the societal and economic impact of cancer health disparities · an g eir ge lopm te cancer a...

The Societal and Economic Impact of

Cancer Health Disparities

1776 I STREET NW 9TH FLOOR, WASHINGTON, DC 20006www.c-changetogether.org

ACKNOWLEDGEMENT C-Change would like to acknowledge the support and guidance its members and partners who serve on the Implementing a National Health Disparities Strategy Advisory Committee and the efforts by staff who all contributed to the development and preparation of this document. ADVISORY COMMITTEE MEMBERS Armin D. Weinberg, PhD, Co-Chair …………………………………..…………………..………..…Life Beyond Cancer Foundation Harold P. Freeman, MD, Co-Chair……………………………………..Ralph Lauren Center for Cancer Care and Prevention Ms. Carolyn Angeleri RiccI…………………………………………………………………………………..……………………Avon Foundation John E. Arradondo, MD, MPH ………………………………………………………..………….Volunteer State Medical Association Joel Beetsch, PhD …………………………………………………………………………………………………………………………Sanofi-Aventis Clem Bezold, PhD ……………………………………………………………..………………………………Institute for Alternative Futures Linda Blount, MPh …………………………………………………………………………………………..……………………………….WFG Equity Sharon Brigner, MS, RN …………………………….…………………………………………………….……………………………………..PhRMA Jimmy Boyd ………………………..……………………………………………………………………….............………Men's Health Network Ahmed Calvo, MD, MPH……………………….........................................Health Resources and Services Administration Erica Childs-Warner ………………………………………………………………..……………………………….Prevent Cancer Foundation Franz Fanuka ……………………………………………………..………National Forum for Heart Disease and Stroke Prevention S. Orlene Grant, RN, BSN, MSN …………………………………………………………………………………………The Grant Group, LLC Deborah Guadalupe Duran, PhD ………………………………………………..……………………………….National Cancer Institute Nikki S. Hayes, MPH ……………………………………………………………..………..Centers for Disease Control and Prevention Cheryl G. Healton, DrPh, MPA ………………………………………………………………….…LEGACY | for Longer Healthier Lives Nina Hill, PhD ………………………………………………………………………………………………….………………Pfizer Pharmaceuticals Marc Hurlbert, PhD ……………………………………………………………………………………..………………………….Avon Foundation Rebecca Johnson ……………………………….…………………….….National Association of City and County Health Officials Lovell Jones, PhD …………………………………………………….……...……..University of Texas MD Anderson Cancer Center Judith Salmon Kaur, MD ……………………………………………………………….…Mayo Clinic Comprehensive Cancer Center Gerard T. Kennealey, MD …………………………………………………………….……………………………………………….Cephalon Inc. Maureen Y. Lichtveld, MD, MPH …………………….Tulane University School of Public Health and Tropical Medicine Kim Linthicum ……………………………………………………………………………………………........................Myriad Genetics, Inc. Mr. Andy Miller, MHSE, CHES …………………………………………………………………………………………………………LIVESTRONG Neal A. Palafox, MD, MPH ……………………………………………..University of Hawaii-John A. Burns School of Medicine Janet A. Phoenix, MD, MPH ………………………………………………………………………………….…………..The Grant Group, LLC Marcus G. Plescia, MD, MPH …………………………………………………………. Centers for Disease Control and Prevention Amelie G Ramirez, DrPH ……………………………………………University of Texas Health Science Center at San Antonio Wayne Rawlins, MD, MBA …………………………………………………………………………………………………………………………Aetna Rochelle Rollins, PhD ……………………………………………………………….U.S. Department of Health and Human Services Anthony Signorelli ………………………………………………………………………………………………………..Advertising Council, Inc. Elizabeth Thompson ……………………………………………………………………………………………..Susan G. Komen for the Cure Shaan K. Trotter, MS ………………………………………Lurie Comprehensive Cancer Center of Northwestern University Donald Warne, MD, MPH ………………………………………………………….Aberdeen Area Tribal Chairmen's Health Board C-CHANGE STAFF Brian Alexander, Gary Gurian, MPA, and Tasha Tilghman-Bryant, MPA

C-CHThe timeapprleadby o PREFC-ChBlooCancthis d DarrThomPatriJean SPEC

HANGE mission of

e by leveragroximately 1ers collaborane organizat

FACE hange commomberg Schocer Health Ddocument:

rell J. Gaskin,mas A. LaVeiick Richard, G Ford, MD

CIAL ACKNO

f

C-Change isging the exp150 of the ate on issuetion or even

missioned resool of Public

Disparities. W

, PhD ist, PhD PhD

D, MPH

WLEDGEME

C-CSus

for thdeve

s to eliminapertise and Nation’s cans spanning tone sector a

searchers at c Health to

We gratefully

ENT

Changsan G

heir geelopm

ate cancer aresources

ncer leadersthe continuualone.

the Center fdevelop this

y acknowledg

ge woG. Komenero

ment o

as a major of our mems from the

um of resear

for Health Ds Case Statege the effort

ould limen fous coof thi

public healtmbers. A 50

private, purch, preventi

isparities Soement on thts of the foll

ike toor theontribis doc

th problem 01(c)(3), C-Cblic, and noon, and care

olutions (CHDhe Societal aowing CHDS

o thane Curbutioncume

at the earlChange is con-profit see - that cann

DS) at the Joand EconomS staff in dev

nk re n to nt.

iest possibleomprised octors. These

not be solved

ohns Hopkinsmic Impact ovelopment o

e of e d

s f f

3 | P a g e

TABLE OF CONTENTS Executive Summary ...................................................................................................................... 1 A. Introduction: Purpose & Overview ............................................................................. 1 B. Key Findings ................................................................................................................1 C. Methodology ................................................................................................................3 D. Recommendations ........................................................................................ 3

Section I Data & Methodology ...................................................................................................... 5

A. Factors Associated with Cancer Prevalence ....................................................... 5 B. Impact of Cancer Health Disparities ................................................................. 5 C. Computing the Costs of Premature Cancer Death .............................................. 5 D. Direct Medical Costs of Cancer Disparities ....................................................... 7 E. Estimating the Indirect Costs of Cancer ............................................................ 9 F. Limitations ................................................................................................... 9

Section II Literature Review ........................................................................................................ 10

A. Disparities in Cancer Incidence, Mortality and Survival ..................................... 10 B. Socioeconomic Disparities in Cancer Burden .................................................... 14 C. Disparities in Lung Cancer .............................................................................. 15 D. Disparities in Colorectal Cancer ....................................................................... 16 E. Disparities in Prostate Cancer .......................................................................... 17 F. Disparities in Breast Cancer ............................................................................ 17 G. Cost of Cancer ............................................................................................... 18

Section III Current Trends in Disparities in Cancer Incidence and Mortality Rates (Cancer Incidence and Mortality by Race/Ethnicity and Gender) .................................................................... 2

A. Factors Associated with Cancer Prevalence ....................................................... 26 B. Results for Any Cancer ................................................................................... 26 C. Colon, Breast, Lung, and Prostate Cancer .......................................................... 30 D. Discussion .....................................................................................................

Section IV Societal and Economic Impact of Cancer Health Disparities .......................................... 40

A. Cost of Premature Death ................................................................................. 40 B. The Direct Cost of Cancer Disparities ............................................................... 43 C. The Indirect Cost of Cancer Disparities ............................................................. 45

Recommendations ........................................................................................................ 4 References ................................................................................................................... 50 Cancer Disparities Publications 2000-2010 .................................................................... 55

39

8

0

,

4 | P a g e

EXECUTIVE SUMMARY Purpose and Overview

This case statement describes the societal & economic impact [e.g., Years of Potential Life Lost, etc.] of current cancer morbidity and mortality rates among minority and other disparate populations in the hope that the detailed information and data will help spur a national effort to address cancer health disparities. Racial and economic disparities in cancer are prevalent in the United States. While morally unacceptable, cancer disparities exact a substantial toll on society in terms of premature death, lower productivity and the costs of medical care. This case statement outlines the significant societal and economic costs cancer health disparities impose on the national community. A review and summary of the cancer health disparities literature from 2000-2010 highlights documented disparities and identifies areas where more research and data are needed. To better understand the burden of cancer in the population, the case statement also explores the association between cancer prevalence and demographic, socioeconomic, obesity and other health factors. Most importantly, the report estimates the economic costs of cancer disparities to society. There are other types of costs that we don’t consider in this report. There are the emotional and psychic costs for cancer victims and their families and friends. There are the costs borne by caregivers of cancer victims. This would include the value of hours of service family members and friends provide cancer victims. What’s indisputable is that cancer exacts a heavy toll on the United States, and the disparities in cancer incidence and deaths not only add to that toll but also underscore the need to examine why such disparities exist. Key Findings The literature on cancer health disparities shows:

• For all cancers and the most common cancer sites, African-American men have the highest incidence and mortality rates compared to white, Asian and Hispanic men. (See section 3, pages 12-16.)

• Asian and Hispanic men have lower incidence and mortality rates than white men for all

cancers and for most individual cancer sites, with the exception of stomach, and liver and bile duct cancers. (See section 3, pages 12-13.)

• White women have the highest incidence rates for all cancers, but African-American

women have the highest mortality rates. (See section 3, pages 12-16.)

• Persons with low socioeconomic status have higher rates of cancer incidence and mortality and lower rates of cancer survival compared to more affluent persons. (See section 3, pages 16-18.)

1 | P a g e

5 | P a g e

• The lack of individual income, insurance and education data limits researchers’ ability to study the association between socioeconomic status and cancer. (See section 3, page 16.)

• Despite higher rates of cancer, African Americans and the poor were more likely to

receive suboptimal cancer care compared to whites and affluent persons. (See section 3, pages 18-21.)

An analysis of recent trends in cancer shows:

• Recent federal data show that racial and ethnic disparities in cancer persist. (See section 4, pages 21-25.)

• Socioeconomic disparities in cancer incidence rates do not show a consistent pattern, such that persons living in poorer communities or less educated communities are consistently disadvantaged. (See section 4, pages 25-28.)

• However, cancer mortality rates do show a consistent pattern — poorer and less educated

communities have higher rates of death compared to affluent and better educated communities. (See section 4, pages 25-28.)

• Traditionally disadvantaged groups tended to have lower prevalence rates of cancer. This probably reflects lower screening rates and survival rates for vulnerable groups. (See section 4, pages 29-31.)

The estimates of the costs of cancer health disparities show:

• Disparities in cancer health are costly. (See section 5, page 43.)

• The annual costs of racial/ethnic disparities are almost $197 billion: o $193 billion for premature death (see section 5, page 44); o $2.3 billion for direct medical costs (see section 5, pages 45-46); and o $471.5 million for lost productivity (see section 5, page 47-49).

• Because whites have higher cancer incidence and mortality rates than Asians and

Hispanics, they bear 80% of the costs of cancer disparities. African Americans bear 17% of the costs of cancer disparities. (See section 5, page 44.)

• The costs of socioeconomic cancer disparities are almost $37 billion. These estimates are conservative because they represent disparities between counties by socioeconomic status versus disparities between individual by socioeconomic status. (See section 5, page 43 and section 2, pages 6-7.)

2 | P a g e

6 | P a g e

Methodology The case statement consists of four analyses:

• A systematic review of the recent literature published from 2000-2010 on racial, ethnic and socioeconomic disparities in cancer.

• An analysis of recent trends in cancer incidence and mortality using data from the Centers for Disease Control and Prevention (CDC), National Vital Statistics System, National Cancer Institute (NCI) Surveillance, Epidemiology and End Results (SEER) Program.

• An analysis of cancer prevalence using the Medical Expenditure Panel Survey (MEPS)

2002-2007.

• An analysis of these data sources to compute the economic burden of cancer incidence and mortality and its variance with race/ethnicity and socioeconomic status. ◦ Economic costs of racial/ethnic disparities were calculated by simulating the

decline in the number of cancer deaths and new cancer cases if racial/ethnic differences in cancer burden were eliminated.

◦ To estimate the potential cost savings associated with eliminating socioeconomic disparities in cancer burden, cancer rates and their associated costs were compared between counties with high poverty rates against counties with low poverty rates. A similar analysis was also conducted for counties with high and low levels of educational attainment.

Notable limitations include:

• Lack of Good Data: ◦ Cancer registries do not collect/report socioeconomic information. ◦ Population-based surveys do not collect survival or stage information. ◦ Administrative data lacks clinical, demographic and socioeconomic information.

• The inability to look at vulnerable subgroups of interest:

◦ Socioeconomics: poor and near poor, uninsured and underinsured, & poorly educated;

◦ Race/ethnicity: Hispanic subgroups, Asian subgroups, & Native Americans and Alaska Natives.

Recommendations

• Society needs to focus on cancer prevention and cancer treatment that result in long-term survival. ◦ Programs designed to lower individual risk of cancer, if they are successful in

reducing incidence and mortality among whites and African Americans to levels experienced by Asians and Hispanics, would yield substantial economic value to the nation.

3 | P a g e

7 | P a g e

• The nation should invest in policies and programs designed to reduce cancer risks associated with environmental factors and unhealthy behaviors. ◦ Regardless of socioeconomic status, individuals from racial and ethnic minority

populations experience a myriad of stressors in the context of their social location. The effects of social location are mediated by lifetime exposure to social and environment factors, as well as individually modifiable behavioral risks. Such social and individual-level exposures translate into differential health care quality and access.

• Strategies to reduce race disparities in the quality of cancer-related care must also address the need for care coordination to promote rational utilization of preventive and therapeutic care, including management of comorbid conditions. ◦ The available evidence points to opportunities to intervene to reduce disparities in

cancer-related care, through health risk management, access to a usual source of care, access to primary and secondary treatment, and adjuvant therapies. However, the evidence also suggests that despite health care coverage, many members of racial and ethnic minority populations underutilize services for cancer detection, diagnosis and treatment.

• There are lessons to be learned from the advantage Asians and Hispanics have over whites and African Americans that can help society address modifiable risk factors. ◦ Society should address modifiable risk factors such as physical inactivity, obesity,

heavy alcohol consumption, diets high in red or processed meats, diets that lack fruits and vegetables, diets high in animal fat, smoking, and environmental hazards such as radon, asbestos, air pollution, and certain metals (chromium, cadmium, arsenic) (American Cancer Society, 2009).

• Community-level interventions are warranted. ◦ At the community level, such interventions might address community norms and

lifestyles that promote adverse health outcomes in racial/ethnic minority populations.

◦ Community-level interventions should also address potential adverse effects of the physical and social environment, as well community differences in health care availability.

• We need better data.

◦ Cancer registries should collect socioeconomic data (income and education). ◦ Cancer registries should link existing data to social security records. ◦ Population surveys should develop cancer supplements to collect stage and

survival information. ◦ Possible national survey of new cancer patients based on a sample of new cancer

registrants.

4 | P a g e

8 | P a g e

I. Data and Methodology Factors Associated with Cancer Prevalence

Medical Expenditure Panel Survey (MEPS) data from 1996-2007 were used to explore the association between cancer prevalence and socioeconomic status and access to care. Five sets of models were estimated to identify the association between the likelihood of having a cancer diagnosis and several factors, particularly race/ethnicity, income, education, health insurance and obesity. The dependent variables of interest were whether the survey respondent reported having any cancer, lung, colorectal, breast or prostate cancer. Models were estimated for any type of cancer combined, and for breast, colon and rectal, lung and prostate cancer separately. Other factors were used as control variables in the analysis: rural-urban location and year indicators variables. The models were estimated using the logistic regression techniques. Because MEPS has a complex sampling design and data was pooled across 11 years, the longitudinal weights and designed factors were used to conduct the models. The models were estimated using STATA version 10. Impact of Cancer Health Disparities

There are several sources of societal and economic impact of cancer health disparities. This case statement attempts to quantify three types of costs: the cost of premature death, the cost of medical care to cancer patients, and the indirect cost of cancer on economic productivity through lost wages and hours worked. The case statement’s computation is based on three sets of analyses: the cost of premature death due to cancer disparities, the direct medical costs of cancer disparities, and the indirect costs of cancer disparities. The two types of disparities considered are racial/ethnic disparities and socioeconomic disparities. For racial/ethnic disparities, the report examines differences in cancer incidence and mortality across whites, African Americans, Asians, American Indians and Alaskan Natives, and Hispanics. For socioeconomic disparities, the report used two measures — poverty status and educational status. Counties were ranked and then compared by their level of poverty and educational attainment. Computing the Costs of Premature Cancer Death

To estimate the costs of premature death due to cancer disparities, the years of life lost as a result of disparities were estimated and valued at society’s willingness to pay for a quality adjusted life year. An elimination of disparities would likely result in a reduction of cancer mortality rates within each age cohort to the rate for the race/ethnic group with the lowest rate. We selected the lowest rate for age cohorts. This rate may be attainable across the entire age cohort because it is an actual rate for one of the race/ethnic groups within the age cohort. Presumably, if society can determine why one racial/ethnic group within the age cohort has low cancer mortality, then it may be replicated for other racial/ethnic groups in that age cohort. For the most part, Asians had the lowest cancer mortality rates. First, the number of excess cancer deaths was computed by age cohort using cancer mortality data published by the Centers for Disease Control and Prevention National Vital Statistics System (www.cdc.gov/nchs/nvss/mortality_tables.htm). The number of deaths and crude death rates were obtained by age and race for 2002 to 2006 (there were ten age groups: 1-4, 5-14, 14-24, 25-34, 35-44, 45-54, 55-64, 65-74, 75-84 and 85 and over). Then the number of deaths that would have occurred for each racial/ethnic group was estimated if every group’s death rate were equal to the racial/ethnic group with the lowest death rate within the age

5 | P a g e

9 | P a g e

category. The difference between the actual number of deaths and the estimated deaths represents “excess deaths.” If an age-race cohort did not have sufficient numbers of persons in the population to compute a valid death rate, this age-race cohort was excluded from the analysis.

Second, number of years of life lost in each racial/ethnic group was computed assuming that all persons would live the life expectancy of that particular age category. For example, for persons 1-4, the life expectancy is 76.8 additional years; for persons 35-44, the life expectancy is 41.3 additional years; and for 75-84, the life expectancy is 10 additional years. The life expectancy data is published by the CDC in the 2008 National Vital Statistics report (Arias, 2008). Third, low, moderate and high estimates were computed for the cost of premature cancer deaths by valuing the years of a life lost at society’s willingness to pay for a quality adjusted life year. For the conservative estimate, each year of life loss was quantified at $50,000, the standard value used to evaluate medical intervention using cost effectiveness analysis (Hirth et al., 2000). The lower and upper bounds recommended by Braithwaite and colleagues (2008) were used to compute the moderate and high estimates. In a relatively recent study, they estimated the value of a quality adjusted life year to range from $95,000 to $264,000 (Braithwaite et al., 2008). Data from the Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute were used to compute the costs of socioeconomic disparities in cancer mortality (specifically, SEER/STAT to compute the number of deaths and crude death rates for counties by the degree of poverty status and the level of educational attainment. These are two separate analyses). Analysis was based on poverty status by dividing the counties evenly into five poverty status groups: counties with 0.0-6.23% poverty rate; 6.24%-8.31%; 8.32%-10.72%; 10.73%-14.60%; and 14.61%-55.74%. A second analysis used educational status by dividing the counties evenly into five educational attainment groups based on the percentage of adults in the county with less than a high school education: 3.04-15.04%; 15.05%-18.90%; 18.91%-23.48%; 23.49%-30.48%; and 30.49%-65.30%. Similar to the race/ethnicity analysis, value of the number of years of life loss due to excess cancer deaths was computed.

After obtaining the number of deaths and crude death rates by socioeconomic status (SES) group for 2002 to 2006, the number of deaths that would have occurred for each SES group, if every county’s death rate were equal to the SES group with the lowest death rate within the age category, was estimated. The difference between the actual number of deaths and the estimated deaths represents “excess deaths.” The number of years of life loss in each SES group is assuming that all persons would live the life expectancy of that particular age category. Based on the number of years of life loss, we estimated low, moderate and high costs of premature death due to SES disparities in cancer mortality. The SES analysis differs from the racial/ethnicity analysis because it is based on area level differences in cancer rates for counties divided by SES. The race/ethnic analysis is based on actual racial/ethnic mortality differences. The SES analysis compares cancer mortality rates of persons living in poor versus affluent counties rather than comparing the rates of poor versus affluent individuals. It is suspected that the variation across counties is smaller than the variation across individuals. Hence, the estimated costs of premature cancer deaths due to SES disparities will be lower because we are equalizing rates across counties by SES, not across individuals by SES.

6 | P a g e

10 | P a g e

Direct Medical Costs of Cancer Disparities The direct medical costs of cancer disparities are equal to the excess number of cancer

cases associated with disparities multiplied by the incremental costs of cancer. This calculation involves two steps: 1) computing the number of excess cancer cases attributable to disparities; and 2) the incremental costs of cancer. The number of excess cancer cases was calculated using cancer incidence data from the CDC and SEER 17 for the years 2002-2006. The SEER 17 represents 17 geographic areas and 26.2% of the U. S. population. The costs estimate for the SEER 17 areas were then multiplied by 3.816 to estimate the national costs. This analysis was conducted by race/ethnic group and by classifying counties into SES groups.

