Social Epidemiologic Methods in International Population Health
1. Social Epidemiologic MethodsSocial Epidemiologic Methods
in International Populationin International Population
Health and Health ServicesHealth and Health Services
ResearchResearch
A Research Agenda Using Cancer
Care as a Sentinel Indicator:
By Kevin M. Gorey
2. Kevin M. GoreyKevin M. Gorey
Kevin is a social epidemiologist and
social welfare researcher interested
in advancing understandings about
how health care policies affect
health. He is particularly interested
in the impacts of various under-
and uninsured statuses in the US.
His web page is:
www.uwindsor.ca/gorey
3. Cancer Survival in CanadianCancer Survival in Canadian
and United Statesand United States
Metropolitan Areas: A SeriesMetropolitan Areas: A Series
of Studiesof Studies
Between-Country Effect
Modification by Socioeconomic
Status
(Health Insurance)
4. Research Team and ReportsResearch Team and Reports
Kevin Gorey, University of Windsor
Eric Holowaty & Gordon Fehringer, CCO
Erich Kliewer, Cancer Care Manitoba
Ethan Laukkanen, WRCC and Colleagues
Study series reports:
Am J Public Health 1997 & 2000
Can J Public Health 1998; Milbank Q 1999
J Public Health Med 2000
J Health Care Poor Underserved 2003
Ann Epidemiol 2003
6. Historical ContextHistorical Context
- Canada: Universal single payer
- US: Multi-tiered—uninsured and
underinsured, Medicaid, Medicare,
continuum of private coverages
- Time of great systemic changes
- Managed care proliferation (US)
- Federal-provincial shift (Canada)
7. Politics Versus SciencePolitics Versus Science
- Political debates tend to mythologize
anecdotal outcomes.
- Rhetoric often not substantiated
(e.g., 2 Manitoba studies)
- Waits for 10 surgical procedures stable
or decreased 5 yrs post-downsizing
- Access to surgery actually increased
after hospital downsizing (maintaining
quality [mortality, readmissions])
8. Cancer Survival is a SentinelCancer Survival is a Sentinel
Health Care OutcomeHealth Care Outcome
- Relatively common over the life course
- Diverse constellation of diseases
- Many with good prognoses and high
quality of survivable life
- Diverse screens (including primary care)
and treatments exist and matter
- Timely access, referral and follow-up
matter
9. Theoretical Context:Theoretical Context:
Systematic Literature ReviewSystematic Literature Review
- In the US, ethnicity and SES are strongly
associated with health insurance
statuses (odds ratios [OR] 2.0 to 15.0).
- All are also strongly associated with
cancer screens, stages at diagnosis and
access to treatments (ORs 2.0 to 5.0).
- Such Canadian associations tend to be
attenuated or nonexistent. For example:
- US SES-cancer survival OR = 1.56
- Canadian OR = 1.04 (NS) to 1.18
10. SES: A Key Effect Modifier?SES: A Key Effect Modifier?
Therefore, any Canada-US cancer outcome
study that does not incorporate SES is
unlikely to observe the truth.
- SES is so intimately connected with
health in North America that it must be
incorporated into all such studies.
- If an interaction exists, interpretations of
main effects alone can be misleading.
11. SES: An Effect Modifier? E.G.SES: An Effect Modifier? E.G.
- One previous study of Canada-US
cancer survival (GAO, 1994)
- Found no between-country differences
- But, did not account for SES
- We have observed a substantially
different picture within SES strata.
- Consistent Canadian advantages
within the lowest SES strata
12. A Country By SES InteractionA Country By SES Interaction
Hypothesis Guided Our SeriesHypothesis Guided Our Series
Relatively poor Canadian cancer patients
(better insured) would enjoy advantaged
survival over their similarly poor
counterparts in the United States.
- We think this a better guide to policy-
interesting and important research
questions in North America than those
provided by main effect country-based
hypotheses.
13. MethodsMethods
A Focused Series of Cancer
Survival Comparisons Among
Relatively Poor Residents of
Canadian and American
Metropolitan Areas
14. Comparative Series OverviewComparative Series Overview
Toronto, Ontario vs Detroit, Michigan
An ecological exemplar
Toronto vs San Francisco, Seattle, Hartford
Adjustment for absolute income
Toronto vs Honolulu, HI
Health insurance hypothesis test
Winnipeg, Manitoba vs Des Moines, Iowa
Replicate among smaller cities
Comparisons of Subsamples < 65 yoa
Health insurance hypothesis test
15. Sampling—Persons/CancerSampling—Persons/Cancer
PatientsPatients
- Ontario and Manitoba Registries, SEER
- First, primary invasive cancer cases
- MC, not DC or autopsy only
- With minimum 5 years follow-up
- Began 15 most common cancers
- Since focused on most significant
- Estimated case ascertainments, MC, and
follow-ups all > 95% (DCO/Autopsy < 1%)
- Even better among the most public
health-significant cancer types
17. Sampling—Places: RationalesSampling—Places: Rationales
For Metropolitan SamplingFor Metropolitan Sampling
- Maximize internal validity
- Higher: MC, follow-up, geocoding rates
- Lower: DCO or autopsy only
- Maximize external validity
- Vast majority of NAs urban residents
- 1 of 3 Ontarians and 1 of 7 Canadians
reside in Toronto
- Control for service availability
18. Sampling—Places: EcologicalSampling—Places: Ecological
Measures of SESMeasures of SES
NeighborhoodsNeighborhoods
No NA registries coded personal SES.
- Census tracts joined cases at diagnosis
to income data (US Census, Stats Can)
- Neighborhood prevalence poor
- Theory, insurance, practical sig.
- Poverty (US), low income (Canada)
- Both household income-based and tied
to the consumer price index
- Though Canadian criterion more liberal
- Used to form relative SES quantiles
19. Comparison of SES Quintiles: 1990/91, US$Comparison of SES Quintiles: 1990/91, US$
WinnipegWinnipeg Des MoinesDes Moines
SESSES Mdn $Mdn $ Mdn $Mdn $
High $47,090 $44,050
39,110 36,370
32,265 30,165
26,043 26,890
Low 17,500 19,570
Lowest US SES quintile: 20% poor, another
45% near poor; estimated (vs highest)
uninsured PR = 10.0, underinsured PR = 15.0
21. SRRs With 95% CIs, 1984 to 1994SRRs With 95% CIs, 1984 to 1994
SESSES TorontoToronto DetroitDetroit
High 1.00 … 1.00 …
1.00 (0.94,1.06) 0.94 (0.88,1.01)
Low 0.98 (0.93,1.04) 0.80 (0.75,0.85)
No significant between-country differences
in the middle or high income areas
Low income areas: Between-country
SRR = 1.30 (1.23,1.38), Canadian patients
advantaged
23. Toronto-Honolulu Between-Toronto-Honolulu Between-
Country Survival OutcomesCountry Survival Outcomes
The only significant decile difference was for
the lowest income area:
SRR = 1.20 (1.06, 1.36)
Canadian patients advantaged
Among those < 65 yoa:
SRR = 1.28 (1.07,1.53)
25. Summary: Health InsuranceSummary: Health Insurance
- Consistent SES-cancer survival
associations in US, but not Canada
- Consistent country-SES interactions
- Canada advantage lowest SES strata
- Particularly among those < 65 yoa
- Consistency of pattern across diverse
contexts—people and places—points
toward a pervasive systemic effect
- 285 of 319 between-country
comparisons were in support of the
health insurance hypothesis
26. AltAlt11—Income Gap or Inequality—Income Gap or Inequality
Larger in the United States?Larger in the United States?
- For some of our studies, the
economic divide is actually larger
in the Canadian sample.
- E.g., Winnipeg vs Des Moines
27. AltAlt22—Ethnic or Cultural—Ethnic or Cultural
Explanations?Explanations?
- Similar pattern of findings observed
among various ethnic mixes
- North American studies of race/ethnicity
and cancer screening have implicated
knowledge (education), rather than race,
per se.
- Consistent indictment of America:
Inequitable distribution of key social
resources—education and health care
28. AltAlt33—Lifestyle Factors (LS): Exercise, Diet,—Lifestyle Factors (LS): Exercise, Diet,
BMI, Tobacco and Alcohol Consumption?BMI, Tobacco and Alcohol Consumption?
- Associations with cancer survival tend
to be extremely small
- Larger associations with incidence
- Survival findings consistent across
cancers with diverse component causes
- Some LS factors very sig., others not
- Income is associated with lifestyle in
both countries, but no income-survival
gradients were observed in Canada
- Little to no Canada-US LS prevalence
differences (2%) have been observed
29. AltAlt44—Different Case Mixes by—Different Case Mixes by
Stage of Disease at Diagnosis?Stage of Disease at Diagnosis?
- Stage differences may account for some,
but probably not all of the between-
country survival differences.
- In within-US stage-adjusted analyses,
treatment differences still account for
roughly 50% of survival variabilities.
30. AltAlt55—Cancer Registry Death Clearance?—Cancer Registry Death Clearance?
National (US) vs Provincial (Canada)National (US) vs Provincial (Canada)
- Over the life of these studied cohorts,
only 1-3% of Toronto residents moved
out-of-province.
- Likely fewer chronically ill moved
- Ontario Cancer Registry comparisons of
national and provincial death clearances
found inconsequential differences.
31. AltAlt66—Competing Causes of Death—Competing Causes of Death
(Observed vs Relative Survival)?(Observed vs Relative Survival)?
- Life expectancy in Honolulu among
both women and men is close to 3
years greater than in Toronto
- Therefore, our between-country
SRRs (Canadian advantage) may
actually underestimate the truth
32. AltAlt77—Lead Time Bias?—Lead Time Bias?
- Our findings were fairly consistent
across different cancers probably with
various pre-clinical phase lengths.
- A systematic review of 87 studies (with
adjustment for lead-time) observed stage
and treatment effects (Richards et al.,
1999, Lancet)
33. AltAlt88—Ecological Fallacy?—Ecological Fallacy?
- Even if it were merely an area effect, the
consistently observed residence-survival
association in the US, but not in Canada
would still be instructive.
- The compositional measure (% poor and
near poor in neighborhoods) is well
known to be intimately associated with
under-and uninsured statuses in the US.
34. Future Research NeedsFuture Research Needs
Health Insurance Hypothesis
Developed and Screened With An
Ecological—Income—Proxy:
More Definitive Testing Needed
35. Central Research NeedsCentral Research Needs
- Study more recent retrospective
and prospective cohorts
- Perform stage-stratified analyses
- Incorporate treatment variables
- Extend generalizability to smaller
urban and rural-remote places
- Develop construct validity of
ecological SES measures in
Canada
36. Our Research Agenda OverOur Research Agenda Over
The Next 5 YearsThe Next 5 Years
Endeavoring to Filling Some of This
Field’s Central Knowledge Gaps
37. Social, Prognostic & TherapeuticSocial, Prognostic & Therapeutic
Factors Associated With CancerFactors Associated With Cancer
Survival in Canada and the USSurvival in Canada and the US
Health Care Access and
Effectiveness in Diverse Urban
and Rural Contexts, 1985 to 2010
38. Research Team Co-InvestigatorsResearch Team Co-Investigators
Kevin Gorey (PI) & Emma Bartfay
(Epidemiology)
Karen Fung (Biostatistics)
Isaac Luginaah (Geography)
Frances Wright (Surgical Oncology)
Caroline Hamm & Sindu Kanjeekal
(Medical Oncology)
Eric Holowaty & William Wright
(Cancer Surveillance & Registration)
39. To Address Identified ResearchTo Address Identified Research
Needs, It Will:Needs, It Will:
- Study more recent retrospective and
prospective cohorts
- Perform stage-stratified analyses
- Incorporate treatment variables
- Extend generalizability to smaller urban
and rural-remote places
- Develop construct & predictive validities
of ecological SES measures in Canada
40. Cohort DesignCohort Design
Incident cohorts: 1985-1990 & 1995-2000
Followed until: 2000 2010
Cox models over 1-, 3-, 5- to 10-years
In Canada and the US
During a policy-interesting period
- Federal-provincial shift in Canada
- For-profit managed care proliferation &
prevalent increases uninsured in US
41. Staged AnalysesStaged Analyses
No Canadian cancer registry routinely codes stage
of disease at diagnosis.
- Thus, no previous study in this field has been
able to account for case-mix.
Stage will be abstracted for this study’s samples.
Allowing for:
- More comparable between-country
comparisons
- Examination of the relative weightiness of
pre- (affect later diagnosis) and post-diagnostic
(affect lack of access to best treatments and
follow-up) social forces
42. Incorporation of TreatmentsIncorporation of Treatments
No Canadian cancer registry routinely codes initial
treatments.
- Thus, no previous study in this field has been
able to account for them in survival analyses.
Detailed treatment variables will be abstracted for this
study’s samples.
- Surgery, radiation, chemotherapy and others
- Initial course and follow-up
- Type, dose, delays, timings/sequence
between various therapies
43. Extending Generalizability:Extending Generalizability:
Contexualizing KnowledgeContexualizing Knowledge
Systematic Replications in:
Ontario California
Large cities Toronto San Fran/Oakland
Small cities Windsor Salinas
Rural/remote areas of Ontario & California
1,060 breast and colon cancer cases for each
incident cohort in each type of place
44. Ecological Measurement ValidityEcological Measurement Validity
Ontarian and Californian cancer cases will be
joined via their residential census tracts to
the following data:
- Income (poverty prevalence) and
- Physician supplies (count/10,000 pop)
- Primary care and specialists
This will provide opportunities to better
understand the meanings of such ecological
measures, particularly in Canada, where little
is yet known about them.
45. Hypotheses Related to SurvivalHypotheses Related to Survival
1. Significant country by SES interaction
(Canadian advantage low-income only)
1a. Advantage significantly increased over time
2. SES-survival significant in US (not in Canada)
2a. Age by SES interaction (Medicare advantage)
2b. US gradient significantly increased over time
3. Physician supplies-survival associations
significant in both Canada & US (for both
primary care and specialists supplies)
Editor's Notes
First (Setting up the Historical and Theoretical Contexts), Slides 3-35:
1) Review of a research agenda on international (Canada and the US) comparative cancer care, and
2) Development of a health insurance hypothesis (explanation for observed Canadian survival advantages)
Second (Presentation of Original Research Proposals), Slides 36-45:
“Social, prognostic and therapeutic factors associated with breast cancer survival in Canada and the US: Health care access and effectiveness in diverse urban and rural places, 1985 to 2010” Canadian Institutes of Health Research/Canadian Breast Cancer Research Alliance (2004-2009)
“Social, prognostic and therapeutic factors associated with colon cancer survival in Canada and the US: Health care access and effectiveness in diverse urban and rural places, 1995 to 2008” National Cancer Institute of Canada (submitted)
Note. CCO = Cancer Care Ontario, WRCC = Windsor Regional Cancer Centre
Selected References:
Gorey KM, Kliewer E, Holowaty EJ, et al. An international comparison of breast cancer survival: Winnipeg, Manitoba and Des Moines, Iowa metropolitan areas. Ann Epidemiol 2003; 13:32-41.
2)Gorey KM, Holowaty EJ, Fehringer G, et al. An international comparison of cancer survival: Metropolitan Toronto, Ontario and Honolulu, Hawaii. Am J Public Health 2000;90:1866-1872.
3)Gorey KM, Holowaty EJ, Fehringer G, et al. An international comparison of cancer survival: Relatively poor areas of Toronto, Ontario and three US metropolitan areas. J Public Health Med 2000;22:343-348.
4)Gorey KM, Holowaty EJ, Laukkanen E, et al. An international comparison of cancer survival: Advantage of Toronto’s poor over the near poor of Detroit. Can J Public Health 1998;89:102-104.
5)Gorey KM, Holowaty EJ, Fehringer G, et al. An international comparison of cancer survival: Toronto, Ontario and Detroit, Michigan metropolitan areas. Am J Public Health 1997;87:1156-1163.
Canada and the United States can make for very policy-interesting comparisons:
Similar (though not the same) on a number of scores: Developed nations in North America, environments, cultural and lifestyle factors
But
2)Very different in the ways that they finance and distribute health care resources
How do we want policy decisions to be steered in Canada (or the US)?
By the mere ebb and flow of political tides (interests of politically influential groups)?
Or
By systematic evidence derived from rational and empirical enquiry (science)?
Manitoban studies referred to in slide:
DeCoster et al., Med Care 1999
Brownell et al., Med Care 1999
Cancer care outcomes then, particularly among the most public health-significant cancers such as breast and colon cancer, can be thought of as sentinel indicators of a health care system’s effectiveness.
Based on the following systematic reviews (300+ study outcomes):
Parsons RR, Gorey KM, Anucha U, Nakhaie R. Institutionalized racism and classism in health care: Meta-analytic evidence of their existence in America, but not in Canada. Paper presented at the 132nd annual meeting of the American Public Health Association. Washington, DC, November, 2004 (accepted). [In preparation for journal submission.]
Gorey KM. Canada-United States comparative cancer care: Systematic review-generated hypotheses and methodological direction for future research. In V Goel (Chair), Canada-US comparisons of health services: Methodological issues and interpretations. Symposium conducted at the North American Congress of Epidemiology. Toronto. June, 2001.
At this point out research group began to wonder about the following phenomena:
What is the best theoretical/methodological fit of social position (broadly defined socioeconomic factors or life chances: education, occupational prestige, income, health insurance) within our planned international cancer outcome comparative studies?
And
Is it possible that previous general population comparisons in this field were misleading?
Given the great diversity of both Canadian and American populations (diverse people living in diverse places), we thought it likely that such gross, general population comparisons had missed many more specific, and potentially more policy-interesting analytic opportunities.
And
We thought it likely that any effect of country would be moderated by social position (e.g., by the ability to pay in the US).
The General Accounting Office study compared all cancer patients in the province of Ontario with a sample of the entire US population of such patients (4 types of cancer over a given period of time).
Such gross comparisons are akin to attempting to find a scientific/policy important needle in a hay stack.
In that they aggregate the outcomes of all patients (the relatively rich and poor, those who live in large metropolitan areas to rural-remote areas, etc.) they are not capable of observing any interesting/important effects among such subgroups.
The Detroit-Toronto comparison was deemed by some as “unfair.” We think it interesting though in demonstrating that one group of patients could be so disadvantaged (Detroit) relative to another (Toronto) living only 200 miles away—both groups in very large metropolitan areas that are generally well endowed with health care (cancer care) resources (services are available, if not accessible/affordable).
Honolulu was chosen for study because of its most “Canada-like” health care system among the states (least prevalent uninsured population). And as hypothesized, its cancer care outcomes were more similar to those observed among Canadian patients.
It was also hypothesized that younger patients in the US (&lt; 65, not Medicare eligible) would be more disadvantaged than their older counterparts. Such was consistently observed across contexts.
Note. SEER = Surveillance, Epidemiology and End Results (the National Cancer Institute’s [US] cancer surveillance program), MC = microscopic confirmation, and DC = death certificate.
We focused on the use of so-called “gold standard” cancer registries of known validity for the study of homogeneous disease entities.
We began by surveying the most common forms of cancer, and then moved to study of the most public health-significant forms of cancer: Relatively common, with relatively good prognoses, where timely diagnosis and access to the best available treatments matter (better survival with a high quality of life over much of that extended period of survival).
