SEARCH

SEARCH BY CITATION

Keywords:

  • cancer stage;
  • disparities;
  • socioeconomic status (SES);
  • outcomes

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

BACKGROUND:

Lower socioeconomic status (SES) is associated with worsened cancer survival. The authors evaluate the impact of SES on stage of cancer at diagnosis and survival in Ontario, Canada.

METHODS:

All incident cases of breast, colon, rectal, nonsmall cell lung, cervical, and laryngeal cancer diagnosed in Ontario during the years 2003-2007 were identified by using the Ontario Cancer Registry. Stage information is captured routinely for patients seen at Ontario's 8 Regional Cancer Centers (RCCs). The Ontario population was divided into quintiles (Q1-Q5) based on community median household income reported in the 2001 census; Q1 represents the poorest communities. Overall survival (OS) and cancer-specific survival (CSS) were determined with Kaplan-Meier methodology. A Cox model was used to evaluate the association between survival and SES, stage, and age.

RESULTS:

Stage at diagnosis was available for 38,431 of 44,802 (85%) of cases seen at RCCs. The authors observed only very small differences in stage distribution by SES. Across all cases in Ontario, the authors found substantial gradients in 5-year OS and 3-year CSS across Q1 and Q5 for breast (7% absolute difference in OS, P < .001; 4% CSS, P < .001), colon (8% OS, P < .001; 3% CSS, P = .002), rectal (9% OS, P < .001; 4% CSS, P = .096), nonsmall cell lung (3% OS, P = .002; 2% CSS, P = .317), cervical (16% OS, P < .001; 10% CSS, P = .118), and laryngeal cancers (1% OS, P = .045; 3% CSS, P = .011). Adjustments for stage and age slightly diminished the survival gradient only among patients with breast cancer.

CONCLUSIONS:

Despite universal healthcare, SES remains associated with survival among patients with cancer in Ontario, Canada. Disparities in outcome were not explained by differences in stage of cancer at time of diagnosis. Cancer 2010. © 2010 American Cancer Society.

Lower socioeconomic status (SES) is known to be associated with worsened survival and increased incidence of cancer.1-9 In Canada, most healthcare for all residents is provided by universal, comprehensive, provincial health insurance plans. Residents make no direct payment for medical services or hospitalization (including surgery, radiotherapy, intravenous chemotherapy), and there is no parallel private sector. Despite having universal health insurance, previous work has demonstrated that disparities in cancer-related outcomes exist in Canada.2, 9, 10 A comparative study has found the magnitude of this association to be only slightly less in Ontario than that observed in the United States.3 Various hypotheses to explain survival differences between social groups have been proposed in the literature including differences in tumor biology, patient comorbidity, stage of disease at diagnosis, access to therapy, and treatment practices.11

Differences in stage of disease is commonly cited as a potential mechanism for the observed relation between SES and cancer outcomes. In their overview of the existing literature, Woods et al conclude that stage of disease at diagnosis and access to optimal treatment explain a portion of disparity in survival of patients with cancer.11 Although much of the literature from the United States has found an association between SES and stage of disease at diagnosis,4, 11, 12 3 large studies conducted in the United Kingdom failed to confirm this observation despite observing disparities in survival across social groups.13-15 It remains unknown whether these inconsistent results are related to the availability of universal health insurance in the United Kingdom. To date, only a single study has evaluated stage of cancer at diagnosis and SES within Canada, where the healthcare system is designed to provide equitable access to equivalent standards of care. In their retrospective, population-based, patterns- of-care study in laryngeal cancer diagnosed during the years 1982-1995 in Ontario, Groome et al found an association between SES and survival of patients with glottic (but not supraglottic) cancer that was only partly explained by stage of disease.10 Given the limitations of the current literature, we designed the current study to evaluate the relations among SES, stage of disease, and survival in a contemporary, population-based cohort of patients diagnosed with cancer in Ontario, Canada.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Study Design

In this population-based study, all incident cases of breast, colon, rectal, nonsmall cell lung (NSCL), cervical, and laryngeal cancer diagnosed in Ontario between January 1, 2003 and October 31, 2007 were identified by using the Ontario Cancer Registry. These diseases were chosen because they are common sites of cancer and represent a mix of diseases that are managed surgically and nonsurgically. Cases that were not confirmed microscopically and benign/in situ neoplasms were excluded. Stage at diagnosis is only captured routinely for patients seen at Ontario's 8 Regional Cancer Centers (RCCs); this represents approximately 64% of the population. To increase the proportion of patients for whom staging information is captured, we restricted our analyses to new cases seen at any 1 of the province's RCCs. Cases were defined as new when they were seen within 3 months of diagnosis or surgery (whichever was most recent). Information about the SES of the community in which the patient resided at the time of diagnosis was linked to the Ontario Cancer Registry from the 2001 Canadian census. The population was divided into quintiles (Q1-Q5) on the basis of community median household income.

Sources of Data

The Ontario Cancer Registry is a passive, population-based, cancer registry that captures diagnostic and demographic information on at least 98% of all incident cases of cancer diagnosed in Ontario, the most populous province of Canada.16, 17 Overall survival (OS) and cancer-specific survival (CSS) were calculated from the registry data; vital status is current through October 2008 and cause of death through December 2006. ICD-9 codes (140.x-208.x) were used to identify death from any cancer. The clinical databases of the RCCs provided information related to stage of disease. Hospital records from the Canadian Institute of Health Information (CIHI) provided information regarding cancer-related surgical procedures. The study was approved by the Queen's University Research Ethics Board.

Creating Stage Groups

Variables available through the stage database included pathologic and clinical tumor, nodal and metastasis stage (pT, pN, pM and cT, cN, cM) as recorded by the treating physician. A user-determined summary stage was also available; however, this was used only if the TNM-derived stage was not available because of missing T, N, or M information. By using a best stage grouping approach, cases were assigned stage based on pathologic TNM when available and clinical TNM otherwise. If M status was missing, it was set by default to M0 in the TNM grouping. Once pathologic or clinical M was identified as M1, then both clinical and pathologic stages were IV. Cases were classified as unstaged when no pathologic or clinical stage was available.

Creating SES Groups

The population of Ontario was divided into quintiles (Q1-Q5)) based on community median household income reported in the 2001 Canadian census. The unit of observation was an electoral enumeration area that contained a maximum of 650 residences in the 2001 census.18 The enumeration area data were linked to the individual study participants based on postal code of residence at the time of diagnosis.19 Quintiles of the median household income were based on the household income distribution for the full province of Ontario. Q1 represents the communities where the poorest 20% of the Ontario population resided.

Statistical Analyses

Comparisons of proportions between SES and age groupings were made using the chi-square test. OS and CSS were determined by using the Kaplan-Meier technique; summary measures are reported as hazard ratios (HRs). To ensure a minimum of 1 year for follow-up, CSS was determined only for those cases diagnosed between the years 2003 and 2005. The Cox proportional hazards model was used to describe associations between SES and survival while controlling for stage of disease at diagnosis and age.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

During the study period, 98,885 patients with breast, colon, rectal, NSCL, cervical, and laryngeal cancer were identified in the Ontario Cancer Registry. Two-thirds (63,650 of 98,885; 64%) of cases were seen at the province's 8 RCCs. Among these, there were 44,802 new cases seen within 3 months of diagnosis at the RCCs during 2003-2007. Subsequent data presented relating to stage of disease pertain to this population. The median household income of the communities from which our study population was derived were: Q1, $24,000; Q2, $31,000; Q3, $36,000; Q4, $44,000; and Q5, $55,000 (all figures in United States dollars).

Stage at Diagnosis

Data shown in Table 1 describe the number of cases in each disease and the proportion of which had available stage information. Patients with breast (18,197 of 19,543; 93%), colon (6312 of 7976; 79%), and rectal cancer (2224 of 3713; 60%) were more likely to undergo surgical resection than nonsmall cell lung (1694 of 12,276; 14%), cervical (187 of 1052; 18%), or laryngeal cancer (82 of 854; 10%). Pathologic stage was more commonly available for diseases managed primarily by surgery. Stage information was available for 85% (38,431 of 44,802) of all cases.

Table 1. Availability of Stage at Diagnosis for Patients Seen at a Regional Cancer Center in Ontario 2003-2007
DiseaseCases With Pathologic StageCases With Clinical StageCases With No StageTotal Cases With Stage
Breast15784/19543 (80.8%)8005/19543 (41.0%)2640/19543 (13.5%)16903/19543 (86.5%)
Colon6401/7976 (80.3%)1658/7976 (20.8%)1380/7976 (17.3%)6596/7976 (82.7%)
Rectum2741/3713 (73.8%)1063/3713 (28.6%)607/3713 (16.3%)3106/3713 (83.7%)
NSCL6479/12276 (52.8%)8465/12276 (69.0%)2213/12276 (18.0%)10063/12276 (82.0%)
Cervix416/1052 (39.5%)648/1052 (61.6%)86/1052 (8.2%)966/1052 (91.8%)
Larynx64/854 (7.5%)762/854 (89.2%)57/854 (6.7%)797/854 (93.3%)
All diseases31885/45414 (70.2%)20601/45414 (45.4%)6983/44802 (15.4%)38431/44802 (84.6%)

Distribution of stage by SES quintile is shown for the complete study population and individual diseases in Figure 1. Data shown in Table 2 depict the proportion of patients from Q1 compared with Q2-Q5 with stage I and stage IV disease. We observed only small differences in stage distribution for breast and rectal cancers. For all 6 cancers combined, cases in Q1 were less likely than Q2-Q5 to have stage 1 disease (26 vs 29%; P < .001) and more likely to have stage 4 disease (23 vs 20%; P < .001).

thumbnail image

Figure 1. Distribution of stage at diagnosis by socioeconomic status is depicted for patients seen at a Regional Cancer Center in Ontario during the years 2003-2007.

Download figure to PowerPoint

Table 2. Proportion of Patients With Stage I and Stage IV Cancer at Time of Diagnosis by SES Among Patients With Stage Information Seen at a Regional Cancer Center in Ontario 2003-2007
 Stage I DiseaseStage IV Disease
Q1aQ2-5aPQ1Q2-5P
  • NSCL indicates nonsmall cell lung (cancer).

  • a

    Q1 refers to patients from the lowest quintile socioeconomic status (SES), and Q2-5 includes patients from all other quintiles.

Breast1241/2820 (44%)6532/14083 (46%).021135/2820 (5%)541/14083 (4%).019
Colon105/1242 (9%)508/5354 (10%).259295/1242 (24%)1321/5354 (25%).497
Rectum107/638 (17%)413/2468 (17%).982160/638 (25%)504/2468 (20%).011
NSCL300/2234 (13%)999/7829 (13%).4061050/2234 (47%)3799/7829 (49%).204
Cervix84/225 (37%)286/741 (39%).73318/225 (8%)84/741 (11%).154
Larynx80/191 (42%)281/606 (46%).27842/191 (22%)102/606 (17%).106
All diseases1917/7350 (26%)9019/31081 (29%)<.0011700/7350 (23%)6351/31081 (20%)<.001

Survival, Socioeconomic Status, and Stage of Disease

Among all cases within the Ontario Cancer Registry (ie, RCC and non-RCC patients), we observed statistically significant and clinically meaningful differences in overall survival (OS) across social groups for all diseases (Table 3). Significant gradients in cancer-specific survival (CSS) were seen in breast, colon, and laryngeal cancers (Table 4). In the subpopulation of cases seen at RCCs we found significant survival differences in breast (OS and CSS) and colon (OS) cancers.

Table 3. Five Year Overall Survival (%) of All Cases in Ontario Cancer Registry (OCR) and of Cases Seen at Regional Cancer Centers (RCC) 2003-2007 by Socioeconomic Quintile
All Cases in OCR
SESBreastColonRectumNSCLCervixLarynx
  • a

    P-value calculated using the log-rank test; 95% CI shown in parentheses.

176.5 (75.0-78.0)52.0 (49.9-54.1)51.5 (48.5-54.4)15.4 (14.2-16.8)63.0 (57.5-68.0)59.0 (53.3-64.2)
279.2 (77.8-80.4)53.3 (51.4-55.2)52.4 (49.4-55.3)16.5 (15.2-17.8)70.5 (65.0-75.3)54.0 (47.8-59.8)
380.9 (79.6-82.1)54.2 (52.2-56.1)56.6 (53.5-59.5)16.4 (15.0-17.8)71.4 (66.3-75.9)60.4 (53.4-66.7)
483.0 (81.7-84.1)57.0 (54.9-59)57.7 (54.7-60.6)17.3 (15.8-18.9)73.3 (67.7-78.1)56.8 (48.5-64.4)
583.6 (82.4-84.8)59.8 (57.7-61.9)60.0 (56.8-63.1)18.6 (16.9-20.4)78.7 (73.5-83.0)60.0 (52.2-66.9)
Pa<.001<.001<.001.002<.001.045
All80.8 (80.2-81.3)55.1 (54.2-56.0)55.5 (54.1-56.8)16.7 (16.0-17.3)71.0 (68.6-73.2)58.0 (55.0-60.9)
New Cases Seen at RCC
SESBreastColonRectumNSCLCervixLarynx
179.3 (77.2-81.3)51.0 (47.4-54.5)50.0 (45.3-54.6)9.0 (7.5-10.7)56.5 (46.6-65.2)63.8 (55.9-70.7)
283.5 (81.9-85.0)55.6 (52.5-58.6)52.1 (47.4-56.6)9.7 (8.2-11.3)54.3 (45.5-62.4)60.4 (51.0-68.6)
385.4 (83.7-86.8)51.6 (48.0-55.1)52.8 (47.7-57.6)10.6 (9.0-12.4)59.6 (50.4-67.6)60.6 (50.4-69.4)
486.3 (84.6-87.8)56.6 (53.0-60.1)55.4 (50.1-60.4)11.2 (9.3-13.4)65.8 (57.8-72.7)61.5 (49.6-71.4)
587.8 (86.2-89.3)60.6 (56.8-64.2)53.4 (47.1-59.2)12.7 (10.5-15)74.7 (66.3-81.3)64.4 (52.2-74.3)
Pa<.001<.001.100.471.076.986
All82.6 (82.0-83.3)51.9 (50.5-53.3)55.6 (53.8-57.4)12.0 (11.3-12.8)67.4 (64.4-70.2)58.5 (55-61.9)
Table 4. Three Year Cancer-Specific Survival (%) of all Cases in Ontario Cancer Registry (OCR) and of Cases Seen at Regional Cancer Centers (RCC) 2003-2005 by Socioeconomic Quintile
All Cases in OCR
SESBreastColonRectumNSCLCervixLarynx
  • a

    P-value calculated using the log-rank test; 95% CI shown in parentheses.

188.2 (86.9-89.4)68.4 (66.3-70.4)67.9 (64.8-70.9)25.8 (24.1-27.6)71.0 (64.7-76.4)73.3 (67.1-78.5)
289.5 (88.4-90.5)67.8 (65.8-69.7)69.4 (66.4-72.1)26.1 (24.5-27.8)77.1 (70.9-82.1)65.3 (58.0-71.6)
390.5 (89.4-91.5)68.5 (66.4-70.5)70.9 (67.7-73.9)27.4 (25.6-29.3)80.5 (75.0-84.9)80.4 (74.3-85.2)
491.3 (90.2-92.3)70.4 (68.2-72.5)71.8 (68.5-74.8)27.3 (25.3-29.3)78.0 (72.0-82.9)75.3 (67.0-81.8)
592.1 (91.0-93.0)71.0 (68.6-73.2)71.8 (68.5-74.8)27.9 (25.7-30.1)81.0 (74.7-85.8)75.8 (68.0-81.9)
Pa<.001.002.096.317.118.011
All90.4 (89.9-90.8)69.1 (68.2-70.1)70.3 (68.9-71.6)26.8 (25.9-27.6)77.3 (74.8-79.7)73.8 (70.8-76.5)
New Cases Seen at RCC
SESBreastColonRectumNSCLCervixLarynx
190.1 (88.3-91.6)67.5 (63.8-71.0)65.4 (60.1-70.1)17.5 (15.2-19.9)69.5 (58.9-77.8)77.9 (69.3-84.3)
291.2 (89.7-92.4)67.8 (64.4-70.9)67.7 (62.8-72.0)16.7 (14.6-18.9)68.0 (58.2-75.9)71.3 (59.9-80.0)
393.4 (92.0-94.5)65.4 (61.5-69.0)69.5 (64.3-74.1)17.7 (15.3-20.1)74.7 (65.0-82.0)80.5 (71.5-86.9)
492.4 (90.8-93.7)68.2 (64.0-72.0)69.2 (63.3-74.4)18.2 (15.5-20.9)66.2 (56.0-74.6)72.1 (59.3-81.5)
594.3 (93.0-95.4)69.7 (65.0-73.9)71.1 (64.8-76.4)19.9 (16.8-23.1)79.5 (69.2-86.7)82.3 (69.5-90.0)
Pa<.001.335.422.606.200.487
All91.1 (90.6-91.7)65.6 (64.1-67.1)71.3 (69.4-73)21.6 (20.6-22.6)74.2 (70.9-77.1)72.2 (68.4-75.6)

Results of Cox analyses suggest that differences among social groups in stage of disease at diagnosis have a modest effect on survival disparity for breast cancer but not other disease sites. In Table 5, we show results of Cox analyses for 3 diseases: a disease with a significant survival gradient that appears to be partly explained by differences in stage distribution (breast cancer); a disease with a significant survival gradient that does not appear to be explained by differences in stage distribution (colon cancer); and a disease with no observed survival gradient (NSCLC). The association between SES and survival in breast cancer appears robust, as there is a clear step-wise gradient of effect across quintiles (HRs, 1.18, 1.27, 1.43, 1.77). The survival difference across social groups seen in univariate analysis is partly attenuated when stage is added to the model. As shown in the multivariate analysis, despite controlling for stage and age, the survival of patients from poorer communities remains significantly inferior to those from more affluent communities. Patients with breast cancer who reside in the poorest communities in Ontario have a 47% increased risk of death compared with patients of the highest quintile (HR, 1.47; 95% confidence interval [CI], 1.27-1.71; P < .001, adjusted analysis). The data for colon cancer also suggests a relation between SES and survival. However, unlike breast cancer, the HRs in multivariate analyses are not substantially different from univariate analyses, thus suggesting that age and stage are not responsible for observed differences in outcome.

Table 5. Univariate and Multivariate Analysis for Overall Survival, Socioeconomic Status, and Age Among Patients With Breast, Colon, and Nonsmall Cell Lung Cancers Seen at a Regional Cancer Center in Ontario 2003-2007
 Univariate AnalysisMultivariate Analysis
HR (95%CI)PHR (95%CI)P
Breast Cancer
 SES    
  11.77 (1.53-2.05)<0.0011.47 (1.27-1.71)<.001
  21.43 (1.24-1.65)<.0011.21 (1.05-1.40).010
  31.27 (1.09-1.47).0021.19 (1.03-1.38).020
  41.18 (1.01-1.37).0361.08 (0.93-1.25).340
  5ReferenceReference
 Stage    
  1ReferenceReference
  22.42 (2.11-2.78)<.0012.54 (2.21-2.91)<.001
  35.99 (5.15-6.97)<.0016.56 (5.64-7.63)<.001
  435.0 (30.2-40.5)<.00133.3 (28.7-38.6)<.001
  Unstaged3.31 (2.81-3.89)<.0013.41 (2.89-4.01)<.001
 Age, y1.04 (1.04-1.05)<.0011.04 (1.04-1.04)<.001
Colon Cancer
 SES    
  11.32 (1.16-1.50)<.0011.36 (1.20-1.55)<.001
  21.16 (1.03-1.32)0.0181.20 (1.06-1.36).004
  31.30 (1.15-1.48)<.0011.25 (1.10-1.42)<.001
  41.16 (1.01-1.32)0.0331.16 (1.02-1.33).027
  5ReferenceReference
 Stage    
  1ReferenceReference
  21.60 (1.20-2.13).0011.56 (1.17-2.07).002
  33.29 (2.51-4.32)<.0013.28 (2.50-4.30)<.001
  417.5 (13.4-23.0)<.00118.4 (14.0-24.0)<.001
  Unstaged5.26 (3.99-6.93)<.0015.43 (4.11-7.15)<.001
 Age, y1.02 (1.01-1.02)<.0011.02 (1.02-1.03)<.001
NSCLC
 SES    
  11.04 (0.98-1.12).2101.09 (1.02-1.16).014
  21.04 (0.97-1.11).2961.05 (0.98-1.12).160
  31.01 (0.95-1.08).7721.03 (0.96-1.10).378
  41.00 (0.93-1.07).8850.98 (0.91-1.05).510
  5ReferenceReference
 Stage    
  1ReferenceReference
  21.53 (1.35-1.74)<.0011.57 (1.39-1.78)<.001
  32.77 (2.53-3.02)<.0012.85 (2.61-3.11)<.001
  45.30 (4.87-5.78)<.0015.56 (5.10-6.06)<.001
  Unstaged3.69 (3.36-4.04)<.0013.79 (3.45-4.15)<.001
 Age, y1.01 (1.01-1.01)<.0011.01 (1.01-1.01)<.001

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Several important findings have emerged from this contemporary population-based study of SES, stage of cancer at diagnosis, and survival. First, among patients seen at Ontario's Regional Cancer Centers, we have observed a stage capture rate of 85%. Our study includes stage data for 38,431 patients, which is the largest such study reported to date outside of the United States. Second, we have found only very modest differences in stage at time of diagnosis across social groups. Third, consistent with work by our group in the 1990s,2, 3 we have found that important differences in survival across socioeconomic quintiles persist in Ontario in the 21st century. Finally, we have found that stage of cancer at time of diagnosis accounts for only a modest portion of the gradient in survival among patients with breast cancer.

The existing literature has consistently found socioeconomic differences in survival and incidence for multiple cancers. These variations have been observed across studies despite differing methods for assigning SES and are greatest for cancers of the breast, colon, bladder, and uterus.1, 11 Our own data for breast and colon cancer are consistent with this observation. More than 1 decade ago, we described large and significant differences in survival for cancers of the head and neck, cervix, uterus, breast, prostate, and bladder within Ontario.2 Smaller, but significant, associations were seen in cancers of the lung and rectum. No significant association between community income and survival was observed in cancers of the stomach, colon, pancreas, and ovary. Our current results are consistent with these findings with the exception of the large disparity in survival now observed for colon cancer. The initial study included patients diagnosed during the years 1982-1991. Since that time, management of colon cancer has changed substantially with the widespread adoption of adjuvant chemotherapy.20 It is possible that the differences in survival we are now observing for this disease relate to access to and quality of care during systemic therapy.

The lack of a strong association between SES and stage of cancer at diagnosis contrasts with much of the existing literature. Two recently published reports from the United States confirm differences in outcome across social groups, and both studies found that patients of lower socioeconomic status were more likely to have advanced cancer at the time of diagnosis. In the National Program of Cancer Registries Patterns of Care Study, Byers et al4 evaluated stage of disease, treatment, and outcomes of 13,958 patients with breast, colorectal, and prostate cancer and found that low SES was a risk factor for all-cause mortality, largely because of later stage at diagnosis and less aggressive treatment. Similar findings were reported by Clegg et al21 in their analysis of cancer outcomes and SES. In their systematic review of literature published during 1995-2005, Woods et al found that most studies report an association between SES and stage at diagnosis.11 Three studies were identified that did not find an association between SES and stage; each of these studies were conducted in the United Kingdom where patients would have been included in universal healthcare programs.13-15 The recent study by Byers et al4 and other reports22-24 have found that SES-related disparities in cancer in the United States are much less apparent among populations whose healthcare is provided through universal coverage including Medicare, health maintenance organizations, and in the Department of Defense and Veterans Administration healthcare systems. It is plausible that the lack of a strong association in our current study between SES and stage of cancer at diagnosis relates to the finding that Ontario has universal health coverage, which may facilitate access to cancer screening.

Having observed important socioeconomic differences in survival in Ontario that are not explained by differences in stage of cancer at diagnosis raises the fundamental issue as to what factors are driving the observed disparities in outcome. Differing outcomes may relate to factors at the disease, patient, and/or health-system levels.25 Existing evidence suggests that each of these domains may contribute to socioeconomic disparities in cancer outcomes. Several studies have found that patients from lower socioeconomic and minority groups may develop cancer with more aggressive biology.26-29 At the patient-level, comorbidity and smoking30-32 vary across socioeconomic groups and may influence patient outcomes. Although nutrition and psychosocial supports are known to vary across socioeconomic groups, there is less evidence to determine the extent to which these variables account for disparities in cancer survival.10, 11 Finally, studies from several countries (including Canada) have demonstrated substantial differences in treatment provided to patients with cancer who are from different socioeconomic groups.4, 11, 33, 34

Although the current study is among the largest studies to evaluate the association between SES, stage of disease, and survival, several limitations merit comment. We have used an ecologic measure of SES as individual-level data were not available. Although ecologic SES measures have been validated as proxies for individual values, the literature suggests that aggregate measures may be less sensitive than individual measures.35-37 Because stage data are routinely captured in Ontario only for patients seen at the province's comprehensive Regional Cancer Centers, our study does not include all patients in Ontario. Although patients seen at RCCs tend to be younger than patients not seen at RCCs, the proportion of patients coming from the poorest communities seen at RCCs and non-RCC institutions was not different (data not shown). However, the observed differences in survival between cases seen at RCCs and all cases in the province suggest that there may be important differences in the case mix and/or intent of treatment. Finally, the quality of stage data as recorded by treating physicians is also unknown, although the substantial agreement between stage groupings derived by TNM variables and the user-determined summary stage suggest very good internal validity (kappa 0.97, data not shown).

