DIABETICMedicine DOI: 10.1111/dme.12422
Research: Epidemiology Impact of socio-economic status on breast cancer screening in women with diabetes: a population-based study W. Chan1, L. Yun2, P. C. Austin2,3, R. L. Jaakkimainen2,4,5, G. L. Booth2,3,6,7, J. Hux2,3,6, P. A. Rochon1,2,3,6 and L. L. Lipscombe1,2,3,6 1 Women’s College Research Institute, Women’s College Hospital, 2Institute for Clinical Evaluative Sciences, 3Institute of Health Policy, Management and Evaluation, University of Toronto, 4Department of Family and Community Medicine, University of Toronto, 5Sunnybrook Health Sciences Centre, 6Department of Medicine, University of Toronto and 7St Michael’s Hospital, Toronto, ON, Canada
Accepted 19 February 2014
Abstract Aims There is evidence to suggest that mammography rates are decreased in women with diabetes and in women of lower socio-economic status. Given the strong association between low socio-economic status and diabetes, we explored the extent to which differences in socio-economic status explain lower mammography rates in women with diabetes. Methods A population-based retrospective cohort study in Ontario, Canada, of women aged 50 to 69 years with diabetes between 1999 and 2010 age matched 1:2 to women without diabetes. Main outcome measure is the likelihood of at least one screening mammogram in women with diabetes within a 36-month period, starting as of either 1 January 1999, their 50th birthday, or 2 years after diabetes diagnosis – whichever came last. Outcomes were compared with those in women without diabetes during the same period as their matched counterparts, adjusting for socio-economic status based on neighbourhood income and other demographic and clinical variables.
Of 504 288 women studied (188 759 with diabetes, 315 529 with no diabetes), 63.8% had a screening mammogram. Women with diabetes were significantly less likely to have a mammogram after adjustment for socio-economic status and other factors (odds ratio 0.79, 95% CI 0.78–0.80). Diabetes was associated with lower mammogram use even in women from the highest socio-economic status quintile (odds ratio 0.79, 95% CI 0.75–0.83).
Results
Conclusions The presence of diabetes was an independent barrier to breast cancer screening, which was not explained by differences in socio-economic status. Interventions that target patient, provider, and health system factors are needed to improve cancer screening in this population.
Diabet. Med. 31, 806–812 (2014)
Introduction There is mounting evidence of an association between diabetes and cancer outcomes. Women with diabetes have been shown to have a higher incidence of post-menopausal breast cancer [1], as well as an increased mortality following a cancer diagnosis [2]. In addition to metabolic factors that influence tumour growth and progression [3], lower breast cancer screening in women with diabetes may also contribute to higher cancer mortality. Screening and early detection through mammography in post-menopausal women has been Correspondence to: Lorraine L. Lipscombe. E-mail: [email protected]
806
associated with increased breast cancer survival [4]. Although a mammogram is recommended every 3 years for women between 50 and 74 years of age [5], a significant proportion of eligible women do not receive routine screening [6]. There is evidence that cancer screening rates are reduced in patients with chronic diseases [7,8] and some studies have documented significantly lower mammography rates in women with diabetes [9–12]. Both healthcare provider and patient factors are known to influence the provision of adequate preventive care and thus may influence screening rates [13]. Indeed, a growing body of literature suggests that the competing demands of complex, chronic diseases such as diabetes interfere with attention to other clinical preventive
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
What’s new? • This study is the first to explore the influence of socio-economic status on the gap in screening mammograms among women with diabetes. • While there is evidence that the presence of diabetes is associated with a decrease in breast cancer screening, no study has examined the extent to which the higher prevalence of low socio-economic status in women with diabetes contributes to this disparity. • Using population-based data, we showed that diabetes is an independent barrier to adequate breast cancer screening even after adjusting for socio-economic status, and that low socio-economic status serves as an additional obstacle to regular screening mammograms. services [13,14]. Patient-related socio-demographic factors that are more common among patients with diabetes, such as recent immigration [15] and low socioeconomic status [16,17], may also contribute to cancer screening disparities. Socio-economic status is a well-known predictor of healthcare utilization, and socio-economic status-based inequities in access to care exist in Canada, despite the fact that most health care is universally funded [18,19]. Many studies have specifically shown lower cancer screening rates among low socio-economic status populations in Canada and elsewhere [20,21]. The extent to which lower socio-economic status among patients with diabetes contributes to disparities in cancer screening is unclear. A better understanding of the patientrelated factors that contribute to lower screening in women with diabetes will help to guide interventions aimed at improving cancer screening and detection in this population. This issue is particularly important in healthcare settings such as Canada, where breast cancer screening is universally subsidized by provincial health plans either via physician referrals or self-referral in Ontario through the Ontario Breast Cancer Screening Program. The objective of this population-based study was to compare receipt of a screening mammogram in Ontario between screen-eligible women with diabetes and age-matched women without diabetes, and to determine the influence of socio-economic status on this relationship.
Subjects and methods Study design and population
We conducted a retrospective, population-based cohort study of women aged 50–69 years with and without diabetes who were living in the province of Ontario, Canada from 1 January 1999 to 31 December 2010. This age group was chosen to reflect the Canadian breast cancer screening
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
DIABETICMedicine
guidelines during that time, which recommended a bilateral mammogram every 24 months in all women between 50 and 69 years of age [5]. We only included women who had at least one primary care visit in the year prior to cohort entry to ensure that all subjects had a similar opportunity to be referred for mammography screening during the follow-up period. These visits may have included both ‘walk-in’ and scheduled appointments, which cannot be distinguished within our data sets. We excluded women with a history of breast cancer, those living in a long-term care facility, those diagnosed with diabetes less than 2 years before the end of their maximum screening interval, and/or those who had less than 36 months of follow-up (because of death or a move from the province).
Data sources
We used linked, population-based healthcare databases, which record data on all Ontario residents covered by the universal provincial healthcare plan. We obtained information on demographics and deaths from the Registered Persons Database. Mammogram records were obtained from the Ontario Health Insurance Plan database, which provides information on physician service and radiologic procedure claims, and the Ontario Breast Cancer Screening Program database, which contains data on screening mammograms for women enrolled in this programme [22]. The Ontario Breast Cancer Screening Program database is a province-wide self-referral breast cancer-screening programme for screen-eligible women that provides routine mammograms, follow-up, and referral services for abnormal tests, and reminders for subsequent screening intervals [22]. We identified co-morbid diagnoses during hospital admissions from the Canadian Institute for Health Information and Discharge Abstract Database and breast cancer cases from the Ontario Cancer Registry [23]. We determined diabetes status from the Ontario Diabetes Database, a previously validated registry of Ontarians diagnosed with diabetes [24]. We linked administrative healthcare databases anonymously using encrypted health card numbers.
Procedures
We first identified all women aged 50–67 years between 1 January 1999 and 31 December 2007 who met inclusion criteria. We then selected all women with diabetes based on inclusion in the Ontario Diabetes Database during that time period. We assigned each subject an index date to begin follow-up, which was either 1 January 1999, their 50th birthday, or 2 years after a diabetes diagnosis, whichever came last (i.e. follow-up began once all three criteria were met). We began following women once they had diabetes for at least 2 years (i.e. prevalent diabetes) to avoid capturing mammograms that were ordered before diabetes diagnosis among newly diagnosed patients. We
807
DIABETICMedicine
Impact of socio-economic status on mammograms in women with diabetes W. Chan et al.
derived a cohort of eligible women without diabetes from the Ontario population, age matched 2:1 to women with diabetes and assigned them the same index date as their counterpart with diabetes.
Breast cancer screening and follow-up
The primary outcome was at least one screening mammogram within a 36-month period after the index date. While the recommended screening interval during our study era was 24 months, we chose 36 months to reflect the updated maximum screening interval in the 2011 Canadian screening guidelines [5]. We defined screening mammograms as either a bilateral mammogram recorded in the Ontario Health Insurance Plan database procedure claims database or a mammogram performed through the Ontario Breast Cancer Screening Program database.
