44
Research paper
Health-related lifestyle factors and mammography screening attendance in a Swedish cohort study Magdalena Lagerlunda, Isabel Drakeb, Elisabet Wirfa¨ltb, Jessica M. Sontropc and Sophia Zackrissona To determine whether health-related lifestyle factors are associated with attendance at a population-based invitational mammography screening program in southern Sweden, data on health-related lifestyle factors (smoking, alcohol use, physical activity, BMI, diet, self-rated health, and stress) were obtained from the Malmo¨ Diet and Cancer Study and linked to the Malmo¨ mammography register (Sweden, 1992–2009). Women (n = 11 409) who were free from breast cancer at study entry were included in the cohort, and mammography attendance was followed from cohort entry to 31 December 2009. Generalized estimating equations were used to account for repeated measures within patients. Adjusted odds ratios (OR) and 95% confidence intervals (CI) are reported. Nonattendance occurred in 8% of the 69 746 screening opportunities that were observed. Nonattendance was more common among women who were current or former smokers [OR = 1.60 (1.45–1.76) and OR = 1.15 (1.05–1.28)], had not used alcohol in the past year [OR = 1.55 (1.32–1.83)], were less physically active outside of work [OR = 1.10 (1.00–1.20)], had high physical activity at work (OR = 1.13, 95% CI: 1.00–1.28), were vegetarians or vegans [OR = 1.49 (1.11–1.99)], had not used dietary supplements [OR = 1.11
(1.01–1.21)], had poor self-rated health [OR = 1.24 (1.14–1.36)], and were experiencing greater stress [OR = 1.25 (1.14–1.36)]. In this cohort, nonattendance was associated with smoking, alcohol abstinence, physical activity, poor self-rated health, stress, and following a vegetarian/vegan diet. These findings generally support the notion that women with less healthy lifestyles are less likely to engage in mammography screening. European c 2014 Wolters Journal of Cancer Prevention 24:44–50 Kluwer Health | Lippincott Williams & Wilkins.
Introduction
does not unanimously support this assumption. Many health-related factors have been reasonably well studied in association with mammography screening, and the general message emerging from those studies is that screening attendance is associated positively with nonsmoking and greater physical activity (for example Edwards and Boulet, 1997; Aro et al., 1999; Lagerlund et al., 2000; Maxwell et al., 2001; Lin, 2008; Gierisch et al., 2009; Caleffi et al., 2010; Lopez-de-Andres et al., 2010), but not BMI (Sutton et al., 1994; Staniscia et al., 2003; Lin, 2008; Caleffi et al., 2010) or self-rated health (for example Rodriguez et al., 1995; Maxwell et al., 1997a; Ore et al., 1997; Lagerlund et al., 2000; Lin, 2008). Associations between screening attendance and alcohol, diet, and stress are less clear. In Sweden, there are no studies reporting on associations with BMI, physical activity, and diet.
Regular mammography screening can reduce breast cancer mortality by 20–50% (Nystro¨m et al., 2002; Tabar et al., 2003). Despite this evidence, some women do not attend screening. Throughout Sweden, attendance rates range between 66 and 91%, and are generally lower in metropolitan regions (Olsson et al., 2000; Swedish Association of Local Authorities and Regions, 2013). To further improve participation rates, which is crucial for the public health impact of population-based mammography screening programs, it is important to understand and consider the factors influencing attendance. Female residents of Malmo¨, Sweden (the setting of the current study), have been invited to a population-based mammography screening program since 1990 (Zackrisson et al., 2004). Initial attendance rates [65% between 1990 and 1993 (Zackrisson et al., 2004)] were among the lowest in Sweden, making this geographical setting particularly interesting for studying factors affecting screening attendance. It seems logical to assume that women who are health conscious and lead healthier lifestyles are more likely to attend mammography screening. However, the literature c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0959-8278
European Journal of Cancer Prevention 2015, 24:44–50 Keywords: alcohol, BMI, breast cancer, diet, health behavior, lifestyle, mammography screening, physical activity, self-rated health, smoking a Department of Clinical Sciences in Malmo¨, Diagnostic Radiology, bDepartment of Clinical Sciences in Malmo¨, Research group in Nutritional Epidemiology, Lund University, Lund, Sweden and cDepartment of Epidemiology and Biostatistics, Western University, London, Ontario, Canada
Correspondence to Magdalena Lagerlund, PhD, Department of Clinical Sciences in Malmo¨, Diagnostic Radiology, Lund University, Ska˚ne University Hospital Malmo¨, Inga Marie Nilssons gata 49, SE 205 02 Malmo¨, Sweden Tel: + 1 519 433 0031; fax: + 46 40 39 10 04; e-mail: [email protected] Received 26 November 2013 Accepted 28 February 2014
Several lifestyle factors are established risk factors for breast cancer, such as high intake of alcohol (Cogliano et al., 2011), low physical activity (Monninkhof et al., 2007), and overweight and obesity (Reeves et al., 2007). Recent research shows some evidence for smoking being DOI: 10.1097/CEJ.0000000000000025
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Health factors and mammography attendance Lagerlund et al. 45
a risk factor (Cogliano et al., 2011), but findings are limited and inconclusive with respect to diet (World Cancer Research Fund/American Institute for Cancer Research, 2010). Stress has been shown to be associated with breast cancer survival, but not with incidence (Chida et al., 2008). Further, obese women may benefit more from mammography screening than women of normal weight in terms of breast cancer survival (Olsson et al., 2009a, 2009b). Regular mammography screening to increase chances of diagnosis of breast cancer at an early stage appears especially important among women with one or more of these risk factors. In the current study, we prospectively followed a large cohort of Swedish women who received regular invitations to attend mammography screening. We examined whether pre-existing health-related lifestyle factors were associated with mammography attendance over the subsequent two decades.
Methods Design, setting, and population
In this community-based cohort study, we linked data from the Malmo¨ Diet and Cancer Study (MDCS) to the Malmo¨ Mammographic Screening Register. Data sources and sample selection are described below. This study was approved by the ethics committee at Lund University.
questionnaire, which was checked for missing answers during the second study visit a few weeks later (Manjer et al., 2001, 2002). Sample
For the purpose of this study, we selected women who completed the MDCS baseline questionnaire between 17 February 1992 and 25 September 1996 (second and third questionnaire versions), and who had been invited to the mammographic screening program in Malmo¨ between baseline and the end of follow-up (date of death, date of emigration from Sweden, date of breast cancer diagnosis, or 31 December 2009, whichever came first). Women who had been diagnosed with breast cancer before baseline were excluded. Details of the different steps of exclusion, resulting in a final sample of 11 409 individuals, are presented elsewhere (Lagerlund et al., 2013). Measures and definitions Outcome: screening attendance
Nonattendance at mammography screening is the outcome variable of interest in this study and the individual mammography invitation was used as the unit of analysis. Among the 11 409 participants included in the study sample, there were a total of 69 746 screening opportunities (invitations) during follow-up. Of these, 5552 (8%) resulted in nonattendance and 64 194 (92%) resulted in attendance.
