ARCHIVED - Organized Breast Cancer Screening Programs in Canada - Report on Program Performance in 2003 and 2004
Special Topic: Time Intervals between Screening Events
It is necessary for mammography to be repeated at regular intervals in order to maximize the benefits of participation in an organized breast cancer screening program (12). Yet the likelihood of returning to screening at a regular interval has been found to vary considerably across provincial programs (13). There are many factors that might influence the likelihood that a woman will make a timely return to screening, including demographic characteristics, the presence/ absence of breast cancer risk factors, previous screening experiences such as false-positive screening test results, and provincial/territorial program factors such as re-invitation and appointment reminder systems. There may be complex relationships among these factors, with one factor influencing another possibly negating or amplifying its influence on rescreening behaviour. In order to simultaneously examine these influences, a series of three longitudinal multivariate analyses were developed.
A total of 314,677 women aged 50 to 68 years who underwent a screening mammogram in British Columbia, Alberta, Manitoba, New Brunswick, and Newfoundland and Labrador in the years 2000 or 2001 were followed for at least 36 months, and then categorized into a series of groups for analysis (Figure 10). Women who were diagnosed with breast cancer at the initial screen were not included in these analyses. With each rescreening grouping (i.e. Group I – women who were early to rescreen; Group II - women who were late to rescreen; and, Group III – women who did not make a timely rescreen) designated as a dependant variable, and the potentially confounding factors, including demographic characteristics, breast cancer risk factor profiles, previous screening experiences, and the provincial programs, treated as independent variables, a series of multivariate logistic regression analyses were conducted. This statistical technique estimates the unique influence of each independent variable while controlling for the competing influence of the other potentially confounding independent variables. The results of these analyses are expressed as ‘odds ratios’, with value 1.0 representing equal odds and larger or smaller numbers representing greater or lesser likelihood. The percentages of program participants with a given dependant variable characteristic in each rescreening grouping were also calculated (see Table 9).
Figure 10 – Sampling framework for screening interval categories
Demographics
Demographic characteristics including age, country of birth, urban/rural residence, and educational attainment did not contribute very substantially to the likelihood of when or whether women returned for screening (see Table 9). The contribution of these characteristics.whether they were associated with greater or lesser likelihood -- was generally consistent, however, with some of the underlying risks often associated with demographic factors. For example, the increased risk of breast cancer associated with increasing age and Western lifestyles which may be more common among women born in Canada was associated with a modest increase in the odds of returning early (Odds Ratioadjusted (ORadj) 1.2 (95% Confidence Interval (CI): 1.17 - 1.23), p≤0.0001). Being aware of one's risk of developing breast cancer, and the potential benefits of health protective behaviours such as screening tend to be associated with increasing levels of education. Increasing levels of education were associated with a lower likelihood of not making a timely return to screening (ORadj 0.9 (CI: 0.87 - 0.92), p.0≤0001). Residing in a rural area is often associated with less convenient access to screening clinics. This can make timely rescreening more challenging, and this appears to be reflected by a slightly greater likelihood of not making a timely return to screening (ORadj 1.1 (CI: 1.07 - 1.12), p≤0.0001)).
Previous Screening Experiences
Of women who attended breast cancer screening for their first time in 2000 or 2001, 39% did not make a timely return compared to 19% of women who had screening mammography on at least one prior occasion at the time of the 2000 or 2001 index screen (ORadj 2.6 (CI: 2.57 - 2.68), p≤0.0001). This is consistent with previous reports (12) and underscores the importance of focusing attention on the retention of women undergoing first time screening (see Figure 11).