Socio-economic status
We defined socio-economic status based on each individual’s neighbourhood income quintile recorded in the Registered Persons Database. We categorized individuals into neighbourhood income quintiles by linking their postal code data with Canadian census data on median household income levels by neighbourhood of residence. This method has been widely used to estimate socio-economic status [16,19] and has been shown to correlate well with health outcomes [19]. We excluded persons whose postal code could not be linked.
and co-morbidity. As a secondary analysis, we tested for an interaction between diabetes and socio-economic status quintile to assess the effect of socio-economic status quintile within the groups with diabetes and without diabetes, and the effect of diabetes within each socio-economic status quintile.
Ethics
This study was approved by the Research Ethics Board of Sunnybrook Health Sciences Centre in Toronto.
Results A total of 188 759 women with diabetes met the inclusion criteria and they were matched to 315 529 eligible women without diabetes. The index date was 1 January 1999 in 22.4%, as of age 50 years in 21.0%, and 2 years after diabetes diagnosis in 62.6% of women. Baseline characteristics for the groups with diabetes and without diabetes are presented in Table 1. The mean age at index date was 57.2 5.6 years. Women with diabetes were more likely to live in a low socio-economic status neighbourhood, to have immigrated to the province within the last 10 years, and to have a higher weighted co-morbidity score than those without diabetes (Table 1). Table 1 Baseline characteristics—diabetes compared with no diabetes among those who had a primary care visit within previous year Characteristic
Other covariates
We recorded the following baseline characteristics: age, rural residence defined as a dissemination area with less than 10 000 residents, recent immigration into Ontario based on entry into the province less than 10 years before index date (as determined by new enrolment in the provincial healthcare programme), and co-morbidity based on diagnoses recorded on physician claims in the 2 years prior to cohort entry. We derived the weighted co-morbidity score from the Johns Hopkins Adjusted Clinical Group Case-Mix assignment software, which assigns diagnoses into 32 Aggregated Diagnosis Groups. This method has been validated to predict mortality in the general adult population [25].
Statistical analyses
We used descriptive statistics (v2-tests) to compare baseline variables between women with and without diabetes. We used univariate and multivariable conditional logistic regression to compare the likelihood of a screening mammogram between women with and without diabetes. For our primary analysis, we assessed the association between diabetes and mammogram use, adjusting for age, socio-economic status quintile, rural residence, recent immigration into Ontario
808
Mean age at index date (years) Socio-economic status quintile*, n (%) Socio-economic status 1 (lowest) Socio-economic status 2 Socio-economic status 3 Socio-economic status 4 Socio-economic status 5 (highest) Rural residence†, n (%) Recent immigrant (resident < 10 years)‡, n (%) Weighted co-morbidity score > 10§, n (%)
Diabetes
No diabetes
n = 188 759 57.1 5.6
n = 315 529 57.2 5.6
46 283 (24.5%)
56 590 (17.9%)
42 097 (22.3%)
62 912 (19.9%)
39 414 (20.9%)
64 866 (20.6%)
32 775 (17.4%)
63 203 (20.0%)
28 190 (14.9%)
67 958 (21.5%)
24 283 (12.9%) 21 365 (11.3%)
43 447 (13.8%) 26 229 (8.3%)
31 686 (16.8%)
30 968 (9.8%)
*Based on neigbhourhood median household income derived from census data and postal codes. † Rural residence defined as a dissemination area with less than 10 000 residents. ‡ Recent immigrant defined as entry into province less than 10 years prior to index date. § Based on physician claims in the previous 2 years listing > 10 diagnostic groups from Johns Hopkins Aggregated Diagnosis Groups (ADG).
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
Screening mammograms
During the 36-month follow-up period, a screening mammogram was recorded in 321 564 (63.8%) women from our cohort: 113 873 (60.3%) in those with diabetes and 207 691 (65.8%) in women without diabetes (P < 0.001). Among those who received mammograms, 61.1% were screened through physician referrals (Ontario Health Insurance Plan database) and 38.9% were screened through the Ontario Breast Cancer Screening Program. Women with diabetes were slightly more likely to be tested through the Ontario Breast Cancer Screening Program database than those without diabetes (39.8% vs. 38.4% of tests), but the rate of mammograms remained significantly lower for women with diabetes regardless of the service used (diabetes vs. no diabetes: 36.3% vs. 40.6% for Ontario Health Insurance Plan database, P < 0.001; and 24.0% vs. 25.3% for Ontario Breast Cancer Screening Program database, P < 0.001). The majority of women were followed for 36 months (97.9% with diabetes, 94.2% without diabetes). Among those women whose follow-up was terminated earlier, 0.8% were diagnosed with breast cancer and 0.5% lost Ontario Health Insurance Plan database eligibility in each group, 0.8% with diabetes and 0.2% without diabetes entered a long-term care facility, and 4.4% of women without diabetes were subsequently diagnosed with diabetes. The rate of screening mammograms increased with increasing socio-economic status quintile, both overall and within women with and without diabetes (Fig 1). As illustrated in Fig 1, the proportion of women who had a mammogram was lower in women with diabetes across all socio-economic status groups, with the lowest rate in women with diabetes from the lowest socio-economic status quintile
FIGURE 1 Rate of screening mammograms among screen-eligible women within 36 months with increasing socio-economic status quintile, both overall and within women with ( ) and without ( ) diabetes.
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
DIABETICMedicine
(55.9%) and the highest rate in women without diabetes in the highest socio-economic status quintile (71.8%). Women with diabetes were significantly less likely to have a screening mammogram in the subsequent 36-month period compared with women without diabetes (unadjusted odds ratio 0.79, 95% CI 0.78–0.80). This relationship persisted after adjustment for age, socio-economic status quintile, recent immigration, rural residence, and co-morbidity score (adjusted odds ratio 0.86, 95% CI 0.85–0.87) (Table 2). As shown in Table 2, lower socio-economic status was also significantly associated with lower mammogram use on multivariable analysis (odds ratio 0.61, 95% CI 0.60–0.63 for socio-economic status quintile 1 vs. 5; odds ratio 0.88, 95% CI 0.86–0.90 for socio-economic status quintile 4 vs. 5). While socio-economic status was associated with a significant decrease in mammograms in women with and without diabetes, a significant two-way interaction was found between diabetes and socio-economic status 1, 2, and 3 versus 4 and 5 (P < 0.0001). The influence of socio-economic status was relatively more pronounced in women without diabetes (socio-economic status 1 vs. 5: odds ratio 0.67, 95% CI 0.63–0.71 in patients with diabetes; odds ratio 0.58, 95% CI 0.55–0.61 in patients without diabetes). In addition, the influence of diabetes was stronger in the highest income quintile compared with the lowest quintile (diabetes vs. no diabetes: odds ratio 0.79, 95% CI 0.75–0.83 in socio-economic status 5; odds ratio 0.91, 95% CI 0.88– 0.94 in socio-economic status 1) (Table 2).
Discussion This large, population-based study found that women with diabetes were 14% less likely to receive a mammogram during the recommended screening interval compared with those without diabetes. This disparity was not explained by differences in socio-economic status, immigration status, region of residence, or co-morbidity. Lower socio-economic status was also an independent predictor of decreased mammogram use in both women with and without diabetes, with the lowest screening rate found in women with both low socio-economic status and diabetes. Diabetes remained a significant barrier to screening mammogram tests even among women from the highest socio-economic status quintile. Indeed, the effect of diabetes was relatively stronger in women of higher socio-economic status, and the effect of low socio-economic status was more salient among women without diabetes. These findings suggest that while both diabetes and socio-economic status independently affect breast cancer screening, the relative contribution of each factor to screening disparities may be greater in populations without competing risk factors. This study illustrates the complex relationship between patient and disease-specific factors and healthcare utilization. Diabetes represents an important barrier to adequate cancer screening, which is not attributable to socio-demographic factors. Rather, low
809
DIABETICMedicine
Impact of socio-economic status on mammograms in women with diabetes W. Chan et al.