Data sources
The Malmo¨ Mammographic Screening Program is a population-based mammography screening program that was established in Malmo¨, Sweden, in 1990. Invitations to attend mammography screening are mailed at intervals of 1.5–2 years. Inclusion of all eligible female residents is ensured by continuous updates from the population register. Because of changes in national recommendations, the age groups invited have varied somewhat over the years. Between 1990 and 1998, women in the age range 50–69 years were invited. In 1999, the upper age limit was extended to 74 years, and in 2009, the lower age limit was decreased to 40 years. The MDCS is a prospective cohort study, carried out in Malmo¨, a city in southern Sweden with a growing population of about 308 000 in 2012 (Statistics Sweden, 2014). The primary goal of the MDCS is to investigate the association between diet and different types of cancer. Recruitment started in 1991 and was primarily performed by postal invitation at random from the source population of Malmo¨ residents born between 1926 and 1945. In 1995, recruitment was extended to include women born between 1923 and 1950. Approximately 18% of the respondents joined the study spontaneously as a result of a passive recruitment campaign. A total of 17 035 women had joined the study at the end of recruitment in the autumn of 1996, resulting in a final response rate of 42.6%. At baseline, all participants received a health
Health-related lifestyle factors
From the baseline MDCS health questionnaire and a dietary assessment, we retrieved information on the following health-related lifestyle factors: smoking, alcohol use, physical activity, BMI, diet, self-rated health, and stress. Smoking was categorized into current (regular and occasional), former, and never. Information on alcohol intake was drawn from the questionnaire and a 7-day menu book (Wallstrom et al., 2000). We categorized alcohol consumption into no use in the past year, low (< 15 g/day), moderate (15–30 g/day), and high (> 30 g/day) (Mattisson et al., 2004). Leisure-time physical activity was measured using a modified questionnaire, adapted from the Minnesota Leisure Time Physical Activity Questionnaire (Taylor et al., 1978). Participants reported the number of minutes per week spent on each of 18 different activities during the four seasons. These numbers were multiplied by activity-specific intensity coefficients and added to create an overall score (Wallstrom et al., 2000). Individuals were ranked on the basis of their scores and divided into quartiles, which were dichotomized into low physical activity (Q1) and physically active (Q2–Q4) (Calling et al., 2006). We also included a variable on physical activity at work where respondents estimated the usual level of physical activity required for work (very low, low, medium, high, very high), categorized into low, medium, and high.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
46
European Journal of Cancer Prevention 2015, Vol 24 No 1
Height and weight were measured directly at baseline examination in light clothing without shoes and BMI was calculated in kg/m2. We analyzed the following WHOdefined BMI categories: underweight (< 18.5), normal (18.5–24.9), preobese (25.0–29.9), and obese (Z 30). Three different variables related to diet were considered. Information on participants being on a special diet (none, vegetarian or vegan, diabetic, or other) was collected from the baseline MDCS health questionnaire. The detailed dietary assessment (using a modified diet history methodology) included a 7-day menu book, a 168-item quantitative food frequency questionnaire, and a 1-h interview (Wirfalt et al., 2002). Information on dietary supplements was based on reported use (any/none) in the 7-day menu book. To assess overall diet, we used a diet quality index, which has been described in detail elsewhere (Drake et al., 2011). The diet quality index was based on adherence to the Swedish nutrition recommendations for the following components: saturated fatty acids (r 14% of nonalcohol energy), polyunsaturated fatty acids (5–10% of nonalcohol energy), fish and shellfish (Z 300 g/week), dietary fiber (2.4–3.6 g/MJ), fruit and vegetables (Z 400 g/day), and sucrose (r 10% of nonalcohol energy). Adherence to the recommended intake for each component was assigned 1 point, for a total score of 0–6. The index was further categorized into low quality (0–1 points), medium quality (2–3 points), and high quality (4–6 points). Self-rated physical and mental health was rated on a seven-point scale, dichotomized into poor health (1–4) and good health (5–7). Finally, participants were asked whether they had experienced recent stress or mental pressure outside of work (yes/no). Statistical analysis
We estimated odds ratios (ORs) and 95% confidence intervals (CIs) for mammography attendance from binary generalized estimating equation models where adjustments are made for the correlation of repeated measures within participants (with an autoregressive correlation structure). We first carried out age-adjusted analyses to determine the effects of the different health-related lifestyle factors on the odds of nonattendance. These effects were then further adjusted by baseline year, season, number of screening invitations after baseline, having been invited to screening before baseline, and sociodemographic factors, including education, employment status, cohabitation, and country of birth (multivariate model 1). The dietary index was also adjusted for total energy intake. In a final step, variables were further adjusted for all health-related variables (multivariate model 2). Interaction terms were tested in model 2 to examine possible effect modification by stress and selfreported health, and between menopausal status and BMI.
Results A total of 69 746 screening invitations were sent out to study participants, ranging between 1 and 12 per participant.
Overall, the attendance rate for the entire study period was 92% and for individual calendar years, the mammography attendance rate varied between 88 and 95% (data not shown). Descriptive sample characteristics are presented in Table 1. Women were on average 55 years old (range: 44–70) at baseline. The mean age at first screening opportunity was 56 years (range: 45–71). About one-third of women had completed high school, two-thirds were living with a partner, two-thirds were employed, and 88% were born in Sweden. In terms of occupational status (present or most recent position), 55% were nonmanual workers and 37% were manual workers. Relationship between risk factors and mammography nonattendance
Associations between health-related lifestyle factors and mammography nonattendance are presented in Table 2. Sociodemographic and screening characteristics of the study sample (n = 11 409), Malmo¨, Sweden (1992–2009)
Table 1
Sociodemographic variables Mean age at baseline (SD) Mean age at first subsequent screening invitation (SD) Age group at baseline (years) 44–49 50–54 55–59 60–64 65–70 Education level High school or higher Less than high school Missing Cohabitation (with partner) Yes No Missing Occupation (present or latest job) Self-employed/employer/farmer Higher nonmanual Middle nonmanual Lower nonmanual Skilled manual Unskilled manual Missing Employment status Employed Not employed Missing Country of birth Sweden Other Missing Invited to screening program before baseline Yes No Number of screening invitations after baseline 1 2 3 4 5 6 7 8 9 10 11 12
N (%) 54.9 (6.7) 56.7 (6.4) 3442 2470 2086 2155 1256
(30.2) (21.6) (18.3) (18.9) (11.0)
3717 (32.7) 7665 (67.3) 27 7871 (69.0) 3531 (31.0) 7 885 (7.8) 741 (6.6) 2049 (18.1) 3489 (30.9) 803 (7.1) 3341 (29.5) 101 7695 (67.6) 3681 (32.4) 33 10 066 (88.3) 1336 (11.7) 7 8459 (74.1) 2950 (25.9) 793 602 568 808 1271 1626 1896 1921 1247 609 67 1
(7.0) (5.3) (5.0) (7.1) (11.1) (14.3) (16.6) (16.8) (10.9) (5.3) (0.6) (0.0)
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Health factors and mammography attendance Lagerlund et al. 47
Table 2
Health-related lifestyle factors in relation to mammography screening nonattendance (Malmo¨, Sweden, 1992–2009) Odds ratios for screening nonattendance (95% CI)
Health-related lifestyle factors Smoking Current Former Never Missing Alcohol use None in past year Low (< 15 g/day) Medium (15–30 g/day) High (> 30 g/day) Physical activity level (leisure) Low (Q1) Active (Q2–Q4) Missing Physical activity at work Low Medium High Missing BMI Below normal (< 18.5) Normal (18.5–24.99) Preobese (25.0–29.99) Obese (Z 30) Missing Special diet None Vegetarian or vegan Diabetic or other Missing Dietary supplement use No Yes Dietary index score (0–6) Low (0–1) Medium (2–3) High (4–6) Self-rated health Poor (1–4) Good (5–7) Missing Stress (non-work-related) Yes No Missing
Age adjusted
Multivariate Model 1a
Multivariate Model 2b
3369 (29.5) 3280 (28.8) 4759 (41.7) 1
1.72 (1.57–1.89) 1.15 (1.04–1.27) Ref
1.65 (1.50–1.81) 1.14 (1.03–1.25) Ref
1.60 (1.45–1.76) 1.15 (1.05–1.28) Ref
738 8613 1767 291
(6.6) (75.5) (15.5) (2.6)
1.78 (1.53–2.06) Ref 1.03 (0.93–1.15) 1.31 (1.05–1.64)
1.63 (1.40–1.89) Ref 1.06 (0.96–1.18) 1.24 (0.99–1.57)
1.55 (1.32–1.83) Ref 1.06 (0.95–1.18) 1.14 (0.91–1.44)
2838 (25.0) 8516 (75.0) 55
1.18 (1.08–1.29) Ref
1.18 (1.08–1.29) Ref
1.10 (1.00–1.20) Ref
6551 (58.5) 3376 (30.1) 1275 (11.4) 207
Ref 1.05 (0.96–1.15) 1.35 (1.20–1.52)
Ref 1.03 (0.94–1.13) 1.24 (1.10–1.40)
Ref 1.00 (0.92–1.10) 1.13 (1.00–1.28)
148 (1.3) 5979 (52.5) 3733 (32.8) 1534 (13.5) 15
1.41 (1.00–2.00) Ref 0.89 (0.90–1.07) 1.20 (1.07–1.36)
1.37 (0.96–1.94) Ref 0.96 (0.88–1.05) 1.14 (1.01–1.29)
1.27 (0.89–1.81) Ref 0.97 (0.89–1.07) 1.10 (0.97–1.25)
10 771 (94.8) 171 (1.5) 423 (3.7) 44
Ref 1.83 (1.40–2.40) 1.34 (1.11–1.62)
Ref 1.40 (1.07–1.84) 1.18 (0.97–1.42)
Ref 1.49 (1.11–1.99) 1.12 (0.92–1.36)
2951 (25.9) 8458 (74.1)
1.08 (0.99–1.18) Ref
1.09 (1.00–1.19) Ref
1.11 (1.01–1.21) Ref
1783 (15.6) 6143 (53.8) 3483 (30.5)
1.18 (1.04–1.33) 1.12 (1.03–1.23) Ref
1.13 (1.00–1.28) 1.11 (1.01–1.21) Ref
1.02 (0.90–1.16) 1.05 (0.96–1.15) Ref
3352 (29.5) 8015 (70.5) 42
1.50 (1.39–1.63) Ref
1.37 (1.26–1.49) Ref
1.24 (1.14–1.36) Ref
3830 (33.7) 7548 (66.3) 31
1.49 (1.38–1.61) Ref
1.35 (1.24–1.46) Ref
1.25 (1.14–1.36) Ref
N (%) in study cohort
CI, confidence interval. a Adjusted for baseline year, season, screening sequence number after baseline, having been invited to screening before baseline, and sociodemographic factors (age, education, cohabitation, employment status, and country of birth). The dietary index score was further adjusted by the total energy intake. b Adjusted for all other included variables in addition to baseline year, season, invitation number, having been invited to screening before baseline, and sociodemographic factors (age, education, cohabitation, employment status, and country of birth).