Table 9 - Factors associated with retention and screening intervals, including the odds (95% CI) of annual return, late return and failure to return (within 30 months)
Annual Return (<16 months) (n = 38,265) |
Late Return (30 to 36 months) (n = 11,958) |
Did not make a timely rescreen (within 30 months) (n = 75,522) |
||||
---|---|---|---|---|---|---|
% | ORadj (95% CI)Table 1 - Footnote 1 a |
% | ORadj (95% CI)Table 1 - Footnote 1 a |
% | ORadj (95% CI)Table 1 - Footnote 1 a |
|
Demographic Characteristics | ||||||
Age (Index Screen) | ||||||
50 to 54 years | 17.2 | 1.0 | 5.7 | 1.0 | 25.9 | 1.0 |
55 to 59 years | 15.8 | 0.9 (0.87 - 0.93)Table 1 - Footnote 6 *** |
5.1 | 1.0 (0.91 - 1.00)Table 1 - Footnote 4 * |
22.8 | Table 1 - Footnote 3 § |
60 to 64 years | 15.4 | 0.9 (0.89 - 0.94)Table 1 - Footnote 6 *** |
4.5 | 0.8 (0.80 . 0.88)Table 1 - Footnote 6 *** |
21.1 | 0.9 (0.86 - 0.90)Table 1 - Footnote 6 *** |
65 to 68 years | 14.8 | 0.9 (0.88 - 0.95)Table 1 - Footnote 6 *** |
3.9 | 0.7 (0.69 - 0.79)Table 1 - Footnote 6 *** |
21.6 | 0.9 (0.88 - 0.93)Table 1 - Footnote 6 *** |
Born In Canada | ||||||
no | 13.8 | 1.0 | 5.1 | 1.0 | 24.2 | 1.0 |
yes | 17.4 | 1.2 (1.17 - 1.23)Table 1 - Footnote 6 *** |
5.0 | 1.0 (1.00 - 1.09)Table 1 - Footnote 4 * |
23.0 | Table 1 - Footnote 3 § |
Residence | ||||||
urban | 16.0 | 1.0 | 5.0 | 1.0 | 23.2 | 1.0 |
rural | 16.4 | 1.1 (1.08 - 1.15)Table 1 - Footnote 6 *** |
5.0 | 1.2 (1.10 - 1.22)Table 1 - Footnote 6 *** |
24.9 | 1.1 (1.07 - 1.12)Table 1 - Footnote 6 *** |
Education | ||||||
< high school | 15.2 | 1.0 | 5.7 | 1.0 | 24.9 | 1.0 |
high school & some post | 16.7 | 1.2 (1.15 - 1.21)Table 1 - Footnote 6 *** |
4.6 | Table 1 - Footnote 3 § | 22.6 | 0.9 (0.91 - 0.94)Table 1 - Footnote 6 *** |
university degree | 16.3 | 1.2 (1.16 - 1.25)Table 1 - Footnote 6 *** |
4.8 | Table 1 - Footnote 3 § | 22.4 | 0.9 (0.87 - 0.92)Table 1 - Footnote 6 *** |
Breast Cancer Risk Factors | ||||||
Breast DensityTable 1 - Footnote 2 b | ||||||
low density | 15.1 | 1.0 | 5.1 | 1.0 | 23.7 | 1.0 |
high density | 20.4 | 1.3 (1.23 - 1.30)Table 1 - Footnote 6 *** |
5.0 | 0.9 (0.86 - 0.95)Table 1 - Footnote 6 *** |
22.5 | 0.9 (0.90 - 0.94)Table 1 - Footnote 6 *** |
Family History of Breast Cancer | ||||||
no | 14.2 | 1.0 | 5.1 | 1.0 | 23.9 | 1.0 |
yes | 30.1 | 2.5 (2.46 - 2.60)Table 1 - Footnote 6 *** |
4.3 | 0.8 (0.75 - 0.84)Table 1 - Footnote 6 *** |
20.3 | 0.8 (0.77 - 0.81)Table 1 - Footnote 6 *** |
Age at First Birth | ||||||
< 30 years | 16.1 | 1.0 | 5.0 | 1.0 | 23.6 | 1.0 |
≥ 30 years | 15.7 | Table 1 - Footnote 3 § | 5.0 | Table 1 - Footnote 3 § | 22.0 | 0.9 (0.87 - 0.93)Table 1 - Footnote 6 *** |
Parity | ||||||
≥ one live birth | 15.4 | 1.0 | 5.1 | 1.0 | 23.9 | 1.0 |
nulliparity | 18.5 | 1.2 (1.16 - 1.23)Table 1 - Footnote 6 *** |
4.9 | 0.9 (0.83 - 0.91)Table 1 - Footnote 6 *** |
21.9 | 0.9 (0.87 - 0.91)Table 1 - Footnote 6 *** |
Hormone Replacement Therapy (Index Screen) | ||||||
no current use | 14.3 | 1.0 | 5.1 | 1.0 | 24.3 | 1.0 |
current use | 19.4 | 1.4 (1.32 - 1.39)Table 1 - Footnote 6 *** |
5.0 | 0.9 (0.90 - 0.97)Table 1 - Footnote 5 ** |
21.9 | 0.9 (0.84 - 0.87)Table 1 - Footnote 6 *** |
Menopausal Status (Index Screen) | ||||||
post-menopausal | 16.0 | 1.0 | 5.0 | 1.0 | 22.9 | 1.0 |
pre-menopausal | 16.8 | 1.1 (1.03 - 1.10)Table 1 - Footnote 5 ** |
5.3 | 0.9 (0.90 - 0.99)Table 1 - Footnote 4 * |
26.0 | Table 1 - Footnote 3 § |
Past Screening Experiences | ||||||
Initial Screen (Index Screen) | ||||||
no | 16.3 | 1.0 | 4.5 | 1.0 | 19.2 | 1.0 |
yes | 15.0 | 0.8 (0.81 - 0.86)Table 1 - Footnote 6 *** |
7.8 | 1.8 (1.69 - 1.85)Table 1 - Footnote 6 *** |
39.4 | 2.6 (2.57 - 2.68)Table 1 - Footnote 6 *** |
Previous False-Positive Screen | ||||||
no | 15.5 | 1.0 | 4.9 | 1.0 | 22.5 | 1.0 |
yes | 26.7 | 1.8 (1.77 - 1.92)Table 1 - Footnote 6 *** |
6.9 | 1.3 (1.21 - 1.39)Table 1 - Footnote 6 *** |
36.0 | 1.8 (1.75 - 1.86)Table 1 - Footnote 6 *** |
Program-specific Factors | ||||||
Province | ||||||
British Columbia | 15.8 | 1.0 | 5.0 | 1.0 | 22.1 | 1.0 |
Alberta | 13.8 | 0.9 (0.84 - 0.92)Table 1 - Footnote 6 *** |
2.4 | 0.4 (0.37 - 0.45)Table 1 - Footnote 6 *** |
31.4 | 1.4 (1.32 - 1.40)Table 1 - Footnote 6 *** |
Manitoba | 11.1 | 0.8 (0.82 - 0.88)Table 1 - Footnote 6 *** |
4.9 | 0.8 (0.75 - 0.84)Table 1 - Footnote 6 *** |
22.5 | 0.8 (0.74 - 0.78)Table 1 - Footnote 6 *** |