Table 2 Number and proportion of women who had a screening mammogram, and results of logistic regression analysis of associations between diabetes, socio-economic status, and mammogram use Total (n)
Per cent screened
Overall Diabetes Yes 188 759 60.3% No 315 529 65.8% Socio-economic status quintile Socio-economic status 1 (lowest) 102 873 57.3% Socio-economic status 2 105 009 61.5% Socio-economic status 3 104 280 64.5% Socio-economic status 4 95 978 67.4% Socio-economic status 5 (referent) 96 148 70.2% Effect of socio-economic status in women with diabetes Socio-economic status quintile Socio-economic status 1 (lowest) 46 283 55.9% Socio-economic status 2 42 097 59.4% Socio-economic status 3 36 824 61.9% Socio-economic status 4 32 775 63.9% Socio-economic status 5 (referent) 28 190 66.2% Effect of socio-economic status in women without diabetes Socio-economic status quintile Socio-economic status 1 (lowest) 56 590 58.5% Socio-economic status 2 62 912 63.0% Socio-economic status 3 63 223 66.1% Socio-economic status 4 63 203 69.2% Socio-economic status 5 (referent) 67 958 71.8% Effect of diabetes within socio-economic status groups Socio-economic status quintile 1 (lowest) Diabetes 46 283 55.9% No diabetes 56 590 58.5% Socio-economic status quintile 2 Diabetes 42 097 59.4% No diabetes 62 912 63.0% Socio-economic status quintile 3 Diabetes 36 824 61.9% No diabetes 63 223 66.1% Socio-economic status quintile 4 Diabetes 32 775 63.9% No diabetes 63 203 69.2% Socio-economic status quintile 5 (highest) Diabetes 28 190 66.2% No diabetes 67 958 71.8%
Unadjusted odds ratio (95% CI)
Adjusted odds ratio* (95% CI)
0.79 (0.78–0.80) 1.00
0.86 (0.85–0.87) 1.00
0.56 0.67 0.72 0.86 1.00
(0.54–0.57) (0.65–0.69) (0.70–0.74) (0.84–0.88)
0.61 0.71 0.75 0.88 1.00
(0.60–0.63) (0.69–0.73) (0.74–0.77) (0.86–0.90)
0.64 0.74 0.75 0.89 1.00
(0.60–0.68) (0.70–0.78) (0.71–0.79) (0.85–0.93)
0.67 0.75 0.77 0.90 1.00
(0.63–0.71) (0.72–0.79) (0.73–0.81) (0.86–0.94)
0.55 0.67 0.74 0.87 1.00
(0.52–0.58) (0.65–0.69) (0.65–0.69) (0.85–0.89)
0.58 0.69 0.75 0.88 1.00
(0.55–0.61) (0.67–0.71) (0.73–0.77) (0.86–0.90)
0.89 (0.86–0.92) 1.00
0.91 (0.88–0.94) 1.00
0.85 (0.82–0.88) 1.00
0.87 (0.83–0.90) 1.00
0.78 (0.75–0.82) 1.00
0.81 (0.78–0.84) 1.00
0.79 (0.75–0.82) 1.00
0.81 (0.77–0.84) 1.00
0.77 (0.73–0.81) 1.00
0.79 (0.75–0.83) 1.00
*Adjusted for socio-economic status, co-morbidity, rural residence, immigration status.
socio-economic status exerts an additive influence on the quality of care for this already vulnerable population. Our findings support those of other studies showing lower preventive care in patients with chronic diseases [7,8,26]. However, evidence regarding breast cancer screening among women with diabetes has been mixed [9–12,27,28]. In two studies examining USA survey data from the general population, diabetes was not associated with lower mammogram use [27,28]. It is possible that the greater healthcare access among US women with diabetes compared with the general population may have led to more opportunities for cancer screening in those studies, thus offsetting any differences between groups. In contrast, diabetes was associated with a significant reduction in mammograms when populations were limited to women with similar healthcare access, as in this study and in others [9–12,26]. For example, one study
810
from Spain (where eligible women are systematically invited for mammograms) found that women with diabetes were 16% less likely to report receiving a mammogram than their counterparts without diabetes (37.9% vs. 53.8%) [26]. The overall screening rates were lower than in the current study, possibly because mammography was based on self-report rather than claims. Our earlier study among Ontario women between 1999 and 2002 also had lower mammogram rates than this study (38.1% and 47.3% in women with and without diabetes, respectively), and found a 30% lower likelihood of a mammogram in women with diabetes compared to control subjects without [9]. However, that study was missing data on approximately 38% of mammograms performed through the Ontario Breast Cancer Screening Program database, which might have been used more frequently by women with diabetes. Nonetheless, our present
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
findings are consistent with our earlier evidence of lower mammogram use among women with diabetes, even after accounting for tests performed through the Ontario Breast Cancer Screening Program database and controlling for factors such as socio-economic status, immigration status, and co-morbidity. This study specifically explored whether low socio-economic status explained the relationship between diabetes and breast cancer screening, as this factor is a well-known predictor of both diabetes [16,17] and cancer screening [20,21]. Indeed, we found that populations of lower socio-economic status were significantly less likely to undergo a mammogram during the recommended screening interval. This finding supports other studies that have shown substantial socio-economic disparities in cancer screening practices in Canada and European countries despite universal health coverage [12,20,26]. Socio-economically disadvantaged groups may have lower preventive healthcare utilization as a result of lower health literacy, less advocacy, discrimination, and impaired access, for example [21]. Although women with diabetes were more likely to have a lower socio-economic status in our study, we found that the disparities in mammogram use persisted after adjusting for socio-economic status. Moreover, diabetes remained a significant barrier to breast cancer screening even in higher socio-economic status populations. We note that the gap in screening because of diabetes has narrowed in recent years compared with our previous study [9]. This improvement may be attributable in part to differences in health system factors, such as the introduction of a screening incentive programme for physicians [29] and/or the present study’s inclusion of tests performed through the Ontario Breast Cancer Screening Program database. Office based diabetes management has become significantly more complex and time-consuming as a result of a shift toward more intensive therapeutic interventions [30]. Given these increasing demands on physicians and greater time constraints because of larger patient loads, preventive issues such as cancer screening may be more often overlooked. Programmes that offer incentives and reminders for cancer screening or allow for self-referral may lead to greater adherence, particularly for patients with multiple healthcare needs. Indeed, the difference in mammograms between the groups with and without diabetes was much smaller among women screened through the Ontario Breast Cancer Screening Program database. It is possible that such organized screening programmes may alleviate some of the competing demands faced by providers caring for complex patients with diabetes. Patients with diabetes also often see multiple physicians, including specialists who are less familiar with guidelines outside of their field. This fragmentation of care might hamper the effective use of healthcare services. Interestingly, one study in patients with diabetes within a US managed care setting found that primary care physicians with better diabetes-related processes of care were also more likely to perform mammo-
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
DIABETICMedicine
grams [31]. These findings highlight the role of both health system and provider factors and suggest that better diabetes care may correlate rather than compete with cancer screening. Interventions to improve cancer screening in patients with diabetes should therefore be multifaceted, focusing on more standardized health system measures, increased support and incentives for diabetes care providers, and greater education for more socially disadvantaged populations. Strengths of our study include the use of population-based data for a large and diverse jurisdiction, and a validated algorithm to identify diabetes. Some limitations should be noted. As with all observational studies, we cannot exclude the role of selection bias. While we were able to adjust for certain patient factors that may influence screening practices, we lacked data on other factors, including family history of breast cancer, lifestyle, ethnicity, education, obesity, and attitudes toward screening. Moreover, in cases where mammograms were not obtained, we cannot determine whether the physician failed to refer the patient or the patient chose not to obtain the recommended test. Finally, in the absence of individual-level income data, we based our measure of socio-economic status on neighbourhood-level household income, which may be less accurate for more mixed neighbourhoods, or for certain groups such as retired persons or live-in domestic workers. Nonetheless, neighbourhood income is widely used as a measure of socio-economic status that has been shown to correlate well with individual level measures and health outcomes [16,18,19]. In conclusion, findings from this large, population-based study suggest that women with diabetes are a unique population at risk for inadequate breast cancer screening, and that low socio-economic status is an additional obstacle to preventive care in an already disadvantaged population. These findings were not explained by differences in co-morbidity, immigration status, or socio-economic status in women with diabetes. These results are of particular importance as women with diabetes are at higher risk of breast cancer and of poorer survival once diagnosed. Interventions to improve cancer screening that target patient, provider, and health system factors need to be developed to ensure ongoing provision of comprehensive care for patients with diabetes. Previous presentation
An abstract of this work was presented at the 15th Annual Canadian Diabetes Association Conference (10–13 October 2012) in Vancouver, British Columbia.
Funding sources
LLL is supported by a Canadian Institutes of Health Research (CIHR) New Investigator Award. PCA is supported in part by a Career Investigator award from the Heart and Stroke Foundation. This study was conducted with the support from a CIHR operating grant (MOP no. 123263),
811
DIABETICMedicine
Impact of socio-economic status on mammograms in women with diabetes W. Chan et al.
and from the Ontario Institute for Cancer Research and Cancer Care Ontario. This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred. The funding bodies had no involvement in the study design, data collection, data analysis, manuscript preparation, and/or publication decisions.
13
14
15
16
17
Competing interests
None declared.
18
Acknowledgements
19
The authors would like to thank Hadas Fischer MD MSc, Institute for Clinical Evaluative Sciences, for assistance with design and implementation of the study, and Lauren McNicol MA, Women’s College Research Institute, for help with manuscript preparation.
20
References 1 Boyle P, Boniol M, Koechlin A, Robertson C, Valentini F, Coppens K et al. Diabetes and breast cancer risk: a meta-analysis. Br J Cancer 2012; 107: 1608–1617. 2 Peairs KS, Barone BB, Snyder CF, Yeh HC, Stein KB, Derr RL et al. Diabetes mellitus and breast cancer outcomes: a systematic review and meta-analysis. J Clin Oncol 2011; 29: 40–46. 3 Godsland IF. Insulin resistance and hyperinsulinaemia in the development and progression of cancer. Clin Sci 2010; 118: 315–332. 4 Tabar L, Yen MF, Vitak B, Chen HH, Smith RA, Duffy SW. Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening. Lancet 2003; 361: 1405–1410. 5 Canadian Task Force on Preventive Health Care. Screening for Breast Cancer: Recommendations for Clinicians and Policymakers. Ottawa: Health Canada, 2011. 6 Cancer Quality Council of Ontario. Cancer System Quality Index 2012: Breast Cancer Screening (Mammography) Participation. Ontario: Cancer Quality Council of Ontario, 2012. 7 Fontana SA, Baumann LC, Helberg C, Love RR. The delivery of preventive services in primary care practices according to chronic disease status. Am J Public Health 1997; 87: 1190–1196. 8 Coughlin SS, Uhler RJ, Hall HI, Briss PA. Non-adherence to breast and cervical cancer screening: what are the linkages to chronic disease risk? Prev Chronic Dis 2004; 1: A04. 9 Lipscombe LL, Hux JE, Booth GL. Reduced screening mammography among women with diabetes. Arch Intern Med 2005; 165: 2090–2095. 10 McBean AM, Yu X. The underuse of screening services among elderly women with diabetes. Diabetes Care 2007; 30: 1466–1472. 11 Beckman TJ, Cuddihy RM, Scheitel SM, Naessens JM, Killian JM, Pankratz VS. Screening mammogram utilization in women with diabetes. Diabetes Care 2001; 24: 2049–2053. 12 Jimenez-Garcia R, Hernandez-Barrera V, Carrasco-Garrido P, Gil A. Prevalence and predictors of breast and cervical cancer screening
812
21
22
23
24
25
26
27
28
29
30
31
among Spanish women with diabetes. Diabetes Care 2009; 32: 1470–1472. Jaen CR, Stange KC, Nutting PA. Competing demands of primary care: a model for the delivery of clinical preventive services. J Fam Pract 1994; 38: 166–171. Stange KC, Fedirko T, Zyzanski SJ, Jaen CR. How do family physicians prioritize delivery of multiple preventive services? J Fam Pract 1994; 38: 231–237. Chiu M, Austin PC, Manuel DG, Tu JV. Cardiovascular risk factor profiles of recent immigrants vs long-term residents of Ontario: a multi-ethnic study. Can J Cardiol 2012; 28: 20–26. Lysy Z, Booth GL, Shah BR, Austin PC, Luo J, Lipscombe LL. The impact of income on the incidence of diabetes: a population-based study. Diabetes Res Clin Pract 2013; 99: 372–379. Choi BC, Shi F. Risk factors for diabetes mellitus by age and sex: results of the National Population Health Survey. Diabetologia 2001; 44: 1221–1231. Kapral MK, Wang H, Mamdani M, Tu JV. Effect of socioeconomic status on treatment and mortality after stroke. Stroke 2002; 33: 268–273. Alter DA, Naylor CD, Austin P, Tu JV. Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction. N Engl J Med 1999; 341: 1359–1367. Katz SJ, Zemencuk JK, Hofer TP. Breast cancer screening in the United States and Canada, 1994: socioeconomic gradients persist. Am J Public Health 2000; 90: 799–803. Calle EE, Flanders WD, Thun MJ, Martin LM. Demographic predictors of mammography and Pap smear screening in US women. Am J Public Health 1993; 83: 53–60. Chiarelli AM, Halapy E, Nadalin V, Shumak R, O’Malley F, Mai V. Performance measures from 10 years of breast screening in the Ontario Breast Screening Program, 1990/91 to 2000. Eur J Cancer Prev 2006; 15: 34–42. Robles SC, Marrett LD, Clarke EA, Risch HA. An application of capture–recapture methods to the estimation of completeness of cancer registration. J Clin Epidemiol 1988; 41: 495–501. Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care 2002; 25: 512–516. Austin PC, van Walraven C, Wodchis WP, Newman A, Anderson GM. Using the Johns Hopkins Aggregated Diagnosis Groups (ADGs) to predict mortality in a general adult population cohort in Ontario. Canada. Med Care 2011; 49: 932–939. Martinez-Huedo MA, Lopez de Andres A, Hernandez-Barrera V, Carrasco-Garrido P, Martinez Hernandez D, Jimenez-Garcia R. Adherence to breast and cervical cancer screening in Spanish women with diabetes: associated factors and trend between, and 2010. Diabetes Metab 2006 and 2012; 38: 142–148. Marshall JG, Cowell JM, Campbell ES, McNaughton DB. Regional variations in cancer screening rates found in women with diabetes. Nurs Res 2010; 59: 34–41. Zhao G, Ford ES, Ahluwalia IB, Li C, Mokdad AH. Prevalence and trends of receipt of cancer screenings among US women with diagnosed diabetes. J Gen Intern Med 2009; 24: 270–275. Jaakkimainen RL, Barnsley J, Klein-Geltink J, Kopp A, Glazier RH. Did changing primary care delivery models change performance? A population-based study using health administrative data. BMC Fam Pract 2011; 12: 44. Grant RW, Pirraglia PA, Meigs JB, Singer DE. Trends in complexity of diabetes care in the United States from 1991 to 2000. Arch Intern Med 2004; 164: 1134–1139. Tabaei BP, Herman WH, Jabarin AF, Kim C. Does diabetes care compete with the provision of women’s preventive care services? Diabetes Care 2005; 28: 2644–2649.