When all variables were entered into model 2, many remained significantly associated with nonattendance, but effect sizes were generally reduced compared with the age-adjusted estimates and model 1 estimates. However, the effects of diet quality and BMI were no longer statistically significant. Nonattendance was more common among current smokers (OR = 1.60, 95% CI: 1.45–1.76), former smokers (OR = 1.15, 95% CI: 1.05–1.28), and those who had not consumed any alcohol in the past year (OR = 1.55, 95% CI: 1.32–1.83); however, high daily consumption of alcohol was not associated with nonattendance. Nonattendance was associated with low
leisure-time physical activity (OR = 1.10, 95% CI: 1.00–1.20) and high physical activity at work (OR = 1.13, 95% CI: 1.00–1.28). Furthermore, nonattendance was more common among vegetarians and vegans (OR = 1.49, 95% CI: 1.11–1.99), among those not taking dietary supplements (OR = 1.11, 95% CI: 1.01–1.21), among those with poor selfrated health (OR = 1.24, 95% CI: 1.14–1.36), and those experiencing higher levels of non-work-related stress (OR = 1.25, 95% CI: 1.14–1.36). Statistically significant interactions were found between stress and BMI (P < 0.01), and self-rated health and BMI (P < 0.01). Obesity was associated with nonattendance
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
48
European Journal of Cancer Prevention 2015, Vol 24 No 1
among women who were stressed (OR = 1.32, 95% CI: 1.09–1.60) and among women with poor self-rated health (OR = 1.39, 95% CI: 1.15–1.67). Among women who were stressed, nonattendance was also more common among those who were underweight (OR = 1.81, 95% CI: 1.09–2.99). No interaction was found between BMI and menopausal status (data not shown).
Discussion In this Swedish cohort of women who received regular invitations to attend mammography screening, a higher odds of nonattendance were observed among smokers, alcohol abstainers, vegetarians/vegans, women who did not use dietary supplements, women who had low leisuretime physical activity, physically demanding jobs, poor self-rated health, and those who were stressed. Underweight and obesity were found to be associated with nonattendance in some groups of women, whereas dietary quality was unrelated to attendance. Previous studies have found that health-related lifestyle factors including physical inactivity, obesity, alcohol use, stress, and possibly smoking are associated with breast cancer incidence and/or mortality, and therefore, these factors warrant extra attention. Similar to other studies, we found that nonattendance was associated with smoking (Rakowski et al., 1993; Beaulieu et al., 1996; Martin et al., 1996; Edwards and Boulet 1997; Maxwell et al., 1997b, 2001; Aro et al., 1999; Fredman et al., 1999; Lagerlund et al., 2000; Bancej et al., 2005; Jelinski et al., 2005; Lin, 2008; Gierisch et al., 2009; Caleffi et al., 2010; Lopez-de-Andres et al., 2010) and lower leisuretime physical activity (Fajardo et al., 1992; Rakowski et al., 1993; Maxwell et al., 1997b, 2001; Lin, 2008; Lopez-deAndres et al., 2010). Although there is some evidence that obesity may be associated with mammography attendance (Lopez-de-Andres et al., 2010), we, along with most other studies (Sutton et al., 1994; Staniscia et al., 2003; Lin, 2008; Caleffi et al., 2010), did not find an association between BMI and adherence to mammography screening when analyzing the entire sample. However, we did find statistically significant associations with BMI in subgroups of women who had poor self-rated health and who were stressed. Therefore, there is reason also to consider obesity when targeting groups whose screening attendance needs improvement. Contrary to some studies that found no association between the use of alcohol and mammography attendance (Maxwell et al., 1997b; Aro et al., 1999; Lopez-deAndres et al., 2010), but in corroboration with other studies (Rakowski et al., 1993; Sutton et al., 1994; Fredman et al., 1999; Lagerlund et al., 2000), we found that alcohol abstainers were less likely to attend mammography screening than those drinking some alcohol. Alcohol abstinence may be associated with previous alcohol abuse, poor health, or religious beliefs. Many studies
show a J-shaped curve between alcohol consumption and health outcomes, where both abstainers and heavy drinkers have worse health outcomes than light drinkers. The absence of an association between greater alcohol intake and mammography attendance in this study may be because of under-reporting of alcohol intake, which is a common problem in surveys (Feunekes et al., 1999). The reported absolute intakes of total alcohol were considerably lower than the reference method used to evaluate the relative validity of the MDCS method (Riboli et al., 1997). In addition, heavy drinkers may be less likely to participate in research studies, resulting in less variability on this measure. Hence, we do not dismiss the possibility that nonattendance could be associated with high alcohol consumption. To our knowledge, no study has specifically investigated how vegetarianism or veganism relates to mammography screening attendance. Our finding that vegetarians and vegans were less likely to attend screening may appear counterintuitive as these diets may indicate a certain level of health consciousness; however, the decision to follow a vegetarian or a vegan diet may reflect an individual’s ethical beliefs and is not necessarily health related. It is important to point out that the number of vegetarians and vegans in this study was small [n = 171 (6%)] and the data were based on self-reported information. An alternative explanation for higher nonattendance among vegetarians/vegans may be that these women feel a false sense of protection from disease in general or breast cancer in particular because of their health behaviors. However, we found no association between diet quality and mammography attendance. Although some studies have found associations with varying aspects of diet (Maclean et al., 1984; Fajardo et al., 1992; Flamant et al., 2006; Lopez-de-Andres et al., 2010), others have not (Sutton et al., 1994; Ore et al., 1997; Hagoel et al., 1999). We found that women with high stress and poor self-rated health were less likely to attend mammography screening. In the case of self-rated health, our finding contradicts a large body of research (Lin et al., 1990; Fajardo et al., 1992; Sutton et al., 1994; Rodriguez et al., 1995; Beaulieu et al., 1996; Rimer et al., 1996; Maxwell et al., 1997b; Ore et al., 1997; Michielutte et al., 1999; Lagerlund et al., 2000), although some studies have reported similar findings (Calnan 1984; Fredman et al., 1999). For both poor selfrated health and stress, lack of time and energy may explain the lower odds of attending mammography screening. As stress is associated with lower breast cancer survival (Chida et al., 2008), it is important to explore ways of encouraging women with higher levels of stress to lend higher priority to regular screening.
Strengths and limitations
The overall yearly attendance rate of 92% in this cohort is considerably higher than the rates reported in earlier
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Health factors and mammography attendance Lagerlund et al. 49
studies, which were closer to 65% (Matson et al., 2001; Zackrisson et al., 2004). This is likely because the MDCS cohort is selected toward better health, and possibly higher socioeconomic status and a healthier lifestyle than the general population (Manjer et al., 2001). Further, because of the prospective nature of this study and the long follow-up period (median follow-up: 15 years; range: 4–18), it is possible that women’s health behaviors may have changed between baseline and screening. However, this type of exposure misclassification would likely attenuate rather than inflate associations with the outcome. Despite these limitations, our study has many strengths, including its large size and prospective follow-up of screening attendance. Information on health-related lifestyle factors was collected independently of mammography screening among women who were free of breast cancer at baseline, which reduces the potential for recall bias. Further, although many studies rely on self-reported mammography attendance, we obtained this information from register data, which would minimize measurement error and misclassification. Conclusion
Our study of Swedish women who received regular invitations for mammography screening generally supports the notion that those with less healthy lifestyles are less likely to engage in mammography screening. In particular, nonattendance was associated with smoking, leisure-time physical inactivity, poor self-rated health, and stress. Also, women who were vegetarian or vegan were significantly less likely to attend screening; however, given the small size of this subgroup, this finding should be investigated more closely in future research. Dietary quality was unrelated to attendance, and underweight and obesity were associated with nonattendance only in some groups of women. Efforts to optimize mammography screening should consider methods of targeting screening referrals especially to women who smoke, are less physically active and obese, as these constitute risk factors for breast cancer.
Acknowledgements The authors are grateful to Dr Jonas Manjer for guidance on the database, Anders Dahlin for MDCS database management, and Darek Gozdzik for his work on the dataset that formed the basis for this study. This research was funded by the Swedish Cancer Society. Conflicts of interest
There are no conflicts of interest.
References Aro AR, de Koning HJ, Absetz P, Schreck M (1999). Psychosocial predictors of first attendance for organised mammography screening. J Med Screen 6:82–88. Bancej CM, Maxwell CJ, Onysko J, Eliasziw M (2005). Mammography utilization in Canadian women aged 50 to 69: identification of factors that predict initiation and adherence. Can J Public Health 96:364–368.
Beaulieu MD, Beland F, Roy D, Falardeau M, Hebert G (1996). Factors determining compliance with screening mammography. CMAJ 154:1335–1343. Caleffi M, Ribeiro RA, Bedin AJ Jr, Viegas-Butzke JM, Baldisserotto FD, Skonieski GP, et al. (2010). Adherence to a breast cancer screening program and its predictors in underserved women in southern Brazil. Cancer Epidemiol Biomarkers Prev 19:2673–2679. Calling S, Hedblad B, Engstrom G, Berglund G, Janzon L (2006). Effects of body fatness and physical activity on cardiovascular risk: risk prediction using the bioelectrical impedance method. Scand J Public Health 34:568–575. Calnan M (1984). The health belief model and participation in programmes for the early detection of breast cancer: a comparative analysis. Soc Sci Med 19:823–830. Chida Y, Hamer M, Wardle J, Steptoe A (2008). Do stress-related psychosocial factors contribute to cancer incidence and survival? Nat Clin Pract Oncol 5:466–475. Cogliano VJ, Baan R, Straif K, Grosse Y, Lauby-Secretan B, El Ghissassi F, et al. (2011). Preventable exposures associated with human cancers. J Natl Cancer Inst 103:1827–1839. Drake I, Gullberg B, Ericson U, Sonestedt E, Nilsson J, Wallstrom P, et al. (2011). Development of a diet quality index assessing adherence to the Swedish nutrition recommendations and dietary guidelines in the Malmo Diet and Cancer cohort. Public Health Nutr 14:835–845. Edwards N, Boulet J (1997). Implementing breast cancer screening guidelines: results of the Ontario Health Status Survey. Am J Prev Med 13:143–149. Fajardo LL, Saint-Germain M, Meakem TJ 3rd, Rose C, Hillman BJ (1992). Factors influencing women to undergo screening mammography. Radiology 184:59–63. Feunekes GI, van’t Veer P, van Staveren WA, Kok FJ (1999). Alcohol intake assessment: the sober facts. Am J Epidemiol 150:105–112. Flamant C, Gauthier E, Clavel-Chapelon F (2006). Determinants of noncompliance to recommendations on breast cancer screening among women participating in the French E3N cohort study. Eur J Cancer Prev 15:27–33. Fredman L, Sexton M, Cui Y, Althuis M, Wehren L, Hornbeck P, et al. (1999). Cigarette smoking, alcohol consumption, and screening mammography among women ages 50 and older. Prev Med 28:407–417. Gierisch JM, O’Neill SC, Rimer BK, DeFrank JT, Bowling JM, Skinner CS (2009). Factors associated with annual-interval mammography for women in their 40s. Cancer Epidemiol 33:72–78. Hagoel L, Ore L, Neter E, Shifroni G, Rennert G (1999). The gradient in mammography screening behavior: a lifestyle marker. Soc Sci Med 48:1281–1290. Jelinski SE, Maxwell CJ, Onysko J, Bancej CM (2005). The influence of breast self-examination on subsequent mammography participation. Am J Public Health 95:506–511. Lagerlund M, Sparen P, Thurfjell E, Ekbom A, Lambe M (2000). Predictors of nonattendance in a population-based mammography screening programme; socio-demographic factors and aspects of health behaviour. Eur J Cancer Prev 9:25–33. Lagerlund M, Sontrop JM, Zackrisson S (2013). Do reproductive and hormonal risk factors for breast cancer associate with attendance at mammography screening? Cancer Causes Control 24:1687–1694. Lin DY, Wei LJ, Zing Z (1990). Checking the Cox model with cumulative sums of martingale-based residuals. Biometrika 80:557–572. Lin SJ (2008). Factors influencing the uptake of screening services for breast and cervical cancer in Taiwan. J R Soc Promot Health 128:327–334. Lopez-de-Andres A, Martin-Lopez R, Hernandez-Barrera V, Carrasco-Garrido P, Gil-de-Miguel A, Esteban y Pena MM, et al. (2010). Predictors of breast and cervical cancer screening in a Spanish metropolitan area. J Womens Health (Larchmt) 19:1675–1681. Maclean U, Sinfield D, Klein S, Harnden B (1984). Women who decline breast screening. J Epidemiol Community Health 38:278–283. Manjer J, Carlsson S, Elmstahl S, Gullberg B, Janzon L, Lindstrom M, et al. (2001). The Malmo Diet and Cancer Study: representativity, cancer incidence and mortality in participants and non-participants. Eur J Cancer Prev 10:489–499. Manjer J, Elmstahl S, Janzon L, Berglund G (2002). Invitation to a populationbased cohort study: differences between subjects recruited using various strategies. Scand J Public Health 30:103–112. Martin LM, Calle EE, Wingo PA, Heath CW Jr (1996). Comparison of mammography and Pap test use from the 1987 and 1992 National Health Interview Surveys: are we closing the gaps? Am J Prev Med 12:82–90. Matson S, Andersson I, Berglund G, Janzon L, Manjer J (2001). Nonattendance in mammographic screening: a study of intraurban differences in Malmo, Sweden, 1990–1994. Cancer Detect Prev 25:132–137. Mattisson I, Wirfalt E, Wallstrom P, Gullberg B, Olsson H, Berglund G (2004). High fat and alcohol intakes are risk factors of postmenopausal breast cancer: a prospective study from the Malmo diet and cancer cohort. Int J Cancer 110:589–597.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
50
European Journal of Cancer Prevention 2015, Vol 24 No 1
Maxwell AE, Bastani R, Warda US (1997a). Breast cancer screening and related attitudes among Filipino-American women. Cancer Epidemiol Biomarkers Prev 6:719–726. Maxwell CJ, Kozak JF, Desjardins-Denault SD, Parboosingh J (1997b). Factors important in promoting mammography screening among Canadian women. Can J Public Health 88:346–350. Maxwell CJ, Bancej CM, Snider J (2001). Predictors of mammography use among Canadian women aged 50–69: findings from the 1996/97 National Population Health Survey. CMAJ 164:329–334. Michielutte R, Dignan MB, Smith BL (1999). Psychosocial factors associated with the use of breast cancer screening by women age 60 years or over. Health Educ Behav 26:625–647. Monninkhof EM, Elias SG, Vlems FA, van der Tweel I, Schuit AJ, Voskuil DW, et al. (2007). Physical activity and breast cancer: a systematic review. Epidemiology 18:137–157. Nystro¨m L, Andersson I, Bjurstam N, Frisell J, Nordenskjold B, Rutqvist L (2002). Long-term effects of mammography screening: updated overview of the Swedish randomised trials. Lancet 359:909–919. Olsson A, Garne JP, Tengrup I, Zackrisson S, Manjer J (2009a). Body mass index and breast cancer survival in relation to the introduction of mammographic screening. Eur J Surg Oncol 35:1261–1267. Olsson A, Garne JP, Tengrup I, Zackrisson S, Manjer J (2009b). Overweight in relation to tumour size and axillary lymph node involvement in postmenopausal breast cancer patients: differences between women invited to vs. not invited to mammography in a randomized screening trial. Cancer Epidemiol 33:9–15. Olsson S, Andersson I, Karlberg I, Bjurstam N, Frodis E, Hakansson S (2000). Implementation of service screening with mammography in Sweden: from pilot study to nationwide programme. J Med Screen 7:14–18. Ore L, Hagoel L, Shifroni G, Rennert G (1997). Compliance with mammography screening in Israeli women: the impact of a pre-scheduled appointment and of the letter-style. Isr J Med Sci 33:103–111. Rakowski W, Rimer BK, Bryant SA (1993). Integrating behavior and intention regarding mammography by respondents in the 1990 National Health Interview Survey of Health Promotion and Disease Prevention. Public Health Rep 108:605–624. Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D (2007). Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study. BMJ 335:1134.
Riboli E, Elmstahl S, Saracci R, Gullberg B, Lindgarde F (1997). The Malmo Food Study: validity of two dietary assessment methods for measuring nutrient intake. Int J Epidemiol 26 (Suppl 1):S161–S173. Rimer BK, Conaway MR, Lyna PR, Rakowski W, Woods-Powell CT, Tessaro I, et al. (1996). Cancer screening practices among women in a community health center population. Am J Prev Med 12:351–357. Rodriguez C, Plasencia A, Schroeder DG (1995). Predictive factors of enrollment and adherence in a breast cancer screening program in Barcelona (Spain). Soc Sci Med 40:1155–1160. Staniscia T, Manzoli LM, Di Giovanni P, Dragani V, Testa P, Cavaliere D, et al. (2003). Factors related to the uptake of breast cancer screening (mammography and breast ultrasound): a retrospective survey on a sample of resident women, 50–70 years aged, from Abruzzo region. Ann Ig 15:1063–1075. Statistics Sweden (2014). Population by municipality, marital status and sex. Years 1968–2013. Stockholm, Sweden: Statistics Sweden. Sutton S, Bickler G, Sancho-Aldridge J, Saidi G (1994). Prospective study of predictors of attendance for breast screening in inner London. J Epidemiol Community Health 48:65–73. Swedish Association of Local Authorities and Regions (2013). Efforts for early detection. Even better cancer care – Interim report 4. Stockholm, Sweden: Swedish Association of Local Authorities and Regions. Tabar L, Yen MF, Vitak B, Chen HH, Smith RA, Duffy SW (2003). Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening. Lancet 361:1405–1410. Taylor HL, Jacobs DR Jr, Schucker B, Knudsen J, Leon AS, Debacker G (1978). A questionnaire for the assessment of leisure time physical activities. J Chronic Dis 31:741–755. Wallstrom P, Wirfalt E, Janzon L, Mattisson I, Elmstahl S, Johansson U, et al. (2000). Fruit and vegetable consumption in relation to risk factors for cancer: a report from the Malmo Diet and Cancer Study. Public Health Nutr 3:263–271. Wirfalt E, Mattisson I, Johansson U, Gullberg B, Wallstrom P, Berglund G (2002). A methodological report from the Malmo Diet and Cancer study: development and evaluation of altered routines in dietary data processing. Nutr J 1:3. World Cancer Research Fund/American Institute for Cancer Research (2010). Continuous Update Project Report summary. Food, nutrition, physical activity, and the prevention of breast cancer. Zackrisson S, Andersson I, Manjer J, Janzon L (2004). Non-attendance in breast cancer screening is associated with unfavourable socio-economic circumstances and advanced carcinoma. Int J Cancer 108:754–760.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Research paper
Health-related lifestyle factors and mammography screening attendance in a Swedish cohort study Magdalena Lagerlunda, Isabel Drakeb, Elisabet Wirfa¨ltb, Jessica M. Sontropc and Sophia Zackrissona To determine whether health-related lifestyle factors are associated with attendance at a population-based invitational mammography screening program in southern Sweden, data on health-related lifestyle factors (smoking, alcohol use, physical activity, BMI, diet, self-rated health, and stress) were obtained from the Malmo¨ Diet and Cancer Study and linked to the Malmo¨ mammography register (Sweden, 1992–2009). Women (n = 11 409) who were free from breast cancer at study entry were included in the cohort, and mammography attendance was followed from cohort entry to 31 December 2009. Generalized estimating equations were used to account for repeated measures within patients. Adjusted odds ratios (OR) and 95% confidence intervals (CI) are reported. Nonattendance occurred in 8% of the 69 746 screening opportunities that were observed. Nonattendance was more common among women who were current or former smokers [OR = 1.60 (1.45–1.76) and OR = 1.15 (1.05–1.28)], had not used alcohol in the past year [OR = 1.55 (1.32–1.83)], were less physically active outside of work [OR = 1.10 (1.00–1.20)], had high physical activity at work (OR = 1.13, 95% CI: 1.00–1.28), were vegetarians or vegans [OR = 1.49 (1.11–1.99)], had not used dietary supplements [OR = 1.11
(1.01–1.21)], had poor self-rated health [OR = 1.24 (1.14–1.36)], and were experiencing greater stress [OR = 1.25 (1.14–1.36)]. In this cohort, nonattendance was associated with smoking, alcohol abstinence, physical activity, poor self-rated health, stress, and following a vegetarian/vegan diet. These findings generally support the notion that women with less healthy lifestyles are less likely to engage in mammography screening. European c 2014 Wolters Journal of Cancer Prevention 24:44–50 Kluwer Health | Lippincott Williams & Wilkins.
