Overall, among women diagnosed with breast cancer in Ontario between 2010 and 2017, those that were screened according to OBSP guidelines were less likely to be diagnosed with later stage disease. In particular, non-screeners had an almost 7-fold higher odds of being diagnosed with stage IV disease. However, as expected, these women were less likely to be diagnosed with DCIS (OR = 0.90, 95% CI 0.85, 0.96). In addition to this, we found that Ontario women residing in urban areas with lower neighborhood income had higher odds of being a non-screener.
Among women with breast cancer who were eligible to be screened in the OBSP (i.e., women ages 50–74 years), most individuals were adherent screeners (n = 33,821, 65.5%), while a minority did not screen at all, or were exclusively screened outside of OBSP (n = 17,796, 34.5%). This is consistent with previous research [14]. Notably, very few women (N = 333) engaged in non-adherent screening, (i.e., had a screening mammogram that did not follow OBSP guidelines). This suggests that individuals who get screened, tend to participate fully in the program, following OBSP guidelines. Focus on efforts to increase screening initiation in women living in larger urban areas with lower neighborhood income will increase screening rates among screen-eligible women, at least amongst those diagnosed with breast-cancer. Further, increasing screening rates is likely to reduce the number of late-stage cancers diagnosed, improving cancer outcomes.
Prior work from the OBSP has shown that about 83% of OBSP women who initiated screening returned for a subsequent screen, and that this proportion consistently increased from 1992–2001 [14]. However, more recent trends suggest that screening retention has actually decreased in Ontario from 83% in 2012 to 77% in 2018 [3]. In particular, it has been shown that the odds of returning for a second screen are highest for those living in rural compared to urban areas [15]. However, when compared to the broader literature, the impact of rural versus urban living on screening behaviour is mixed. Some studies based in the United States have found that access to breast screening is more available in urban centers. This is reflected in the higher screening rates for women living in urban versus rural areas [7, 16, 17]. Other studies, in Australia and Croatia (both with publicly funded screening), have shown similar screening rates in women living in rural and urban settings [18, 19]. The literature suggests that access to screening services does tend to be lower for women living in rural areas [18, 20]. On the contrary, in our study, we found that women living in rural areas were less likely to be non-screeners. This suggests that having an organized, province-wide, publicly funded, screening program mitigates some of the rural-urban disparities in screening rates observed in other jurisdictions.
Data has also shown that those in the lowest neighborhood income category tend to have higher odds of not being screened compared to those in the middle- or highest- neighborhood income categories [8, 21, 22]. Consistent with the evidence, in our study, lower- and middle-income individuals were more like to be non-screeners when compared to those in the highest income quintile. This is despite the fact that OBSP screening is publicly available, without the need for referral from primary care, and has no associated cost for the patient [23]. Prior research of low-income African American women in the United States has shown that mistrust of the medical system, inadequate education about screening, and the presence of barriers (e.g., lack of childcare and transportation) may limit the ability of some individuals to attend screening [24, 25]. Work is needed to determine if similar barriers to screening exist in Ontario, preventing low- and middle-income women from attaining the same degree of screening as their higher-income counterparts.
Notably, while we did observe differences in screening rates with income, this effect was limited to women living in an urban setting, with no observed differences in screening behavior based on income for individuals living in a rural area. Conversely, in urban areas, differences in screening behaviour for individuals with low, middle, and high income were observed. Here, the odds of being a non-screener were highest for individuals in the lowest income category (OR = 1.45 95% CI 1.32, 1.58 compared to high income individuals). This suggests that screening behaviour differences associated with the effects of income may only exist for those living in urban environments and not in rural areas. Similar research into the interaction of income and community size on breast screening has been performed in low-income countries [26]. Here, it was found that both urban and rural residing low-income women had significantly lower odds of mammography attendance, however the effect size for rural residing women was smaller in comparison to urban residing women [26]. Further research into breast cancer screening behaviour should focus on what specific barriers to screening exist for low-income, urban residing women that do not exist for low income rural residing women.
The effectiveness of breast cancer screening programs in reducing the incidence of advanced stage breast cancer has been shown in multiple studies [4, 5, 27–29]. Consistent with these findings, the results from this analysis show a clear gradient of an increasing odds of stage II (OR = 1.76, 95% CI 1.68, 1.85), III (OR = 2.74, 95% CI 2.56, 2.92), and IV (OR = 6.93, 95% CI 6.24, 7.69) cancers in non-screeners. These results suggest that regular breast cancer screening, within OBSP guidelines, is effective in reducing later stage cancer diagnoses. While performance measures of the OBSP have been analyzed [14, 30, 31], to our knowledge, this is the first analysis to look at differences in cancer stage for breast cancer patients who did or did not participate in the OBSP during this time-period. These results highlight the effectiveness of breast screening in Ontario in achieving the goal of reducing the incidence of later stage disease.
Rates of diagnosis of DCIS are known to be higher in a screening population [32]. Accordingly, in this study we found non-screeners to be less likely to be diagnosed with DCIS compared to adherent screeners (OR = 0.90, 95% CI 0.85, 0.96). It was expected that more cases of DCIS would be found in adherent screeners (15.5%) compared to non-screeners (9.9%).
This study has numerous strengths, including the use of a large population-based cohort of women diagnosed with breast cancer identified through a provincial cancer registry. This allowed for robust comparisons between adherent screeners and non-screeners. The OCR also includes detailed information on tumour characteristics (e.g., stage, ER-status) that can be linked to demographic characteristics of the women within this population.
Limitations of this study include a lack of information on personal income levels, race/ethnicity and immigration status of women diagnosed with breast cancer in Ontario. The absence of these key demographic variables limited our ability to examine their impact on breast cancer screening behaviours, and therefore limited the conclusions we could draw from this current analysis. Further, the study population only included women who were diagnosed with breast cancer, so when looking at demographic characteristics that associate with breast cancer screening behaviour, these associations may only exist amongst women diagnosed with breast cancer. However, screening behaviours were captured through OBSP records and only included screens conducted prior to diagnosis, and all women known to be at high risk of breast cancer (i.e., screened as part of the High Risk OBSP) were excluded from the analysis. Because the coverage of the OCR and the OBSP is province-wide, it is expected that the screening behaviours in this study population should reflect those of the broader population. Further evidence of this is seen in the similar screening rates (65.5%) in the study population, as compared to the general population of Ontario in 2018 (66.0%) [3]. Another limitation is that eligible women who exclusively screened outside OBSP would have been classified as non-screeners in this sample since only OBSP screening data was available. However, it would be expected that screening adherent women misclassified as non-screening would move the associations seen with cancer stage at diagnosis towards the null hypothesis. The fact that these results still show a strong association of non-screening and higher cancer stage at diagnosis suggests that this population of non-OBSP screeners was not sufficient to move the association towards the null hypothesis.
This research highlights important differences between women who are screened according to the guidelines of a province-wide, publicly funded screening program, and those who do not get screened at all. Women who followed these screening guidelines were less likely to have been diagnosed with late-stage breast cancer. Furthermore, being younger, urban residing, or having a lower neighbourhood income were all associated with a greater likelihood of not undergoing any breast cancer screening. Notably, few women were non-adherent screeners, suggesting that once women initiate screening, most tend to follow the guidelines set out by OBSP. This highlights the need for focused interventions aimed to increase screening initiation urban residing, low-income women, to increase screening rates and ensure that more breast cancers are detected before they progress to more advanced, and serious stages.