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research: Epidemiology Impact of socio-economic status on breast cancer screening in women with diabetes: a population-based study W. Chan1, L. Yun2, P. C. Austin2,3, R. L. Jaakkimainen2,4,5, G. L. Booth2,3,6,7, J. Hux2,3,6, P. A. Rochon1,2,3,6 and L. L. Lipscombe1,2,3,6 1 Women’s College Research Institute, Women’s College Hospital, 2Institute for Clinical Evaluative Sciences, 3Institute of Health Policy, Management and Evaluation, University of Toronto, 4Department of Family and Community Medicine, University of Toronto, 5Sunnybrook Health Sciences Centre, 6Department of Medicine, University of Toronto and 7St Michael’s Hospital, Toronto, ON, Canada
Accepted 19 February 2014
Abstract Aims There is evidence to suggest that mammography rates are decreased in women with diabetes and in women of lower socio-economic status. Given the strong association between low socio-economic status and diabetes, we explored the extent to which differences in socio-economic status explain lower mammography rates in women with diabetes. Methods A population-based retrospective cohort study in Ontario, Canada, of women aged 50 to 69 years with diabetes between 1999 and 2010 age matched 1:2 to women without diabetes. Main outcome measure is the likelihood of at least one screening mammogram in women with diabetes within a 36-month period, starting as of either 1 January 1999, their 50th birthday, or 2 years after diabetes diagnosis – whichever came last. Outcomes were compared with those in women without diabetes during the same period as their matched counterparts, adjusting for socio-economic status based on neighbourhood income and other demographic and clinical variables.
Of 504 288 women studied (188 759 with diabetes, 315 529 with no diabetes), 63.8% had a screening mammogram. Women with diabetes were significantly less likely to have a mammogram after adjustment for socio-economic status and other factors (odds ratio 0.79, 95% CI 0.78–0.80). Diabetes was associated with lower mammogram use even in women from the highest socio-economic status quintile (odds ratio 0.79, 95% CI 0.75–0.83).
Results
Conclusions The presence of diabetes was an independent barrier to breast cancer screening, which was not explained by differences in socio-economic status. Interventions that target patient, provider, and health system factors are needed to improve cancer screening in this population.
Diabet. Med. 31, 806–812 (2014)
Introduction There is mounting evidence of an association between diabetes and cancer outcomes. Women with diabetes have been shown to have a higher incidence of post-menopausal breast cancer [1], as well as an increased mortality following a cancer diagnosis [2]. In addition to metabolic factors that influence tumour growth and progression [3], lower breast cancer screening in women with diabetes may also contribute to higher cancer mortality. Screening and early detection through mammography in post-menopausal women has been Correspondence to: Lorraine L. Lipscombe. E-mail: [email protected]
806
associated with increased breast cancer survival [4]. Although a mammogram is recommended every 3 years for women between 50 and 74 years of age [5], a significant proportion of eligible women do not receive routine screening [6]. There is evidence that cancer screening rates are reduced in patients with chronic diseases [7,8] and some studies have documented significantly lower mammography rates in women with diabetes [9–12]. Both healthcare provider and patient factors are known to influence the provision of adequate preventive care and thus may influence screening rates [13]. Indeed, a growing body of literature suggests that the competing demands of complex, chronic diseases such as diabetes interfere with attention to other clinical preventive
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
What’s new? • This study is the first to explore the influence of socio-economic status on the gap in screening mammograms among women with diabetes. • While there is evidence that the presence of diabetes is associated with a decrease in breast cancer screening, no study has examined the extent to which the higher prevalence of low socio-economic status in women with diabetes contributes to this disparity. • Using population-based data, we showed that diabetes is an independent barrier to adequate breast cancer screening even after adjusting for socio-economic status, and that low socio-economic status serves as an additional obstacle to regular screening mammograms. services [13,14]. Patient-related socio-demographic factors that are more common among patients with diabetes, such as recent immigration [15] and low socioeconomic status [16,17], may also contribute to cancer screening disparities. Socio-economic status is a well-known predictor of healthcare utilization, and socio-economic status-based inequities in access to care exist in Canada, despite the fact that most health care is universally funded [18,19]. Many studies have specifically shown lower cancer screening rates among low socio-economic status populations in Canada and elsewhere [20,21]. The extent to which lower socio-economic status among patients with diabetes contributes to disparities in cancer screening is unclear. A better understanding of the patientrelated factors that contribute to lower screening in women with diabetes will help to guide interventions aimed at improving cancer screening and detection in this population. This issue is particularly important in healthcare settings such as Canada, where breast cancer screening is universally subsidized by provincial health plans either via physician referrals or self-referral in Ontario through the Ontario Breast Cancer Screening Program. The objective of this population-based study was to compare receipt of a screening mammogram in Ontario between screen-eligible women with diabetes and age-matched women without diabetes, and to determine the influence of socio-economic status on this relationship.
Subjects and methods Study design and population
We conducted a retrospective, population-based cohort study of women aged 50–69 years with and without diabetes who were living in the province of Ontario, Canada from 1 January 1999 to 31 December 2010. This age group was chosen to reflect the Canadian breast cancer screening
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
DIABETICMedicine
guidelines during that time, which recommended a bilateral mammogram every 24 months in all women between 50 and 69 years of age [5]. We only included women who had at least one primary care visit in the year prior to cohort entry to ensure that all subjects had a similar opportunity to be referred for mammography screening during the follow-up period. These visits may have included both ‘walk-in’ and scheduled appointments, which cannot be distinguished within our data sets. We excluded women with a history of breast cancer, those living in a long-term care facility, those diagnosed with diabetes less than 2 years before the end of their maximum screening interval, and/or those who had less than 36 months of follow-up (because of death or a move from the province).
Data sources
We used linked, population-based healthcare databases, which record data on all Ontario residents covered by the universal provincial healthcare plan. We obtained information on demographics and deaths from the Registered Persons Database. Mammogram records were obtained from the Ontario Health Insurance Plan database, which provides information on physician service and radiologic procedure claims, and the Ontario Breast Cancer Screening Program database, which contains data on screening mammograms for women enrolled in this programme [22]. The Ontario Breast Cancer Screening Program database is a province-wide self-referral breast cancer-screening programme for screen-eligible women that provides routine mammograms, follow-up, and referral services for abnormal tests, and reminders for subsequent screening intervals [22]. We identified co-morbid diagnoses during hospital admissions from the Canadian Institute for Health Information and Discharge Abstract Database and breast cancer cases from the Ontario Cancer Registry [23]. We determined diabetes status from the Ontario Diabetes Database, a previously validated registry of Ontarians diagnosed with diabetes [24]. We linked administrative healthcare databases anonymously using encrypted health card numbers.
Procedures
We first identified all women aged 50–67 years between 1 January 1999 and 31 December 2007 who met inclusion criteria. We then selected all women with diabetes based on inclusion in the Ontario Diabetes Database during that time period. We assigned each subject an index date to begin follow-up, which was either 1 January 1999, their 50th birthday, or 2 years after a diabetes diagnosis, whichever came last (i.e. follow-up began once all three criteria were met). We began following women once they had diabetes for at least 2 years (i.e. prevalent diabetes) to avoid capturing mammograms that were ordered before diabetes diagnosis among newly diagnosed patients. We
807
DIABETICMedicine
Impact of socio-economic status on mammograms in women with diabetes W. Chan et al.
derived a cohort of eligible women without diabetes from the Ontario population, age matched 2:1 to women with diabetes and assigned them the same index date as their counterpart with diabetes.
Breast cancer screening and follow-up
The primary outcome was at least one screening mammogram within a 36-month period after the index date. While the recommended screening interval during our study era was 24 months, we chose 36 months to reflect the updated maximum screening interval in the 2011 Canadian screening guidelines [5]. We defined screening mammograms as either a bilateral mammogram recorded in the Ontario Health Insurance Plan database procedure claims database or a mammogram performed through the Ontario Breast Cancer Screening Program database.