New Brunswick | 19.3 | 1.2 (1.17 - 1.27)Table 1 - Footnote 6 *** |
7.7 | 1.6 (1.51 - 1.67)Table 1 - Footnote 6 *** |
27.3 | 1.2 (1.13 - 1.21)Table 1 - Footnote 6 *** |
Newfoundland and Labrador | 36.7 | 2.8 (2.64 - 2.90)Table 1 - Footnote 6 *** |
4.7 | 0.8 (0.68 - 0.83)Table 1 - Footnote 6 *** |
18.1 | 0.6 (0.54 - 0.60)Table 1 - Footnote 6 *** |
Figure 11 – Effect of Initial versus Subsequent Screen on Retention and Screening Interval
Among women who experienced a false-positive result in 2000 or 2001, 36% did not make a timely return compared to 23% of women who did not experience a false positive (ORadj:1.8 (CI (1.75 - 1.86)), p≤0.0001). This is also consistent with previous reports (14). It is unclear whether this tendency not to make a timely return is a result of a negative experience with screening, the transfer of women into non-programmatic sector for ongoing care, or delays in returning to regular mammography due to ongoing clinical follow-up. The way in which a typical two-year screening interval is counted before returning to regular mammography may also have an impact (i.e., two years after the completion of follow-up vs. two years from the date of the original screen). Given that false positives in this context can be associated with biopsy and benign breast disease -- both known breast cancer risk factors -- this tendency toward not making a timely return warrants further investigation (15). Women who experienced false-positive results and made timely returns to screening were considerably more likely to have returned for rescreening within 16 months than those women who had not experienced false-positive results (26.7% versus 15.5% respectively -- see Figure 12).
Breast Cancer Risk Factors
In general, women with breast cancer risk factors, including high breast density, current use of hormone replacement therapy (HRT), having a first child after age 30 or never having a baby (nulliparity) were more likely to make a return to screening within 30 months but the absolute differences between groups were small. As expected, women with a family history were considerably more likely to return early (within 16 months) for rescreening than women without a family history (ORadj:2.5 (CI (2.46 - 2.60)), p≤0.0001). Women with a family history who did not return early were slightly less likely to return late, or not make a timely return, than women without a family history (Figure 13).
Figure 12 – Impact of False Positive on Retention and Screening Interval
Program-specific factors
After controlling for variation in population demographics and breast cancer risk profile, there were still considerable differences among each of screening programs in terms of likelihood of women returning. In particular, women attending screening in Newfoundland and Labrador were most likely to return for early rescreening (ORadj:2.8 (CI (2.64 - 2.90)), p≤0.0001) and least likely to not return within 30 months (ORadj:0.6 (CI (0.54 - 0.60)), p≤0.0001). The interprovincial program difference suggest that program capacity, program-specific participant retention strategies, and the availability of mammography from other sources which might draw women outside of the organized program settings. all factors that are unique to individual programs/provinces.would be worthy focal points for retention enhancement efforts.
Summary
This report suggests that among women using organized breast cancer screening services in Canada, most attend on a biennial basis (Main Report: Figure 5, Tables 6-8). These more detailed analyses show that several characteristics may explain some variation in retention and the interval associated with timely retention and early rescreening. These characteristics include: initial versus subsequent screen, previous false positive results, family history of breast cancer, and provincial screening programs. An examination of the associations between these characteristics and screening outcomes represents a worthwhile avenue for future analysis. Variation between provincial programs offers an opportunity to study the effect of differing recall policies on breast cancer detection, morbidity, and mortality, ultimately leading to effective screening policy across Canada.
Figure 13 - Impact of Family History of Breast Cancer on Retention and Screening Interval
Page details
- Date modified: