made for demographic background variables; the adjusted prevalence rate ratio for a screening mammogram in the past 2 years was 0.52 (95% C.I., 0.43,0.64), and for intention to have a mammogram in the next 2 years, 0.75 (0.63.0.90). The results suggest particular program delivery strategies when planning for provision of breast screening information and service to the large subgroup of rural women. o 1992 Academic press, hc.
INTRODUCTION
Breast cancer is one of the leading causes of death for women in developed countries. In the United States and Canada, more than 415,000 deaths from breast cancer are expected to occur over the next year (1, 2). Canadian statistics show breast cancer as the leading cause of potential years of life lost (3) and, as the most important cause of cancer deaths among women, still exceeding even lung cancer in mortality (1). Because of this, and the potential impact mammographic screening could have in altering these statistics, initiatives have been underway in many developed countries to encourage screening or to develop centralized, quality controlled screening programs on a regional basis (4). These vary in their criteria for age and interval of screening, largely due to unresolved questions in the literature regarding appropriate targets and methodologies. However, as noted in the Healthy People 2000 report, “two facts are clear: [I] there is universal agreement on the need for regular mammograms after age 50, and [2] the proportion of women over age 50 following this recommendation is relatively small” (5). Several potential barriers to screening practices have been cited. In jurisdictions where health care insurance does not cover screening mammography costs for all women, the expense is frequently referred to as a barrier either to women’s acceptance of screening or to a physician’s likelihood to recommend it (6-11). ’ This research was carried out as a part of program Health Division, Alberta Health. ’ To whom reprint requests should be addressed.
initiation;
the program
is funded
0091-7435192
$5.00
by the Public
405 Copyright All rights
0 1992 by Academic Press, Inc. of reproduction in any form reserved.
406
BRYANTANDMAH
This has led to initiatives in the United States to provide low cost mammograms to several underscreened populations (12-16). Geographic accessibility of mammographic facilities has also been noted to be a barrier in some areas (17). Other barriers include a lack of awareness of benefit among women and physicians (7, 8, 10, 18, 19), failure of physicians to refer (7, 20-24), lack of a regular physician (18), and fears about perceived risks (18, 19, 22, 25). One major potential sociodemographic determinant of screening which has not been adequately explored is rural versus urban place of residence. This oversight is not a minor one; approximately 24% of the Canadian and U.S. populations live outside of urban areas (27). The consideration of this variable also requires acknowledgement that a geographic barrier to accessibility is not the only distinction to be considered. Because women with higher educational levels and family income tend to live in cities, one must also consider the potential influence of these factors on health knowledge, attitudes, and behaviors. The purpose of this report is to compare the breast cancer awareness and practices of women living in larger cities with those of women living in small towns or rural areas over an hour’s drive from these centers. The overall study was a baseline survey examining mammography utilization practices in Alberta, a Canadian province of 2.4 million people (26), and was carried out as part of the initiation of a province-wide screening program. At the time of the survey, Alberta, like much of North America, was experiencing increased mammography utilization rates. Mammography had been available for several years by physician referral. Cost was not a barrier in this context, as the provincial health care plan generally covered the cost with no expense to individual women. However, there was little knowledge of whether current use of mammography was primarily screening or diagnostic. Further, it was not known whether current patterns of population-based participation rates targeted particular sociodemographic groups of women effectively. The data presented here describe these patterns, with a focus on the contrasts between the urban and rural respondents. METHODS
The Alberta Knowledge, Attitude, and Behavior Study was a telephone interview survey carried out in the month of March 1991. Interviewers, all of whom had some prior health care or research background, were selected and trained by the study coordinator (Z.M.) and were monitored on a weekly basis throughout the survey for consistency with the designated recruitment and interview methods. The survey used a standardized questionnaire which took approximately 20 min to administer. Households in the two major urban centers (Edmonton and Calgary, with metropolitan populations of 790,000 and 680,000, respectively (26)) were selected through random digit dialing (RDD) (28). For our rural sample, we selected names using a randomized population weighted formula from all listed household numbers in two rural health units in Alberta. All of the rural locations selected were between 1 and 3 hr drive from a mammographic facility, but the women could have lived in small towns, in villages, or on farms within the area we designated as “rural.” The utilization of directory lists in these rural areas was necessary as
BREAST
CANCER
SCREENING
PRACTICES
407
lack of saturation of the exchange codes made RDD too inefficient; however, it is estimated that 97% of all households in rural Alberta have telephones, and of these at least 85% have listed numbers (29). Thus, this method would have a negligible effect in biasing the sample. The protocol required that numbers be called up to five times before they were eliminated. Once a household was reached, interviewers ascertained whether it included any females age 40 to 74. If so, we attempted to enroll her in the survey, calling back at a later time if necessary to reach the respondent. If more than one woman of the requisite age lived in the household, a standard procedure was used to decide which of the women could be enrolled as the study participant. The other criterion for eligibility, besides age and gender, was lack of a personal history of breast cancer. Respondents were asked about their own breast health history, their knowledge of breast cancer and breast cancer screening, and their personal breast cancer screening practices. Women were asked questions on their perceptions about mammograms, their perceived susceptibility to breast cancer, and their intentions to have future mammograms. Some information was also collected on their demographic characteristics and their usual sources of health care. Questionnaires were checked throughout the data collection period for completeness and accuracy. They were then entered using SAS-FSP, and verification checks were run on the resultant data set. For this study, the knowledge, attitudes, and behaviors of rural and urban women were compared using the frequency procedures in SAS (30). Mantel-Haenszel estimates were used for comparisons of proportions and the calculation of adjusted prevalence rate ratios and confidence intervals; homogeneity of stratified rates was addressed using Breslow-Day testing for homogeneity. As noted above, there is some discrepancy in the optimal age recommendations for screening behaviors, which may affect the likelihood of screening activities taking place among women in the controversial age ranges (below 50 and over 69). Thus, analyses of breast cancer and health variables were analyzed first using the entire sample and then focusing only on the women for whom screening recommendations are consistent, i.e., women age 50 to 69. RESULTS
The interviewers reached women in 1,741 age-eligible households, of whom 1,350 participated in the study, for a response rate of 78%. Response rate was not significantly different in rural or urban areas. Of the respondents, 38 were ineligible due to a personal history of breast cancer, and 39 either declined to give their age or date of birth at the end of the interview or revealed then that they were not actually in the target age range. Thus, the total sample remaining is 1,273 women, of whom 538 were rural and 735 were urban. Although the rural and urban sample had roughly the same age distribution, the rural women were of lower income and education and were more frequently married (Table 1). Employment status was marginally different between the two groups as well. These data reflect the actual differences between the urban and the
408
BRYANT
AND MAH
TABLE DESCRIPTIVE
40-49 50-59 60-69 70-74 Education Less than high school High school graduate Post-high school University or college degree Marital status % married Employment status % employed out of home Incomeb