For each race/ethnic/SES group, the incidence rate, incidence counts, and population were obtained for ten age cohorts: 1-4, 5-14, 14-24, 25-34, 35-44, 45-54, 55-64, 65-74, 75-84 and 85 and over. After estimating the number of cancer cases that would have occurred for each race/ethnic/SES group if every group’s incidence rate were equal to lowest incident rate within the age category, the report’s estimate of “excess cancer cases” equals the actual number of cancer cases minus the estimated number of cancer cases. This analysis was conducted for all cancer cases, and lung, colorectal, breast and prostate cancer separately. Data from the Medical Expenditure Panel Survey (MEPS) were used to compute the incremental costs of cancer. The MEPS is a longitudinal survey that covers the United States civilian non-institutionalized population. It is co-sponsored by the Agency Health Care Research and Quality (AHRQ) and the National Center for Health Statistics and fielded by AHRQ based on a sampling frame of the National Health Interview Survey. The MEPS is a nationally representative survey of health care use, insurance coverage, medical expenditures, and sources of payment for the U.S. civilian non-institutionalized population. The MEPS is widely used as an authoritative source of information on the nation’s healthcare use. AHRQ uses it to monitor the nation’s progress on healthcare disparities (AHRQ, 2006). More information about the MEPS is available on their website, www.meps.ahrq.gov/mepsweb. Six waves of data were pooled from the 2002-2007 MEPS to increase the sample size to sufficient numbers of respondents with cancer. For this analysis, we used the Household Component (HC) files, which are the core components of the survey; the condition files; and the pooled estimations. HC collects demographic characteristics, health conditions, health status, medical services utilization, access to care, satisfaction with care, health insurance coverage, and income data for each person surveyed. Combining the HC component files with the corresponding Health Condition files measured the different types of cancer conditions. Subsequently, the resulting files were merged with the pooled estimation linkage file from the MEPS to restrict the analytic sample to unique individuals only. The MEPS overlapping design allows repeated observations on the same individuals for several rounds. Hence, it is important to restrict the sample to unique individuals to compute appropriate standard errors in pooled estimations. The analytic sample for direct savings was restricted to a total of 71,549 individuals older than 18 years old with non-missing observations. A model to predict healthcare expenditures data from the 2000-2007 MEPS was developed. Total expenditures in the MEPS include both out-of-pocket and third party payments to health care providers but do not include health insurance premiums. Expenditures for hospital-based services include those for both facility and separately billed physician services. Total expenditures include inpatient, emergency room, outpatient (hospital, clinic and office-based visits), prescription drugs, and other (e.g., home health services, vision care services, dental care, ambulance services, and medical equipment). The prescription drug expenditures do not include over-the-counter purchases.

7 | P a g e

8 | P a g e

11 | P a g e

A modified version of Aday and Andersen’s (1974) Behavioral Model of Health Services Use was used to estimate direct medical costs for patients with cancer compared to those without any cancer. Healthcare expenditures were predicted using demographic, socioeconomic status, location and health status measures. The demographic factors were age, race/ethnicity and gender. The socioeconomic factors were education, income and health insurance status. The location factors were Census region and urban-rural residence. The primary variable of interest is whether the respondent had cancer, and cancer was measured in two ways. The first measure was a simple binary variable that indicated whether the respondent had cancer. The second measures were five binary indicators of different cancer types such as colon, breast, lung, prostate, and other types of cancer. The health condition files were used to measure different types of cancer condition. In the MEPS, health conditions such as cancer were collected from respondents who reported using health care services such as hospital visits or drug purchases.

These conditions were subsequently recorded by the interviewer as verbatim text, which was then coded by professionals using the International Classification of Diseases, Ninth Revision (ICD-9). The ICD-9 codes were subsequently collapsed into meaningful category codes using the Clinical Classification System (CCS) developed by the U.S. Department of Health and Human Services (HHS) (Cohen and Krauss, 1997). CCS codes from 11 to 45 were used to measure the different types of cancer conditions. The other health status measures were self-reported general health status, self-reported mental health status, measures of functional status and limitations, the number of instrumental activities of daily living (IIADL), body mass index, and the presence of other chronic conditions (diabetes, asthma, asthma attack, high blood pressure, heart attack, angina, other heart disease, stroke, emphysema, joint pain and arthritis). To estimate the incremental cost of cancer in general, we estimated the model using the simple any cancers binary variable. To estimate the incremental costs of lung, colorectal, breast and prostate cancers, we estimated the model using the five binary variables for the different sites of cancer. A two-part model approach was used to predict healthcare expenditures. The two-part model accounts for systematic differences in the characteristics of individuals with cancer compared to those without any cancer. The first part of the model consisted of estimating logistic regression models to estimate the probability of having any type of direct expenditures. The second part consisted of using Generalized Linear Models with Log link and Gamma distribution to predict levels of direct expenditures conditional on individuals with positive expenditures. We conducted the different diagnostic and specifications tests as recommended by Manning and Mullahy (1998, 2001). We computed unconditional levels of healthcare expenditures by multiplying the probabilities obtained from the first part of the model by predicted levels of expenditures from the second part of the model. Subsequently, we computed the incremental values for each type of cancer condition. The incremental cost of cancer was computed by using our model to predict medical costs if a person has any type of cancer and subtracting the predicted medical costs if a person does not have cancer (Deb, Manning and Norton, 2006). To do this calculation, the probabilities of having medical costs for persons with cancer and persons without cancer must be taken into account. Similar calculations were computed for each cancer site including colon, breast, lung, prostate, and other sites of cancer, each time comparing them to predicted medical costs if persons did not have cancer.

12 | P a g e

Estimating the Indirect Costs of Cancer Similar to the direct medical costs of cancer, the indirect costs equals the number of

excess cancer cases due to disparities times the estimated productivity loss associated with having cancer. Using MEPS data for the years 2002-2007, the value of productivity loss was computed by estimating the incremental effect of cancer on the number of disability days, hours worked and wages among working age adults. The analytic sample of 50,952 was restricted to individuals between 24 and 65 years old with non-missing observations to capture information on individuals who are active in the labor force. We used a two-part model approach to predict indirect costs due to cancer. The dependent variables were number of disability days, annual works hours, and hourly wages. The first part of the model estimated the probability of having any disability days, hours of work, or any wages. The second part predicted levels of disability days and hourly wages conditional on individuals with positive days or hourly wages. To compute the value of lost wages, the reduction in wages due to cancer was multiplied by the mean number of hours worked by persons with cancer. Limitations

This study has several limitations primarily associated with data problems. There are many studies that focus on African-American/White cancer disparities; however, few studies examine Hispanics, Asians, Native American, Alaskan Natives, Hawaiians and Pacific Islanders. Studies on these minority groups focus exclusively on their disadvantages relative to Whites and ignore data that show their advantages relative to Whites. This bias in the literature is due in part to small sample sizes and the lack of data on Hispanic and Asian subgroups. Relatively few studies focus on socioeconomic disparities, primarily because cancer registries do not routinely collect economic, education or health insurance status information. Consequently, it is expensive, time-consuming and legally difficult to conduct studies of socioeconomic disparities because such studies involve linking confidential socioeconomic information to cancer registry data. Using administrative data from health plans is problematic because these data lack the clinical information required to categorize cancer by severity and survival rate. The national population-based surveys do not offer a satisfactory solution because, while these surveys have detailed socioeconomic data, they lack data on cancer stage, severity and length of survival. These data limitations hinder this analysis as well. We were unable to include Native Americans, Alaska Natives, Hawaiians and Pacific Islanders in our analyses. Similarly, we cannot incorporate the diversity within Hispanics and Asians, specifically looking at subgroups that prior literature has shown to be disadvantaged relative to Whites. The lack of individual socioeconomic information forced us to use area level measures of income and education. This biases estimates of socioeconomic cancer disparities downward. This report’s estimates of cost due to premature death are conservative due to the fact that $50,000 was used to value years of life, as opposed to the higher amounts suggested by Braithwaite (2008). The costs associated with socioeconomic disparities are also conservative. The poverty and education data used in this analysis were measured at the county level as opposed to the individual level because the report examines differences in cancer incidence and mortality rates between low socioeconomic counties and high socioeconomic counties. Because low socioeconomic counties have residents who span the level of socioeconomic status, cancer disparities between counties are smaller than cancer disparities between individuals. This limitation of the CDC and NCI data tends to reduce this report’s estimated costs of socioeconomic cancer disparities.