This slide demonstrates the very high registry validities across the metropolitan socioeconomic strata that we studied. Even in the extreme instance of socioeconomic deciles, microscopic confirmation rates were nearly perfect and death certificate only/autopsy case ascertainment rates were nil from the highest to the lowest strata.
Similar validities were observed in Toronto.
So registry differences did not likely confound between-country survival comparisons.
Note. NA = North America.
We based our initial analyses on the principle that concerns for internal validity ought to precede those for external validity. Without first gaining a confident understanding of the relationships between hypothesized independent and dependent variables, issues of where they may be generalizable to are probably moot.
Consequently we initially focused on metropolitan areas (that also provide some external validation in North America). Though we recognize that the ultimate development of coherent systematic evidence for sound policy decisions will require similar study in Canada and America’s diverse places.
In the construction of our central ecological measure of SES we focused on the prevalence of federally-defined poor people within census tracts for the following reasons.
Census tracts are the most comparable units used by both the US Census Bureau and Statistics Canada.
Much construct validation of such units has been accomplished, particularly by sociologists in the US (e.g., poverty areas; William Julius Wilson, The truly disadvantaged, 1987; Paul Jargowsky, Poverty and Place, 1997).
Much predictive validation of such units has been accomplished, particularly in the US (e.g., Krieger N et al. The Public Health Disparities Geocoding Project. Am J Epidemiol 2002).
4)Such income-based measures are known to be intimately associated with various under- and uninsured statuses in the US, and thus, are close proxies of them.
This slide demonstrates the between-country comparability (median household income in census tracts) across five socioeconomic strata that typically resulted from construction of an ecological SES variable with the previously described method.
It also tends to validate the notion that the prevalence of poor people is very high in the lowest income strata (two-thirds are poor or near poor [200% of federal criterion]) and that their risk of being under- or uninsured is much greater [10- to 15-fold prevalence ratios] than among those who live in the highest income strata.
Exemplary findings of this series of studies (more than 300 within- and between-country survival comparisons, 15 most common cancers) will be encapsulated by focusing on breast cancer.
Note. SRR = survival rate ratio.
The slide depicts the lack of a significant income gradient in Toronto and a significant one in Detroit (low/high income survival rate ratio of 0.80).
And, as hypothesized, a significant country by income interaction was observed (the effect of country was moderated by income level). A Canadian advantaged was observed in the lowest third of income areas (survival rate ratio = 1.30), but not in the other middle to high income areas.
And given the relative commonness of breast cancer such SRRs (reflecting a 20-30% risk ratio) may be thought of as very large in a population attributable risk sense.
Again, this slide depicts the lack of a significant income gradient in Toronto and a significant one in Honolulu.
Note that these are findings across more extreme income deciles. The Honolulu gradient across tertiles was similar, though as hypothesized, attenuated compared with the Detroit gradient..
Again, as hypothesized, a significant country by income interaction was observed. A Canadian advantaged was observed in the lowest tenth of income areas (survival rate ratio = 1.20), but not in the others.
And as hypothesized, the Canadian advantage seemed even greater among patients less than 65 years of age (not yet eligible for Medicare coverage in the US).
Now lets take a look at 8 alternative explanations that could be advanced for the consistently observed Canadian cancer survival advantage. Rationales for ruling them out are also presented (expanded upon in the discussion section of Gorey and Colleagues, Ann Epidemiol 2003).
For example, it could be advanced that the economic divide is, in fact, wider in the US than in Canada, and that it is some social-health phenomena of such relative inequality in the US, rather than any absolute effect of being poor/uninsured, that is responsible for the observed US survival disadvantage.
However, it was clearly observed that such is not always the case. The income distributions displayed in slide 17, for example, clearly show a wider economic divide in Winnipeg than in Des Moines. Yet, cancer patients in Winnipeg remained advantaged.
A very conservative comparison for example, found relatively poor white cancer patients from the 3-county metropolitan Detroit area to be as disadvantaged relative to similarly poor patients in greater metropolitan Toronto (multi-ethnic sample, race/ethnicity not coded in the Ontario Cancer Registry).