In summary, in this large population-based study, we have observed only modest differences in stage of cancer at time of diagnosis across socioeconomic groups. We have also found that large and significant disparities in survival exist in Ontario despite universal healthcare. Whereas stage of disease at diagnosis may account for a small portion of the survival gradient among patients with breast cancer, other factors, including differences in disease biology, comorbidity, access to therapy, and quality of care, may have a greater impact on survival. Further work is needed to better understand such factors and to develop strategies that reduce disparities in the outcome of patients with cancer.

CONFLICT OF INTEREST DISCLOSURES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES

Dr. Booth is supported as a Cancer Care Ontario Research Chair in Health Services.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. CONFLICT OF INTEREST DISCLOSURES
  7. REFERENCES
  • 1
    Kogevinas M, Porta M. Socioeconomic differences in cancer survival: a review of the evidence. In: Kogevinas M, Pearce M, Susser M, Boffetta P, eds. Social Inequalities and Cancer. Lyon, France: International Agency for research on Cancer. IARC Scientific Publications; 1997; 138: 177-206.
  • 2
    Mackillop WJ, Zhang-Salomons J, Groome PA, Paszat L, Holowaty E. Socioeconomic status and cancer survival in Ontario. J Clin Oncol. 1997; 15: 1680-1689.
  • 3
    Boyd C, Zhang-Salomons JY, Groome PA, Mackillop WJ. Associations between community income and cancer survival in Ontario, Canada, and the United States. J Clin Oncol. 1999; 17: 2244-2255.
  • 4
    Byers TE, Wolf HJ, Bauer KR, et al. The impact of socioeconomic status on survival after cancer in the United States: findings from the National Program of Cancer Registries Patterns of Care Study. Cancer. 2008; 113: 582-591.
  • 5
    Parkin DM, Muir CS. Cancer Incidence in Five Continents. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications; 1992.
  • 6
    Williams J, Clifford C, Hopper J, Giles G. Socioeconomic status and cancer mortality and incidence in Melbourne. Eur J Cancer. 1991; 27: 917-921.
  • 7
    Faggiano F, Zanetti R, Costa G. Cancer risk and social inequalities in Italy. J Epidemiol Community Health. 1994; 48: 447-452.
  • 8
    Van Loon AJ, Goldbohm RA, van den Brandt PA. Socioeconomic status and breast cancer incidence: a prospective cohort study. Int J Epidemiol. 1994; 23: 899-905.
  • 9
    Mackillop WJ, Zhang-Salomons J, Boyd CJ, Groome PA. Associations between community income and cancer incidence in Canada and the United States. Cancer. 2000; 89: 901-912.
  • 10
    Groome PA, Schulze KM, Keller S, et al. Explaining socioeconomic status effects in laryngeal cancer. Clin Oncol (R Coll Radiol). 2006; 18: 283-292.
  • 11
    Woods LM, Rachet B, Coleman MP. Origins of socio-economic inequalities in cancer survival: a review. Ann Oncol. 2006; 17: 5-19.
  • 12
    Clegg LX, Reichman ME, Miller BA, et al. Impact of socioeconomic status on cancer incidence and stage at diagnosis: selected findings from the surveillance, epidemiology, and end results: National Longitudinal Mortality Study. Cancer Causes Control. 2009; 20: 417-435.
  • 13
    Wrigley H, Roderick P, George S, Smith J, Mullee M, Goddard J. Inequalities in survival from colorectal cancer: a comparison of the impact of deprivation, treatment, and host factors on observed and cause specific survival. J Epidemiol Community Health. 2003; 57: 301-309.
  • 14
    Thomson CS, Hole DJ, Twelves CJ, Brewster DH, Black RJ. Prognostic factors in women with breast cancer: distribution by socioeconomic status and effect on differences in survival. J Epidemiol Community Health. 2001; 55: 308-315.
  • 15
    Brewster DH, Thomson CS, Hole DJ, Black RJ, Stroner PL, Gillis CR. Relation between socioeconomic status and tumour stage in patients with breast, colorectal, ovarian, and lung cancer: results from 4 national, population based studies. BMJ. 2001; 322: 830-831.
  • 16