Introduction
does not unanimously support this assumption. Many health-related factors have been reasonably well studied in association with mammography screening, and the general message emerging from those studies is that screening attendance is associated positively with nonsmoking and greater physical activity (for example Edwards and Boulet, 1997; Aro et al., 1999; Lagerlund et al., 2000; Maxwell et al., 2001; Lin, 2008; Gierisch et al., 2009; Caleffi et al., 2010; Lopez-de-Andres et al., 2010), but not BMI (Sutton et al., 1994; Staniscia et al., 2003; Lin, 2008; Caleffi et al., 2010) or self-rated health (for example Rodriguez et al., 1995; Maxwell et al., 1997a; Ore et al., 1997; Lagerlund et al., 2000; Lin, 2008). Associations between screening attendance and alcohol, diet, and stress are less clear. In Sweden, there are no studies reporting on associations with BMI, physical activity, and diet.
Regular mammography screening can reduce breast cancer mortality by 20–50% (Nystro¨m et al., 2002; Tabar et al., 2003). Despite this evidence, some women do not attend screening. Throughout Sweden, attendance rates range between 66 and 91%, and are generally lower in metropolitan regions (Olsson et al., 2000; Swedish Association of Local Authorities and Regions, 2013). To further improve participation rates, which is crucial for the public health impact of population-based mammography screening programs, it is important to understand and consider the factors influencing attendance. Female residents of Malmo¨, Sweden (the setting of the current study), have been invited to a population-based mammography screening program since 1990 (Zackrisson et al., 2004). Initial attendance rates [65% between 1990 and 1993 (Zackrisson et al., 2004)] were among the lowest in Sweden, making this geographical setting particularly interesting for studying factors affecting screening attendance. It seems logical to assume that women who are health conscious and lead healthier lifestyles are more likely to attend mammography screening. However, the literature c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0959-8278
European Journal of Cancer Prevention 2015, 24:44–50 Keywords: alcohol, BMI, breast cancer, diet, health behavior, lifestyle, mammography screening, physical activity, self-rated health, smoking a Department of Clinical Sciences in Malmo¨, Diagnostic Radiology, bDepartment of Clinical Sciences in Malmo¨, Research group in Nutritional Epidemiology, Lund University, Lund, Sweden and cDepartment of Epidemiology and Biostatistics, Western University, London, Ontario, Canada
Correspondence to Magdalena Lagerlund, PhD, Department of Clinical Sciences in Malmo¨, Diagnostic Radiology, Lund University, Ska˚ne University Hospital Malmo¨, Inga Marie Nilssons gata 49, SE 205 02 Malmo¨, Sweden Tel: + 1 519 433 0031; fax: + 46 40 39 10 04; e-mail: [email protected] Received 26 November 2013 Accepted 28 February 2014
Several lifestyle factors are established risk factors for breast cancer, such as high intake of alcohol (Cogliano et al., 2011), low physical activity (Monninkhof et al., 2007), and overweight and obesity (Reeves et al., 2007). Recent research shows some evidence for smoking being DOI: 10.1097/CEJ.0000000000000025
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Health factors and mammography attendance Lagerlund et al. 45
a risk factor (Cogliano et al., 2011), but findings are limited and inconclusive with respect to diet (World Cancer Research Fund/American Institute for Cancer Research, 2010). Stress has been shown to be associated with breast cancer survival, but not with incidence (Chida et al., 2008). Further, obese women may benefit more from mammography screening than women of normal weight in terms of breast cancer survival (Olsson et al., 2009a, 2009b). Regular mammography screening to increase chances of diagnosis of breast cancer at an early stage appears especially important among women with one or more of these risk factors. In the current study, we prospectively followed a large cohort of Swedish women who received regular invitations to attend mammography screening. We examined whether pre-existing health-related lifestyle factors were associated with mammography attendance over the subsequent two decades.
Methods Design, setting, and population
In this community-based cohort study, we linked data from the Malmo¨ Diet and Cancer Study (MDCS) to the Malmo¨ Mammographic Screening Register. Data sources and sample selection are described below. This study was approved by the ethics committee at Lund University.
questionnaire, which was checked for missing answers during the second study visit a few weeks later (Manjer et al., 2001, 2002). Sample
For the purpose of this study, we selected women who completed the MDCS baseline questionnaire between 17 February 1992 and 25 September 1996 (second and third questionnaire versions), and who had been invited to the mammographic screening program in Malmo¨ between baseline and the end of follow-up (date of death, date of emigration from Sweden, date of breast cancer diagnosis, or 31 December 2009, whichever came first). Women who had been diagnosed with breast cancer before baseline were excluded. Details of the different steps of exclusion, resulting in a final sample of 11 409 individuals, are presented elsewhere (Lagerlund et al., 2013). Measures and definitions Outcome: screening attendance
Nonattendance at mammography screening is the outcome variable of interest in this study and the individual mammography invitation was used as the unit of analysis. Among the 11 409 participants included in the study sample, there were a total of 69 746 screening opportunities (invitations) during follow-up. Of these, 5552 (8%) resulted in nonattendance and 64 194 (92%) resulted in attendance.