Socio-economic status
We defined socio-economic status based on each individual’s neighbourhood income quintile recorded in the Registered Persons Database. We categorized individuals into neighbourhood income quintiles by linking their postal code data with Canadian census data on median household income levels by neighbourhood of residence. This method has been widely used to estimate socio-economic status [16,19] and has been shown to correlate well with health outcomes [19]. We excluded persons whose postal code could not be linked.
and co-morbidity. As a secondary analysis, we tested for an interaction between diabetes and socio-economic status quintile to assess the effect of socio-economic status quintile within the groups with diabetes and without diabetes, and the effect of diabetes within each socio-economic status quintile.
Ethics
This study was approved by the Research Ethics Board of Sunnybrook Health Sciences Centre in Toronto.
Results A total of 188 759 women with diabetes met the inclusion criteria and they were matched to 315 529 eligible women without diabetes. The index date was 1 January 1999 in 22.4%, as of age 50 years in 21.0%, and 2 years after diabetes diagnosis in 62.6% of women. Baseline characteristics for the groups with diabetes and without diabetes are presented in Table 1. The mean age at index date was 57.2 5.6 years. Women with diabetes were more likely to live in a low socio-economic status neighbourhood, to have immigrated to the province within the last 10 years, and to have a higher weighted co-morbidity score than those without diabetes (Table 1). Table 1 Baseline characteristics—diabetes compared with no diabetes among those who had a primary care visit within previous year Characteristic
Other covariates
We recorded the following baseline characteristics: age, rural residence defined as a dissemination area with less than 10 000 residents, recent immigration into Ontario based on entry into the province less than 10 years before index date (as determined by new enrolment in the provincial healthcare programme), and co-morbidity based on diagnoses recorded on physician claims in the 2 years prior to cohort entry. We derived the weighted co-morbidity score from the Johns Hopkins Adjusted Clinical Group Case-Mix assignment software, which assigns diagnoses into 32 Aggregated Diagnosis Groups. This method has been validated to predict mortality in the general adult population [25].
Statistical analyses
We used descriptive statistics (v2-tests) to compare baseline variables between women with and without diabetes. We used univariate and multivariable conditional logistic regression to compare the likelihood of a screening mammogram between women with and without diabetes. For our primary analysis, we assessed the association between diabetes and mammogram use, adjusting for age, socio-economic status quintile, rural residence, recent immigration into Ontario
808
Mean age at index date (years) Socio-economic status quintile*, n (%) Socio-economic status 1 (lowest) Socio-economic status 2 Socio-economic status 3 Socio-economic status 4 Socio-economic status 5 (highest) Rural residence†, n (%) Recent immigrant (resident < 10 years)‡, n (%) Weighted co-morbidity score > 10§, n (%)
Diabetes
No diabetes
n = 188 759 57.1 5.6
n = 315 529 57.2 5.6
46 283 (24.5%)
56 590 (17.9%)
42 097 (22.3%)
62 912 (19.9%)
39 414 (20.9%)
64 866 (20.6%)
32 775 (17.4%)
63 203 (20.0%)
28 190 (14.9%)
67 958 (21.5%)
24 283 (12.9%) 21 365 (11.3%)
43 447 (13.8%) 26 229 (8.3%)
31 686 (16.8%)
30 968 (9.8%)
*Based on neigbhourhood median household income derived from census data and postal codes. † Rural residence defined as a dissemination area with less than 10 000 residents. ‡ Recent immigrant defined as entry into province less than 10 years prior to index date. § Based on physician claims in the previous 2 years listing > 10 diagnostic groups from Johns Hopkins Aggregated Diagnosis Groups (ADG).
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
Screening mammograms
During the 36-month follow-up period, a screening mammogram was recorded in 321 564 (63.8%) women from our cohort: 113 873 (60.3%) in those with diabetes and 207 691 (65.8%) in women without diabetes (P < 0.001). Among those who received mammograms, 61.1% were screened through physician referrals (Ontario Health Insurance Plan database) and 38.9% were screened through the Ontario Breast Cancer Screening Program. Women with diabetes were slightly more likely to be tested through the Ontario Breast Cancer Screening Program database than those without diabetes (39.8% vs. 38.4% of tests), but the rate of mammograms remained significantly lower for women with diabetes regardless of the service used (diabetes vs. no diabetes: 36.3% vs. 40.6% for Ontario Health Insurance Plan database, P < 0.001; and 24.0% vs. 25.3% for Ontario Breast Cancer Screening Program database, P < 0.001). The majority of women were followed for 36 months (97.9% with diabetes, 94.2% without diabetes). Among those women whose follow-up was terminated earlier, 0.8% were diagnosed with breast cancer and 0.5% lost Ontario Health Insurance Plan database eligibility in each group, 0.8% with diabetes and 0.2% without diabetes entered a long-term care facility, and 4.4% of women without diabetes were subsequently diagnosed with diabetes. The rate of screening mammograms increased with increasing socio-economic status quintile, both overall and within women with and without diabetes (Fig 1). As illustrated in Fig 1, the proportion of women who had a mammogram was lower in women with diabetes across all socio-economic status groups, with the lowest rate in women with diabetes from the lowest socio-economic status quintile
FIGURE 1 Rate of screening mammograms among screen-eligible women within 36 months with increasing socio-economic status quintile, both overall and within women with ( ) and without ( ) diabetes.
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
DIABETICMedicine
(55.9%) and the highest rate in women without diabetes in the highest socio-economic status quintile (71.8%). Women with diabetes were significantly less likely to have a screening mammogram in the subsequent 36-month period compared with women without diabetes (unadjusted odds ratio 0.79, 95% CI 0.78–0.80). This relationship persisted after adjustment for age, socio-economic status quintile, recent immigration, rural residence, and co-morbidity score (adjusted odds ratio 0.86, 95% CI 0.85–0.87) (Table 2). As shown in Table 2, lower socio-economic status was also significantly associated with lower mammogram use on multivariable analysis (odds ratio 0.61, 95% CI 0.60–0.63 for socio-economic status quintile 1 vs. 5; odds ratio 0.88, 95% CI 0.86–0.90 for socio-economic status quintile 4 vs. 5). While socio-economic status was associated with a significant decrease in mammograms in women with and without diabetes, a significant two-way interaction was found between diabetes and socio-economic status 1, 2, and 3 versus 4 and 5 (P < 0.0001). The influence of socio-economic status was relatively more pronounced in women without diabetes (socio-economic status 1 vs. 5: odds ratio 0.67, 95% CI 0.63–0.71 in patients with diabetes; odds ratio 0.58, 95% CI 0.55–0.61 in patients without diabetes). In addition, the influence of diabetes was stronger in the highest income quintile compared with the lowest quintile (diabetes vs. no diabetes: odds ratio 0.79, 95% CI 0.75–0.83 in socio-economic status 5; odds ratio 0.91, 95% CI 0.88– 0.94 in socio-economic status 1) (Table 2).
Discussion This large, population-based study found that women with diabetes were 14% less likely to receive a mammogram during the recommended screening interval compared with those without diabetes. This disparity was not explained by differences in socio-economic status, immigration status, region of residence, or co-morbidity. Lower socio-economic status was also an independent predictor of decreased mammogram use in both women with and without diabetes, with the lowest screening rate found in women with both low socio-economic status and diabetes. Diabetes remained a significant barrier to screening mammogram tests even among women from the highest socio-economic status quintile. Indeed, the effect of diabetes was relatively stronger in women of higher socio-economic status, and the effect of low socio-economic status was more salient among women without diabetes. These findings suggest that while both diabetes and socio-economic status independently affect breast cancer screening, the relative contribution of each factor to screening disparities may be greater in populations without competing risk factors. This study illustrates the complex relationship between patient and disease-specific factors and healthcare utilization. Diabetes represents an important barrier to adequate cancer screening, which is not attributable to socio-demographic factors. Rather, low
809
DIABETICMedicine
Impact of socio-economic status on mammograms in women with diabetes W. Chan et al.