Relative survival analyses are generally not possible in this field of study because life tables are not available within their specific socioeconomic contexts (e.g., for those who live in Canada or America’s poorest neighborhoods).
However, the results of both all-cause and cancer-specific (cases who died from competing causes of death censored) survival models tended to be close systematic replicates.
A study on breast cancer is already funded: Canadian Institutes of Health Research/Canadian Breast Cancer Research Alliance, 2004-9, $374,000.
A study on colon cancer is presently under review: National Cancer Institute of Canada, $293,050.
Breast and colon cancer have been selected for their public health significance (commonness, preventive potential: early diagnosis and access to the best treatments with effective follow-up matter), and for their previously observed social gradients in some international contexts. Also, increasingly effective treatments were introduced over this study’s generational time frame for all stages of both breast and colon cancer. They therefore have scientific and clinical power to expose any within- or between-country social gradients, if they exist. They also offer an interesting contrast, in that breast cancer screens are already common practice, while such screens for colon cancer will probably become much more common practice over the next 5 years or so. And high quality primary care may be thought of as a screen of sorts for both breast and colon cancer.
Institutional affiliations include: University of Windsor, University of Western Ontario, Sunnybrook and Women’s College Health Sciences Centre, Windsor Regional Cancer Centre, Cancer Care Ontario (Ontario Cancer Registry), and the California Department of Health Services (California Cancer Registry).
Studies are powered to detect a relatively small, but potentially public health- significant survival rate difference of 10%, with up to four covariates and an interaction term in models (power of .80 [1 – ] and of .05).
Total samples required are: 3,180 each breast and colon cancer cases for each Canadian and American incident cohort.
Because the US (California ) already codes stage of disease at diagnosis with SEER’s (Surveillance, Epidemiology and End Results program) extent of disease (EOD) coding scheme, to maximize between-country comparability on disease/stage definitions it will be used to stage the samples of Canadian cancer cases.
We will also be able to explore analyses with other staging systems: TNM (tumor-nodes-metastases) for both breast and colon cancer, and Dukes stages and other subsite analyses (e.g., proximal vs. distal) for colon cancer.
Again, to maximize between-country comparability on data collection/validity, SEER codes will be used in the abstraction of investigations and treatments.
Large cites have populations greater than 4 million.
Smaller cities and their surrounding county environs have populations between 300,000 and 400,000.
Rural places are communities of less than 10,000 residents with population densities of less than 150 per km2. Remote places are typically removed from urban centres by 200 miles or more (Health Canada’s Rural Health Initiative definitions).
Ontarian and Californian sampling frames were selected for the high quality of their cancer registry databases as well as for their representation of diverse Canadian and American places.
Socioeconomic data (census tract profiles) is available from Statistics Canada and the US Bureau of the Census.
Physician supply data (census tract profiles) is available from the Canadian Institute for Health Information (CIHI) and the American Medical Association (AMA).
Related to the second point:
In contrast to our group’s findings, another group of researchers has observed small to moderate SES-cancer survival gradients in Ontario (Kingston, ON: Mackillop et al., J Clin Oncol 1997; Boyd et al., J Clin Oncol 1999). We think that their group has posed questions very different than ours (different SES variables used). Area units in our metropolitan analyses have tended to be very small poverty enclaves of around .25 km2. Whereas, in their province wide analyses, the units were often times much larger (particularly in far outlying rural areas), sometimes more than 4,000 times larger (e.g., ecological units larger than 1,000 km2).
We hypothesize that the smaller units are better proxies for personal socioeconomic resources (and more predictive in American contexts), while the larger units are better proxies for community resources or regional health care service endowments (as indexed by physician supplies [and more predictive in Canadian contexts, particularly in other than large metro areas]).
Hypotheses apply to female breast cancer as well as to colon cancer among both women and men.
Similar hypotheses relevant to the specific prediction of stage of disease at diagnosis and the receipt of specific treatments have also been advanced (not shown here for simplicity sake).