Data sources
The Malmo¨ Mammographic Screening Program is a population-based mammography screening program that was established in Malmo¨, Sweden, in 1990. Invitations to attend mammography screening are mailed at intervals of 1.5–2 years. Inclusion of all eligible female residents is ensured by continuous updates from the population register. Because of changes in national recommendations, the age groups invited have varied somewhat over the years. Between 1990 and 1998, women in the age range 50–69 years were invited. In 1999, the upper age limit was extended to 74 years, and in 2009, the lower age limit was decreased to 40 years. The MDCS is a prospective cohort study, carried out in Malmo¨, a city in southern Sweden with a growing population of about 308 000 in 2012 (Statistics Sweden, 2014). The primary goal of the MDCS is to investigate the association between diet and different types of cancer. Recruitment started in 1991 and was primarily performed by postal invitation at random from the source population of Malmo¨ residents born between 1926 and 1945. In 1995, recruitment was extended to include women born between 1923 and 1950. Approximately 18% of the respondents joined the study spontaneously as a result of a passive recruitment campaign. A total of 17 035 women had joined the study at the end of recruitment in the autumn of 1996, resulting in a final response rate of 42.6%. At baseline, all participants received a health
Health-related lifestyle factors
From the baseline MDCS health questionnaire and a dietary assessment, we retrieved information on the following health-related lifestyle factors: smoking, alcohol use, physical activity, BMI, diet, self-rated health, and stress. Smoking was categorized into current (regular and occasional), former, and never. Information on alcohol intake was drawn from the questionnaire and a 7-day menu book (Wallstrom et al., 2000). We categorized alcohol consumption into no use in the past year, low (< 15 g/day), moderate (15–30 g/day), and high (> 30 g/day) (Mattisson et al., 2004). Leisure-time physical activity was measured using a modified questionnaire, adapted from the Minnesota Leisure Time Physical Activity Questionnaire (Taylor et al., 1978). Participants reported the number of minutes per week spent on each of 18 different activities during the four seasons. These numbers were multiplied by activity-specific intensity coefficients and added to create an overall score (Wallstrom et al., 2000). Individuals were ranked on the basis of their scores and divided into quartiles, which were dichotomized into low physical activity (Q1) and physically active (Q2–Q4) (Calling et al., 2006). We also included a variable on physical activity at work where respondents estimated the usual level of physical activity required for work (very low, low, medium, high, very high), categorized into low, medium, and high.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
46
European Journal of Cancer Prevention 2015, Vol 24 No 1
Height and weight were measured directly at baseline examination in light clothing without shoes and BMI was calculated in kg/m2. We analyzed the following WHOdefined BMI categories: underweight (< 18.5), normal (18.5–24.9), preobese (25.0–29.9), and obese (Z 30). Three different variables related to diet were considered. Information on participants being on a special diet (none, vegetarian or vegan, diabetic, or other) was collected from the baseline MDCS health questionnaire. The detailed dietary assessment (using a modified diet history methodology) included a 7-day menu book, a 168-item quantitative food frequency questionnaire, and a 1-h interview (Wirfalt et al., 2002). Information on dietary supplements was based on reported use (any/none) in the 7-day menu book. To assess overall diet, we used a diet quality index, which has been described in detail elsewhere (Drake et al., 2011). The diet quality index was based on adherence to the Swedish nutrition recommendations for the following components: saturated fatty acids (r 14% of nonalcohol energy), polyunsaturated fatty acids (5–10% of nonalcohol energy), fish and shellfish (Z 300 g/week), dietary fiber (2.4–3.6 g/MJ), fruit and vegetables (Z 400 g/day), and sucrose (r 10% of nonalcohol energy). Adherence to the recommended intake for each component was assigned 1 point, for a total score of 0–6. The index was further categorized into low quality (0–1 points), medium quality (2–3 points), and high quality (4–6 points). Self-rated physical and mental health was rated on a seven-point scale, dichotomized into poor health (1–4) and good health (5–7). Finally, participants were asked whether they had experienced recent stress or mental pressure outside of work (yes/no). Statistical analysis
We estimated odds ratios (ORs) and 95% confidence intervals (CIs) for mammography attendance from binary generalized estimating equation models where adjustments are made for the correlation of repeated measures within participants (with an autoregressive correlation structure). We first carried out age-adjusted analyses to determine the effects of the different health-related lifestyle factors on the odds of nonattendance. These effects were then further adjusted by baseline year, season, number of screening invitations after baseline, having been invited to screening before baseline, and sociodemographic factors, including education, employment status, cohabitation, and country of birth (multivariate model 1). The dietary index was also adjusted for total energy intake. In a final step, variables were further adjusted for all health-related variables (multivariate model 2). Interaction terms were tested in model 2 to examine possible effect modification by stress and selfreported health, and between menopausal status and BMI.
Results A total of 69 746 screening invitations were sent out to study participants, ranging between 1 and 12 per participant.
Overall, the attendance rate for the entire study period was 92% and for individual calendar years, the mammography attendance rate varied between 88 and 95% (data not shown). Descriptive sample characteristics are presented in Table 1. Women were on average 55 years old (range: 44–70) at baseline. The mean age at first screening opportunity was 56 years (range: 45–71). About one-third of women had completed high school, two-thirds were living with a partner, two-thirds were employed, and 88% were born in Sweden. In terms of occupational status (present or most recent position), 55% were nonmanual workers and 37% were manual workers. Relationship between risk factors and mammography nonattendance
Associations between health-related lifestyle factors and mammography nonattendance are presented in Table 2. Sociodemographic and screening characteristics of the study sample (n = 11 409), Malmo¨, Sweden (1992–2009)
Table 1
Sociodemographic variables Mean age at baseline (SD) Mean age at first subsequent screening invitation (SD) Age group at baseline (years) 44–49 50–54 55–59 60–64 65–70 Education level High school or higher Less than high school Missing Cohabitation (with partner) Yes No Missing Occupation (present or latest job) Self-employed/employer/farmer Higher nonmanual Middle nonmanual Lower nonmanual Skilled manual Unskilled manual Missing Employment status Employed Not employed Missing Country of birth Sweden Other Missing Invited to screening program before baseline Yes No Number of screening invitations after baseline 1 2 3 4 5 6 7 8 9 10 11 12
N (%) 54.9 (6.7) 56.7 (6.4) 3442 2470 2086 2155 1256
(30.2) (21.6) (18.3) (18.9) (11.0)
3717 (32.7) 7665 (67.3) 27 7871 (69.0) 3531 (31.0) 7 885 (7.8) 741 (6.6) 2049 (18.1) 3489 (30.9) 803 (7.1) 3341 (29.5) 101 7695 (67.6) 3681 (32.4) 33 10 066 (88.3) 1336 (11.7) 7 8459 (74.1) 2950 (25.9) 793 602 568 808 1271 1626 1896 1921 1247 609 67 1
(7.0) (5.3) (5.0) (7.1) (11.1) (14.3) (16.6) (16.8) (10.9) (5.3) (0.6) (0.0)
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Health factors and mammography attendance Lagerlund et al. 47
Table 2
Health-related lifestyle factors in relation to mammography screening nonattendance (Malmo¨, Sweden, 1992–2009) Odds ratios for screening nonattendance (95% CI)
Health-related lifestyle factors Smoking Current Former Never Missing Alcohol use None in past year Low (< 15 g/day) Medium (15–30 g/day) High (> 30 g/day) Physical activity level (leisure) Low (Q1) Active (Q2–Q4) Missing Physical activity at work Low Medium High Missing BMI Below normal (< 18.5) Normal (18.5–24.99) Preobese (25.0–29.99) Obese (Z 30) Missing Special diet None Vegetarian or vegan Diabetic or other Missing Dietary supplement use No Yes Dietary index score (0–6) Low (0–1) Medium (2–3) High (4–6) Self-rated health Poor (1–4) Good (5–7) Missing Stress (non-work-related) Yes No Missing
Age adjusted
Multivariate Model 1a
Multivariate Model 2b
3369 (29.5) 3280 (28.8) 4759 (41.7) 1
1.72 (1.57–1.89) 1.15 (1.04–1.27) Ref
1.65 (1.50–1.81) 1.14 (1.03–1.25) Ref
1.60 (1.45–1.76) 1.15 (1.05–1.28) Ref
738 8613 1767 291
(6.6) (75.5) (15.5) (2.6)
1.78 (1.53–2.06) Ref 1.03 (0.93–1.15) 1.31 (1.05–1.64)
1.63 (1.40–1.89) Ref 1.06 (0.96–1.18) 1.24 (0.99–1.57)
1.55 (1.32–1.83) Ref 1.06 (0.95–1.18) 1.14 (0.91–1.44)
2838 (25.0) 8516 (75.0) 55
1.18 (1.08–1.29) Ref
1.18 (1.08–1.29) Ref
1.10 (1.00–1.20) Ref
6551 (58.5) 3376 (30.1) 1275 (11.4) 207
Ref 1.05 (0.96–1.15) 1.35 (1.20–1.52)
Ref 1.03 (0.94–1.13) 1.24 (1.10–1.40)
Ref 1.00 (0.92–1.10) 1.13 (1.00–1.28)
148 (1.3) 5979 (52.5) 3733 (32.8) 1534 (13.5) 15
1.41 (1.00–2.00) Ref 0.89 (0.90–1.07) 1.20 (1.07–1.36)
1.37 (0.96–1.94) Ref 0.96 (0.88–1.05) 1.14 (1.01–1.29)
1.27 (0.89–1.81) Ref 0.97 (0.89–1.07) 1.10 (0.97–1.25)
10 771 (94.8) 171 (1.5) 423 (3.7) 44
Ref 1.83 (1.40–2.40) 1.34 (1.11–1.62)
Ref 1.40 (1.07–1.84) 1.18 (0.97–1.42)
Ref 1.49 (1.11–1.99) 1.12 (0.92–1.36)
2951 (25.9) 8458 (74.1)
1.08 (0.99–1.18) Ref
1.09 (1.00–1.19) Ref
1.11 (1.01–1.21) Ref
1783 (15.6) 6143 (53.8) 3483 (30.5)
1.18 (1.04–1.33) 1.12 (1.03–1.23) Ref
1.13 (1.00–1.28) 1.11 (1.01–1.21) Ref
1.02 (0.90–1.16) 1.05 (0.96–1.15) Ref
3352 (29.5) 8015 (70.5) 42
1.50 (1.39–1.63) Ref
1.37 (1.26–1.49) Ref
1.24 (1.14–1.36) Ref
3830 (33.7) 7548 (66.3) 31
1.49 (1.38–1.61) Ref
1.35 (1.24–1.46) Ref
1.25 (1.14–1.36) Ref
N (%) in study cohort
CI, confidence interval. a Adjusted for baseline year, season, screening sequence number after baseline, having been invited to screening before baseline, and sociodemographic factors (age, education, cohabitation, employment status, and country of birth). The dietary index score was further adjusted by the total energy intake. b Adjusted for all other included variables in addition to baseline year, season, invitation number, having been invited to screening before baseline, and sociodemographic factors (age, education, cohabitation, employment status, and country of birth).