Table 2 Number and proportion of women who had a screening mammogram, and results of logistic regression analysis of associations between diabetes, socio-economic status, and mammogram use Total (n)
Per cent screened
Overall Diabetes Yes 188 759 60.3% No 315 529 65.8% Socio-economic status quintile Socio-economic status 1 (lowest) 102 873 57.3% Socio-economic status 2 105 009 61.5% Socio-economic status 3 104 280 64.5% Socio-economic status 4 95 978 67.4% Socio-economic status 5 (referent) 96 148 70.2% Effect of socio-economic status in women with diabetes Socio-economic status quintile Socio-economic status 1 (lowest) 46 283 55.9% Socio-economic status 2 42 097 59.4% Socio-economic status 3 36 824 61.9% Socio-economic status 4 32 775 63.9% Socio-economic status 5 (referent) 28 190 66.2% Effect of socio-economic status in women without diabetes Socio-economic status quintile Socio-economic status 1 (lowest) 56 590 58.5% Socio-economic status 2 62 912 63.0% Socio-economic status 3 63 223 66.1% Socio-economic status 4 63 203 69.2% Socio-economic status 5 (referent) 67 958 71.8% Effect of diabetes within socio-economic status groups Socio-economic status quintile 1 (lowest) Diabetes 46 283 55.9% No diabetes 56 590 58.5% Socio-economic status quintile 2 Diabetes 42 097 59.4% No diabetes 62 912 63.0% Socio-economic status quintile 3 Diabetes 36 824 61.9% No diabetes 63 223 66.1% Socio-economic status quintile 4 Diabetes 32 775 63.9% No diabetes 63 203 69.2% Socio-economic status quintile 5 (highest) Diabetes 28 190 66.2% No diabetes 67 958 71.8%
Unadjusted odds ratio (95% CI)
Adjusted odds ratio* (95% CI)
0.79 (0.78–0.80) 1.00
0.86 (0.85–0.87) 1.00
0.56 0.67 0.72 0.86 1.00
(0.54–0.57) (0.65–0.69) (0.70–0.74) (0.84–0.88)
0.61 0.71 0.75 0.88 1.00
(0.60–0.63) (0.69–0.73) (0.74–0.77) (0.86–0.90)
0.64 0.74 0.75 0.89 1.00
(0.60–0.68) (0.70–0.78) (0.71–0.79) (0.85–0.93)
0.67 0.75 0.77 0.90 1.00
(0.63–0.71) (0.72–0.79) (0.73–0.81) (0.86–0.94)
0.55 0.67 0.74 0.87 1.00
(0.52–0.58) (0.65–0.69) (0.65–0.69) (0.85–0.89)
0.58 0.69 0.75 0.88 1.00
(0.55–0.61) (0.67–0.71) (0.73–0.77) (0.86–0.90)
0.89 (0.86–0.92) 1.00
0.91 (0.88–0.94) 1.00
0.85 (0.82–0.88) 1.00
0.87 (0.83–0.90) 1.00
0.78 (0.75–0.82) 1.00
0.81 (0.78–0.84) 1.00
0.79 (0.75–0.82) 1.00
0.81 (0.77–0.84) 1.00
0.77 (0.73–0.81) 1.00
0.79 (0.75–0.83) 1.00
*Adjusted for socio-economic status, co-morbidity, rural residence, immigration status.
socio-economic status exerts an additive influence on the quality of care for this already vulnerable population. Our findings support those of other studies showing lower preventive care in patients with chronic diseases [7,8,26]. However, evidence regarding breast cancer screening among women with diabetes has been mixed [9–12,27,28]. In two studies examining USA survey data from the general population, diabetes was not associated with lower mammogram use [27,28]. It is possible that the greater healthcare access among US women with diabetes compared with the general population may have led to more opportunities for cancer screening in those studies, thus offsetting any differences between groups. In contrast, diabetes was associated with a significant reduction in mammograms when populations were limited to women with similar healthcare access, as in this study and in others [9–12,26]. For example, one study
810
from Spain (where eligible women are systematically invited for mammograms) found that women with diabetes were 16% less likely to report receiving a mammogram than their counterparts without diabetes (37.9% vs. 53.8%) [26]. The overall screening rates were lower than in the current study, possibly because mammography was based on self-report rather than claims. Our earlier study among Ontario women between 1999 and 2002 also had lower mammogram rates than this study (38.1% and 47.3% in women with and without diabetes, respectively), and found a 30% lower likelihood of a mammogram in women with diabetes compared to control subjects without [9]. However, that study was missing data on approximately 38% of mammograms performed through the Ontario Breast Cancer Screening Program database, which might have been used more frequently by women with diabetes. Nonetheless, our present
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
Research article
findings are consistent with our earlier evidence of lower mammogram use among women with diabetes, even after accounting for tests performed through the Ontario Breast Cancer Screening Program database and controlling for factors such as socio-economic status, immigration status, and co-morbidity. This study specifically explored whether low socio-economic status explained the relationship between diabetes and breast cancer screening, as this factor is a well-known predictor of both diabetes [16,17] and cancer screening [20,21]. Indeed, we found that populations of lower socio-economic status were significantly less likely to undergo a mammogram during the recommended screening interval. This finding supports other studies that have shown substantial socio-economic disparities in cancer screening practices in Canada and European countries despite universal health coverage [12,20,26]. Socio-economically disadvantaged groups may have lower preventive healthcare utilization as a result of lower health literacy, less advocacy, discrimination, and impaired access, for example [21]. Although women with diabetes were more likely to have a lower socio-economic status in our study, we found that the disparities in mammogram use persisted after adjusting for socio-economic status. Moreover, diabetes remained a significant barrier to breast cancer screening even in higher socio-economic status populations. We note that the gap in screening because of diabetes has narrowed in recent years compared with our previous study [9]. This improvement may be attributable in part to differences in health system factors, such as the introduction of a screening incentive programme for physicians [29] and/or the present study’s inclusion of tests performed through the Ontario Breast Cancer Screening Program database. Office based diabetes management has become significantly more complex and time-consuming as a result of a shift toward more intensive therapeutic interventions [30]. Given these increasing demands on physicians and greater time constraints because of larger patient loads, preventive issues such as cancer screening may be more often overlooked. Programmes that offer incentives and reminders for cancer screening or allow for self-referral may lead to greater adherence, particularly for patients with multiple healthcare needs. Indeed, the difference in mammograms between the groups with and without diabetes was much smaller among women screened through the Ontario Breast Cancer Screening Program database. It is possible that such organized screening programmes may alleviate some of the competing demands faced by providers caring for complex patients with diabetes. Patients with diabetes also often see multiple physicians, including specialists who are less familiar with guidelines outside of their field. This fragmentation of care might hamper the effective use of healthcare services. Interestingly, one study in patients with diabetes within a US managed care setting found that primary care physicians with better diabetes-related processes of care were also more likely to perform mammo-
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
DIABETICMedicine
grams [31]. These findings highlight the role of both health system and provider factors and suggest that better diabetes care may correlate rather than compete with cancer screening. Interventions to improve cancer screening in patients with diabetes should therefore be multifaceted, focusing on more standardized health system measures, increased support and incentives for diabetes care providers, and greater education for more socially disadvantaged populations. Strengths of our study include the use of population-based data for a large and diverse jurisdiction, and a validated algorithm to identify diabetes. Some limitations should be noted. As with all observational studies, we cannot exclude the role of selection bias. While we were able to adjust for certain patient factors that may influence screening practices, we lacked data on other factors, including family history of breast cancer, lifestyle, ethnicity, education, obesity, and attitudes toward screening. Moreover, in cases where mammograms were not obtained, we cannot determine whether the physician failed to refer the patient or the patient chose not to obtain the recommended test. Finally, in the absence of individual-level income data, we based our measure of socio-economic status on neighbourhood-level household income, which may be less accurate for more mixed neighbourhoods, or for certain groups such as retired persons or live-in domestic workers. Nonetheless, neighbourhood income is widely used as a measure of socio-economic status that has been shown to correlate well with individual level measures and health outcomes [16,18,19]. In conclusion, findings from this large, population-based study suggest that women with diabetes are a unique population at risk for inadequate breast cancer screening, and that low socio-economic status is an additional obstacle to preventive care in an already disadvantaged population. These findings were not explained by differences in co-morbidity, immigration status, or socio-economic status in women with diabetes. These results are of particular importance as women with diabetes are at higher risk of breast cancer and of poorer survival once diagnosed. Interventions to improve cancer screening that target patient, provider, and health system factors need to be developed to ensure ongoing provision of comprehensive care for patients with diabetes. Previous presentation
An abstract of this work was presented at the 15th Annual Canadian Diabetes Association Conference (10–13 October 2012) in Vancouver, British Columbia.