    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.
  • 17
    Clarke EA, Marrett LD, Krieger N. Cancer registration in Ontario: a computer approach. In: Jenson OM, Parkin DM, MacLennan R, eds. Cancer Registration Principles and Methods. Lyon, France: IARC; 1991: 246-257.
  • 18
    Statistics Canada. Census Geographic Units. http://www12.statcan.ca/english/census01/Products/Reference/dict/geo024.htm. Published in 2001. Accessed on December 8, 2009.
  • 19
    Statistics Canada. Postal Code Conversion File. http://www.statcan.gc.ca/bsolc/olc-cel/olc-cel?catno=92F0153XCE&lang=eng. Published in 2002. Accessed on December 8, 2009.
  • 20
    Meyerhardt JA, Mayer RJ. Systemic therapy for colorectal cancer. N Engl J Med. 2005; 352: 476-487.
  • 21
    Clegg LX, Reichman ME, Miller BA, et al. Impact of socioeconomic status on cancer incidence and stage at diagnosis: selected findings from the surveillance, epidemiology, and end results: National Longitudinal Mortality Study. Cancer Causes Control. 2009; 20: 417-435.
  • 22
    Page WF, Kuntz AJ. Racial and socioeconomic factors in cancer survival. A comparison of Veterans Administration results with selected studies. Cancer. 1980; 45: 1029-1040.
  • 23
    Rabeneck L, Souchek J, El-Serag HB. Survival of colorectal cancer patients hospitalized in the Veterans Affairs Health Care System. Am J Gastroenterol. 2003; 98: 1186-1192.
    Direct Link:
  • 24
    Dominitz JA, Samsa GP, Landsman P, Provenzale D. Race, treatment, and survival among colorectal carcinoma patients in an equal-access medical system. Cancer. 1998; 82: 2312-2320.
  • 25
    Mackillop WJ. Health Services research in radiation oncology: toward achieving the achievable. In: Gunderson LL, Tepper JE, eds. Clinical Radiation Oncology. Philadelphia: Churchill Livingstone; 2007: 215-237.
  • 26
    Schrijvers CT, Coebergh JW, van der Heijden LH, Mackenbach JP. Socioeconomic variation in cancer survival in the southeastern Netherlands, 1980-1989. Cancer. 1995; 75: 2946-2953.
  • 27
    Lai H, Lai S, Ma F, Meng L, Trapido E. Prevalence and spectrum of p53 mutations in white Hispanic and non-Hispanic women with breast cancer. Breast Cancer Res. Treat. 2003; 81: 53-60.
  • 28
    Chen VW, Fenoglio-Preiser CM, Wu XC, et al. Aggressiveness of colon carcinoma in blacks and whites. National Cancer Institute Black/White Cancer Survival Study Group. Cancer Epidemiol Biomarkers Prev. 1997; 6: 1087-1093.
  • 29
    Gapstur SM, Dupuis J, Gann P, Collila S, Winchester DP. Hormone receptor status of breast tumors in black, Hispanic, and non-Hispanic white women. An analysis of 13,239 cases. Cancer. 1996; 77: 1465-1471.
  • 30
    Berg JW, Ross R, Latourette HB. Economic status and survival of cancer patients. Cancer. 1977; 39: 467-477.
  • 31
    Polednak AP. Poverty, comorbidity, and survival of colorectal cancer patients diagnosed in Connecticut. J Health Care Poor Underserved. 2001; 12: 302-310.
  • 32
    Prout GR Jr, Wesley MN, McCarron PG, et al. Survival experience of black patients and white patients with bladder carcinoma. Cancer. 2004; 100: 621-630.
  • 33
    Huang J, Zhou S, Groome P, Tyldesley S, Zhang-Solomans J, Mackillop WJ. Factors affecting the use of palliative radiotherapy in Ontario. J Clin Oncol. 2001; 19: 137-144.
  • 34
    Paszat LF, Mackillop WJ, Groome PA, Zhang-Salomons J, Schulze K, Holowaty E. Radiotherapy for breast cancer in Ontario: rate variation associated with region, age and income. Clin Invest Med. 1998; 21: 125-134.
  • 35
    Mustard CA, Derksen S, Berthelot JM, et al. Assessing ecologic proxies for household income: a comparison of household and neighbourhood level income measures in the study of population health status. Health Place. 1999; 5: 157-171.
  • 36
    Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health. 1992; 82: 703-710.
  • 37
    Krieger N, Williams DR, Moss NE. Measuring social class in US public health research: concepts, methodologies, and guidelines. Annu Rev Public Health. 1997; 18: 341-378.