When all variables were entered into model 2, many remained significantly associated with nonattendance, but effect sizes were generally reduced compared with the age-adjusted estimates and model 1 estimates. However, the effects of diet quality and BMI were no longer statistically significant. Nonattendance was more common among current smokers (OR = 1.60, 95% CI: 1.45–1.76), former smokers (OR = 1.15, 95% CI: 1.05–1.28), and those who had not consumed any alcohol in the past year (OR = 1.55, 95% CI: 1.32–1.83); however, high daily consumption of alcohol was not associated with nonattendance. Nonattendance was associated with low
leisure-time physical activity (OR = 1.10, 95% CI: 1.00–1.20) and high physical activity at work (OR = 1.13, 95% CI: 1.00–1.28). Furthermore, nonattendance was more common among vegetarians and vegans (OR = 1.49, 95% CI: 1.11–1.99), among those not taking dietary supplements (OR = 1.11, 95% CI: 1.01–1.21), among those with poor selfrated health (OR = 1.24, 95% CI: 1.14–1.36), and those experiencing higher levels of non-work-related stress (OR = 1.25, 95% CI: 1.14–1.36). Statistically significant interactions were found between stress and BMI (P < 0.01), and self-rated health and BMI (P < 0.01). Obesity was associated with nonattendance
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
48
European Journal of Cancer Prevention 2015, Vol 24 No 1
among women who were stressed (OR = 1.32, 95% CI: 1.09–1.60) and among women with poor self-rated health (OR = 1.39, 95% CI: 1.15–1.67). Among women who were stressed, nonattendance was also more common among those who were underweight (OR = 1.81, 95% CI: 1.09–2.99). No interaction was found between BMI and menopausal status (data not shown).
Discussion In this Swedish cohort of women who received regular invitations to attend mammography screening, a higher odds of nonattendance were observed among smokers, alcohol abstainers, vegetarians/vegans, women who did not use dietary supplements, women who had low leisuretime physical activity, physically demanding jobs, poor self-rated health, and those who were stressed. Underweight and obesity were found to be associated with nonattendance in some groups of women, whereas dietary quality was unrelated to attendance. Previous studies have found that health-related lifestyle factors including physical inactivity, obesity, alcohol use, stress, and possibly smoking are associated with breast cancer incidence and/or mortality, and therefore, these factors warrant extra attention. Similar to other studies, we found that nonattendance was associated with smoking (Rakowski et al., 1993; Beaulieu et al., 1996; Martin et al., 1996; Edwards and Boulet 1997; Maxwell et al., 1997b, 2001; Aro et al., 1999; Fredman et al., 1999; Lagerlund et al., 2000; Bancej et al., 2005; Jelinski et al., 2005; Lin, 2008; Gierisch et al., 2009; Caleffi et al., 2010; Lopez-de-Andres et al., 2010) and lower leisuretime physical activity (Fajardo et al., 1992; Rakowski et al., 1993; Maxwell et al., 1997b, 2001; Lin, 2008; Lopez-deAndres et al., 2010). Although there is some evidence that obesity may be associated with mammography attendance (Lopez-de-Andres et al., 2010), we, along with most other studies (Sutton et al., 1994; Staniscia et al., 2003; Lin, 2008; Caleffi et al., 2010), did not find an association between BMI and adherence to mammography screening when analyzing the entire sample. However, we did find statistically significant associations with BMI in subgroups of women who had poor self-rated health and who were stressed. Therefore, there is reason also to consider obesity when targeting groups whose screening attendance needs improvement. Contrary to some studies that found no association between the use of alcohol and mammography attendance (Maxwell et al., 1997b; Aro et al., 1999; Lopez-deAndres et al., 2010), but in corroboration with other studies (Rakowski et al., 1993; Sutton et al., 1994; Fredman et al., 1999; Lagerlund et al., 2000), we found that alcohol abstainers were less likely to attend mammography screening than those drinking some alcohol. Alcohol abstinence may be associated with previous alcohol abuse, poor health, or religious beliefs. Many studies
show a J-shaped curve between alcohol consumption and health outcomes, where both abstainers and heavy drinkers have worse health outcomes than light drinkers. The absence of an association between greater alcohol intake and mammography attendance in this study may be because of under-reporting of alcohol intake, which is a common problem in surveys (Feunekes et al., 1999). The reported absolute intakes of total alcohol were considerably lower than the reference method used to evaluate the relative validity of the MDCS method (Riboli et al., 1997). In addition, heavy drinkers may be less likely to participate in research studies, resulting in less variability on this measure. Hence, we do not dismiss the possibility that nonattendance could be associated with high alcohol consumption. To our knowledge, no study has specifically investigated how vegetarianism or veganism relates to mammography screening attendance. Our finding that vegetarians and vegans were less likely to attend screening may appear counterintuitive as these diets may indicate a certain level of health consciousness; however, the decision to follow a vegetarian or a vegan diet may reflect an individual’s ethical beliefs and is not necessarily health related. It is important to point out that the number of vegetarians and vegans in this study was small [n = 171 (6%)] and the data were based on self-reported information. An alternative explanation for higher nonattendance among vegetarians/vegans may be that these women feel a false sense of protection from disease in general or breast cancer in particular because of their health behaviors. However, we found no association between diet quality and mammography attendance. Although some studies have found associations with varying aspects of diet (Maclean et al., 1984; Fajardo et al., 1992; Flamant et al., 2006; Lopez-de-Andres et al., 2010), others have not (Sutton et al., 1994; Ore et al., 1997; Hagoel et al., 1999). We found that women with high stress and poor self-rated health were less likely to attend mammography screening. In the case of self-rated health, our finding contradicts a large body of research (Lin et al., 1990; Fajardo et al., 1992; Sutton et al., 1994; Rodriguez et al., 1995; Beaulieu et al., 1996; Rimer et al., 1996; Maxwell et al., 1997b; Ore et al., 1997; Michielutte et al., 1999; Lagerlund et al., 2000), although some studies have reported similar findings (Calnan 1984; Fredman et al., 1999). For both poor selfrated health and stress, lack of time and energy may explain the lower odds of attending mammography screening. As stress is associated with lower breast cancer survival (Chida et al., 2008), it is important to explore ways of encouraging women with higher levels of stress to lend higher priority to regular screening.
Strengths and limitations
The overall yearly attendance rate of 92% in this cohort is considerably higher than the rates reported in earlier
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Health factors and mammography attendance Lagerlund et al. 49
studies, which were closer to 65% (Matson et al., 2001; Zackrisson et al., 2004). This is likely because the MDCS cohort is selected toward better health, and possibly higher socioeconomic status and a healthier lifestyle than the general population (Manjer et al., 2001). Further, because of the prospective nature of this study and the long follow-up period (median follow-up: 15 years; range: 4–18), it is possible that women’s health behaviors may have changed between baseline and screening. However, this type of exposure misclassification would likely attenuate rather than inflate associations with the outcome. Despite these limitations, our study has many strengths, including its large size and prospective follow-up of screening attendance. Information on health-related lifestyle factors was collected independently of mammography screening among women who were free of breast cancer at baseline, which reduces the potential for recall bias. Further, although many studies rely on self-reported mammography attendance, we obtained this information from register data, which would minimize measurement error and misclassification. Conclusion
Our study of Swedish women who received regular invitations for mammography screening generally supports the notion that those with less healthy lifestyles are less likely to engage in mammography screening. In particular, nonattendance was associated with smoking, leisure-time physical inactivity, poor self-rated health, and stress. Also, women who were vegetarian or vegan were significantly less likely to attend screening; however, given the small size of this subgroup, this finding should be investigated more closely in future research. Dietary quality was unrelated to attendance, and underweight and obesity were associated with nonattendance only in some groups of women. Efforts to optimize mammography screening should consider methods of targeting screening referrals especially to women who smoke, are less physically active and obese, as these constitute risk factors for breast cancer.
Acknowledgements The authors are grateful to Dr Jonas Manjer for guidance on the database, Anders Dahlin for MDCS database management, and Darek Gozdzik for his work on the dataset that formed the basis for this study. This research was funded by the Swedish Cancer Society. Conflicts of interest
There are no conflicts of interest.
References Aro AR, de Koning HJ, Absetz P, Schreck M (1999). Psychosocial predictors of first attendance for organised mammography screening. J Med Screen 6:82–88. Bancej CM, Maxwell CJ, Onysko J, Eliasziw M (2005). Mammography utilization in Canadian women aged 50 to 69: identification of factors that predict initiation and adherence. Can J Public Health 96:364–368.