Funding sources
LLL is supported by a Canadian Institutes of Health Research (CIHR) New Investigator Award. PCA is supported in part by a Career Investigator award from the Heart and Stroke Foundation. This study was conducted with the support from a CIHR operating grant (MOP no. 123263),
811
DIABETICMedicine
Impact of socio-economic status on mammograms in women with diabetes W. Chan et al.
and from the Ontario Institute for Cancer Research and Cancer Care Ontario. This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred. The funding bodies had no involvement in the study design, data collection, data analysis, manuscript preparation, and/or publication decisions.
13
14
15
16
17
Competing interests
None declared.
18
Acknowledgements
19
The authors would like to thank Hadas Fischer MD MSc, Institute for Clinical Evaluative Sciences, for assistance with design and implementation of the study, and Lauren McNicol MA, Women’s College Research Institute, for help with manuscript preparation.
20
References 1 Boyle P, Boniol M, Koechlin A, Robertson C, Valentini F, Coppens K et al. Diabetes and breast cancer risk: a meta-analysis. Br J Cancer 2012; 107: 1608–1617. 2 Peairs KS, Barone BB, Snyder CF, Yeh HC, Stein KB, Derr RL et al. Diabetes mellitus and breast cancer outcomes: a systematic review and meta-analysis. J Clin Oncol 2011; 29: 40–46. 3 Godsland IF. Insulin resistance and hyperinsulinaemia in the development and progression of cancer. Clin Sci 2010; 118: 315–332. 4 Tabar L, Yen MF, Vitak B, Chen HH, Smith RA, Duffy SW. Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening. Lancet 2003; 361: 1405–1410. 5 Canadian Task Force on Preventive Health Care. Screening for Breast Cancer: Recommendations for Clinicians and Policymakers. Ottawa: Health Canada, 2011. 6 Cancer Quality Council of Ontario. Cancer System Quality Index 2012: Breast Cancer Screening (Mammography) Participation. Ontario: Cancer Quality Council of Ontario, 2012. 7 Fontana SA, Baumann LC, Helberg C, Love RR. The delivery of preventive services in primary care practices according to chronic disease status. Am J Public Health 1997; 87: 1190–1196. 8 Coughlin SS, Uhler RJ, Hall HI, Briss PA. Non-adherence to breast and cervical cancer screening: what are the linkages to chronic disease risk? Prev Chronic Dis 2004; 1: A04. 9 Lipscombe LL, Hux JE, Booth GL. Reduced screening mammography among women with diabetes. Arch Intern Med 2005; 165: 2090–2095. 10 McBean AM, Yu X. The underuse of screening services among elderly women with diabetes. Diabetes Care 2007; 30: 1466–1472. 11 Beckman TJ, Cuddihy RM, Scheitel SM, Naessens JM, Killian JM, Pankratz VS. Screening mammogram utilization in women with diabetes. Diabetes Care 2001; 24: 2049–2053. 12 Jimenez-Garcia R, Hernandez-Barrera V, Carrasco-Garrido P, Gil A. Prevalence and predictors of breast and cervical cancer screening
812
21
22
23
24
25
26
27
28
29
30
31
among Spanish women with diabetes. Diabetes Care 2009; 32: 1470–1472. Jaen CR, Stange KC, Nutting PA. Competing demands of primary care: a model for the delivery of clinical preventive services. J Fam Pract 1994; 38: 166–171. Stange KC, Fedirko T, Zyzanski SJ, Jaen CR. How do family physicians prioritize delivery of multiple preventive services? J Fam Pract 1994; 38: 231–237. Chiu M, Austin PC, Manuel DG, Tu JV. Cardiovascular risk factor profiles of recent immigrants vs long-term residents of Ontario: a multi-ethnic study. Can J Cardiol 2012; 28: 20–26. Lysy Z, Booth GL, Shah BR, Austin PC, Luo J, Lipscombe LL. The impact of income on the incidence of diabetes: a population-based study. Diabetes Res Clin Pract 2013; 99: 372–379. Choi BC, Shi F. Risk factors for diabetes mellitus by age and sex: results of the National Population Health Survey. Diabetologia 2001; 44: 1221–1231. Kapral MK, Wang H, Mamdani M, Tu JV. Effect of socioeconomic status on treatment and mortality after stroke. Stroke 2002; 33: 268–273. Alter DA, Naylor CD, Austin P, Tu JV. Effects of socioeconomic status on access to invasive cardiac procedures and on mortality after acute myocardial infarction. N Engl J Med 1999; 341: 1359–1367. Katz SJ, Zemencuk JK, Hofer TP. Breast cancer screening in the United States and Canada, 1994: socioeconomic gradients persist. Am J Public Health 2000; 90: 799–803. Calle EE, Flanders WD, Thun MJ, Martin LM. Demographic predictors of mammography and Pap smear screening in US women. Am J Public Health 1993; 83: 53–60. Chiarelli AM, Halapy E, Nadalin V, Shumak R, O’Malley F, Mai V. Performance measures from 10 years of breast screening in the Ontario Breast Screening Program, 1990/91 to 2000. Eur J Cancer Prev 2006; 15: 34–42. Robles SC, Marrett LD, Clarke EA, Risch HA. An application of capture–recapture methods to the estimation of completeness of cancer registration. J Clin Epidemiol 1988; 41: 495–501. Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care 2002; 25: 512–516. Austin PC, van Walraven C, Wodchis WP, Newman A, Anderson GM. Using the Johns Hopkins Aggregated Diagnosis Groups (ADGs) to predict mortality in a general adult population cohort in Ontario. Canada. Med Care 2011; 49: 932–939. Martinez-Huedo MA, Lopez de Andres A, Hernandez-Barrera V, Carrasco-Garrido P, Martinez Hernandez D, Jimenez-Garcia R. Adherence to breast and cervical cancer screening in Spanish women with diabetes: associated factors and trend between, and 2010. Diabetes Metab 2006 and 2012; 38: 142–148. Marshall JG, Cowell JM, Campbell ES, McNaughton DB. Regional variations in cancer screening rates found in women with diabetes. Nurs Res 2010; 59: 34–41. Zhao G, Ford ES, Ahluwalia IB, Li C, Mokdad AH. Prevalence and trends of receipt of cancer screenings among US women with diagnosed diabetes. J Gen Intern Med 2009; 24: 270–275. Jaakkimainen RL, Barnsley J, Klein-Geltink J, Kopp A, Glazier RH. Did changing primary care delivery models change performance? A population-based study using health administrative data. BMC Fam Pract 2011; 12: 44. Grant RW, Pirraglia PA, Meigs JB, Singer DE. Trends in complexity of diabetes care in the United States from 1991 to 2000. Arch Intern Med 2004; 164: 1134–1139. Tabaei BP, Herman WH, Jabarin AF, Kim C. Does diabetes care compete with the provision of women’s preventive care services? Diabetes Care 2005; 28: 2644–2649.
ª 2014 The Authors. Diabetic Medicine ª 2014 Diabetes UK