Beaulieu MD, Beland F, Roy D, Falardeau M, Hebert G (1996). Factors determining compliance with screening mammography. CMAJ 154:1335–1343. Caleffi M, Ribeiro RA, Bedin AJ Jr, Viegas-Butzke JM, Baldisserotto FD, Skonieski GP, et al. (2010). Adherence to a breast cancer screening program and its predictors in underserved women in southern Brazil. Cancer Epidemiol Biomarkers Prev 19:2673–2679. Calling S, Hedblad B, Engstrom G, Berglund G, Janzon L (2006). Effects of body fatness and physical activity on cardiovascular risk: risk prediction using the bioelectrical impedance method. Scand J Public Health 34:568–575. Calnan M (1984). The health belief model and participation in programmes for the early detection of breast cancer: a comparative analysis. Soc Sci Med 19:823–830. Chida Y, Hamer M, Wardle J, Steptoe A (2008). Do stress-related psychosocial factors contribute to cancer incidence and survival? Nat Clin Pract Oncol 5:466–475. Cogliano VJ, Baan R, Straif K, Grosse Y, Lauby-Secretan B, El Ghissassi F, et al. (2011). Preventable exposures associated with human cancers. J Natl Cancer Inst 103:1827–1839. Drake I, Gullberg B, Ericson U, Sonestedt E, Nilsson J, Wallstrom P, et al. (2011). Development of a diet quality index assessing adherence to the Swedish nutrition recommendations and dietary guidelines in the Malmo Diet and Cancer cohort. Public Health Nutr 14:835–845. Edwards N, Boulet J (1997). Implementing breast cancer screening guidelines: results of the Ontario Health Status Survey. Am J Prev Med 13:143–149. Fajardo LL, Saint-Germain M, Meakem TJ 3rd, Rose C, Hillman BJ (1992). Factors influencing women to undergo screening mammography. Radiology 184:59–63. Feunekes GI, van’t Veer P, van Staveren WA, Kok FJ (1999). Alcohol intake assessment: the sober facts. Am J Epidemiol 150:105–112. Flamant C, Gauthier E, Clavel-Chapelon F (2006). Determinants of noncompliance to recommendations on breast cancer screening among women participating in the French E3N cohort study. Eur J Cancer Prev 15:27–33. Fredman L, Sexton M, Cui Y, Althuis M, Wehren L, Hornbeck P, et al. (1999). Cigarette smoking, alcohol consumption, and screening mammography among women ages 50 and older. Prev Med 28:407–417. Gierisch JM, O’Neill SC, Rimer BK, DeFrank JT, Bowling JM, Skinner CS (2009). Factors associated with annual-interval mammography for women in their 40s. Cancer Epidemiol 33:72–78. Hagoel L, Ore L, Neter E, Shifroni G, Rennert G (1999). The gradient in mammography screening behavior: a lifestyle marker. Soc Sci Med 48:1281–1290. Jelinski SE, Maxwell CJ, Onysko J, Bancej CM (2005). The influence of breast self-examination on subsequent mammography participation. Am J Public Health 95:506–511. Lagerlund M, Sparen P, Thurfjell E, Ekbom A, Lambe M (2000). Predictors of nonattendance in a population-based mammography screening programme; socio-demographic factors and aspects of health behaviour. Eur J Cancer Prev 9:25–33. Lagerlund M, Sontrop JM, Zackrisson S (2013). Do reproductive and hormonal risk factors for breast cancer associate with attendance at mammography screening? Cancer Causes Control 24:1687–1694. Lin DY, Wei LJ, Zing Z (1990). Checking the Cox model with cumulative sums of martingale-based residuals. Biometrika 80:557–572. Lin SJ (2008). Factors influencing the uptake of screening services for breast and cervical cancer in Taiwan. J R Soc Promot Health 128:327–334. Lopez-de-Andres A, Martin-Lopez R, Hernandez-Barrera V, Carrasco-Garrido P, Gil-de-Miguel A, Esteban y Pena MM, et al. (2010). Predictors of breast and cervical cancer screening in a Spanish metropolitan area. J Womens Health (Larchmt) 19:1675–1681. Maclean U, Sinfield D, Klein S, Harnden B (1984). Women who decline breast screening. J Epidemiol Community Health 38:278–283. Manjer J, Carlsson S, Elmstahl S, Gullberg B, Janzon L, Lindstrom M, et al. (2001). The Malmo Diet and Cancer Study: representativity, cancer incidence and mortality in participants and non-participants. Eur J Cancer Prev 10:489–499. Manjer J, Elmstahl S, Janzon L, Berglund G (2002). Invitation to a populationbased cohort study: differences between subjects recruited using various strategies. Scand J Public Health 30:103–112. Martin LM, Calle EE, Wingo PA, Heath CW Jr (1996). Comparison of mammography and Pap test use from the 1987 and 1992 National Health Interview Surveys: are we closing the gaps? Am J Prev Med 12:82–90. Matson S, Andersson I, Berglund G, Janzon L, Manjer J (2001). Nonattendance in mammographic screening: a study of intraurban differences in Malmo, Sweden, 1990–1994. Cancer Detect Prev 25:132–137. Mattisson I, Wirfalt E, Wallstrom P, Gullberg B, Olsson H, Berglund G (2004). High fat and alcohol intakes are risk factors of postmenopausal breast cancer: a prospective study from the Malmo diet and cancer cohort. Int J Cancer 110:589–597.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
50
European Journal of Cancer Prevention 2015, Vol 24 No 1
Maxwell AE, Bastani R, Warda US (1997a). Breast cancer screening and related attitudes among Filipino-American women. Cancer Epidemiol Biomarkers Prev 6:719–726. Maxwell CJ, Kozak JF, Desjardins-Denault SD, Parboosingh J (1997b). Factors important in promoting mammography screening among Canadian women. Can J Public Health 88:346–350. Maxwell CJ, Bancej CM, Snider J (2001). Predictors of mammography use among Canadian women aged 50–69: findings from the 1996/97 National Population Health Survey. CMAJ 164:329–334. Michielutte R, Dignan MB, Smith BL (1999). Psychosocial factors associated with the use of breast cancer screening by women age 60 years or over. Health Educ Behav 26:625–647. Monninkhof EM, Elias SG, Vlems FA, van der Tweel I, Schuit AJ, Voskuil DW, et al. (2007). Physical activity and breast cancer: a systematic review. Epidemiology 18:137–157. Nystro¨m L, Andersson I, Bjurstam N, Frisell J, Nordenskjold B, Rutqvist L (2002). Long-term effects of mammography screening: updated overview of the Swedish randomised trials. Lancet 359:909–919. Olsson A, Garne JP, Tengrup I, Zackrisson S, Manjer J (2009a). Body mass index and breast cancer survival in relation to the introduction of mammographic screening. Eur J Surg Oncol 35:1261–1267. Olsson A, Garne JP, Tengrup I, Zackrisson S, Manjer J (2009b). Overweight in relation to tumour size and axillary lymph node involvement in postmenopausal breast cancer patients: differences between women invited to vs. not invited to mammography in a randomized screening trial. Cancer Epidemiol 33:9–15. Olsson S, Andersson I, Karlberg I, Bjurstam N, Frodis E, Hakansson S (2000). Implementation of service screening with mammography in Sweden: from pilot study to nationwide programme. J Med Screen 7:14–18. Ore L, Hagoel L, Shifroni G, Rennert G (1997). Compliance with mammography screening in Israeli women: the impact of a pre-scheduled appointment and of the letter-style. Isr J Med Sci 33:103–111. Rakowski W, Rimer BK, Bryant SA (1993). Integrating behavior and intention regarding mammography by respondents in the 1990 National Health Interview Survey of Health Promotion and Disease Prevention. Public Health Rep 108:605–624. Reeves GK, Pirie K, Beral V, Green J, Spencer E, Bull D (2007). Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study. BMJ 335:1134.
Riboli E, Elmstahl S, Saracci R, Gullberg B, Lindgarde F (1997). The Malmo Food Study: validity of two dietary assessment methods for measuring nutrient intake. Int J Epidemiol 26 (Suppl 1):S161–S173. Rimer BK, Conaway MR, Lyna PR, Rakowski W, Woods-Powell CT, Tessaro I, et al. (1996). Cancer screening practices among women in a community health center population. Am J Prev Med 12:351–357. Rodriguez C, Plasencia A, Schroeder DG (1995). Predictive factors of enrollment and adherence in a breast cancer screening program in Barcelona (Spain). Soc Sci Med 40:1155–1160. Staniscia T, Manzoli LM, Di Giovanni P, Dragani V, Testa P, Cavaliere D, et al. (2003). Factors related to the uptake of breast cancer screening (mammography and breast ultrasound): a retrospective survey on a sample of resident women, 50–70 years aged, from Abruzzo region. Ann Ig 15:1063–1075. Statistics Sweden (2014). Population by municipality, marital status and sex. Years 1968–2013. Stockholm, Sweden: Statistics Sweden. Sutton S, Bickler G, Sancho-Aldridge J, Saidi G (1994). Prospective study of predictors of attendance for breast screening in inner London. J Epidemiol Community Health 48:65–73. Swedish Association of Local Authorities and Regions (2013). Efforts for early detection. Even better cancer care – Interim report 4. Stockholm, Sweden: Swedish Association of Local Authorities and Regions. Tabar L, Yen MF, Vitak B, Chen HH, Smith RA, Duffy SW (2003). Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening. Lancet 361:1405–1410. Taylor HL, Jacobs DR Jr, Schucker B, Knudsen J, Leon AS, Debacker G (1978). A questionnaire for the assessment of leisure time physical activities. J Chronic Dis 31:741–755. Wallstrom P, Wirfalt E, Janzon L, Mattisson I, Elmstahl S, Johansson U, et al. (2000). Fruit and vegetable consumption in relation to risk factors for cancer: a report from the Malmo Diet and Cancer Study. Public Health Nutr 3:263–271. Wirfalt E, Mattisson I, Johansson U, Gullberg B, Wallstrom P, Berglund G (2002). A methodological report from the Malmo Diet and Cancer study: development and evaluation of altered routines in dietary data processing. Nutr J 1:3. World Cancer Research Fund/American Institute for Cancer Research (2010). Continuous Update Project Report summary. Food, nutrition, physical activity, and the prevention of breast cancer. Zackrisson S, Andersson I, Manjer J, Janzon L (2004). Non-attendance in breast cancer screening is associated with unfavourable socio-economic circumstances and advanced carcinoma. Int J Cancer 108:754–760.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