9 | P a g e

10 | P a g e

13 | P a g e

II. Literature Review

A systematic review of recent literature on racial, ethnic and socioeconomic disparities in cancer specifically focused on cancer incidence, mortality, survival, treatment and costs for the four most prevalent cancers: lung, colorectal, breast and prostate. A search using www.pubmed.gov of the National Library of Medicine, National Institute of Health led to a literature review of 122 peer-reviewed articles published from 1999 through 2010. These articles were identified using the following search terms: cancer prevalence, cancer incidence, cancer mortality, cancer treatment, cancer diagnoses, cancer survival, socioeconomic status, race/ethnicity, minority, racial differences/disparities, lung cancer, prostate cancer, colorectal cancer, breast cancer, cancer treatment, and cancer costs. Also reviewed: reports published by the American Cancer Society, the National Cancer Institute, the Agency for Healthcare Quality and Research, and information published on the website of the National Cancer Institute and the Office of Minority Health. Based on the review of the research literatures, this report concluded the following:

1. African Americans and persons with low socioeconomic status bear a disproportionate cancer burden. They experience higher incidence and mortality rates.

2. Despite a disproportionate cancer burden, African Americans and persons of low socioeconomic status are more likely to receive suboptimal cancer care compared to Whites and more affluent cancer patients.

3. Cancer imposes a significant burden on the United States in terms of loss of life, quality of life and economic activity.

4. The cost of cancer care varies by site, with lung and colorectal cancers being the most expensive, followed by prostate and breast cancers.

Disparities in Cancer Incidence, Mortality and Survival

While cancer is a leading cause of death among all racial and ethnic groups, the mortality burden is not distributed evenly among the various groups (ACS, 2009). African Americans and persons of low socioeconomic status bear the greatest cancer burdens. Despite improvements in cancer survival rates for all racial and ethnic groups since 1988, race disparities persist (Clegg et al., 2002). For all cancers and the most common cancer sites, African-American men have the highest incidence and mortality rates, higher than those of Whites, Hispanics and Asians. Hispanic and Asian/Pacific Islander men tend to have lower incidence and mortality rates than white men for all cancers combined, and for most individual cancers. (See table 2.1) Hispanic and Asian/Pacific Islander men’s rates only exceed white men’s for stomach, and liver and bile duct cancers (Horner, Ries, Krapcho et al., 2009). The racial differences among women are not so straightforward. While white women have the highest incidence rates for all cancers, African-American women have the highest mortality rates (ACS, 2009). Compared to white women, African-American women have a lower incidence of breast cancer, a similar incidence of lung cancer and a higher incidence of colorectal cancer. Despite lower incidence rates for breast cancer, African-American women have higher breast cancer mortality rates than white women. African-American women have higher rates of mortality for colorectal cancer but lower mortality rates for lung cancer compared to white women. Hispanic and Asian/Pacific Islander women uniformly have lower incidence and mortality rates for all cancers and for the three most common cancers compared to both African-American and white women. Because the population of Pacific Islanders is relatively small, they

14 | P a g e

are combined with Asians. Combining Pacific Islanders with Asians may obscure their higher cancer burden relative to Whites. Table 2.1 Age Adjusted SEER Cancer Incidence and Mortality Rates by Site, Race and Ethnicity, US 2002-2006 Incidence Rates White African

American Asian & Pacific Islander

Hispanic

Men All sites 544.3 633.7 349.1 409.7 Lung 77.6 104.3 53.4 43.2 Colorectal 56.9 69.3 46.9 46.3 Prostate 153.0 239.8 91.1 133.4 Women All Sites 420.5 398.9 287.5 312.5 Lung 54.8 54.7 28.1 24.7 Colorectal 42.1 53.5 34.6 32.2 Breast 127.8 117.7 89.5 88.3 Mortality Rates Men All sites 226.4 304.2 135.4 154.7 Lung 69.9 90.1 36.9 33.9 Colorectal 21.4 31.4 13.8 16.1 Prostate 23.6 56.3 10.6 19.6 Women All Sites 157.3 183.7 95.1 103.9 Lung 41.9 40.0 18.2 14.4 Colorectal 14.9 21.6 10.0 10.7 Breast 23.9 33.0 12.5 15.5 Per 100,000, age adjusted to the 2000 U.S. standard population. Hispanic category is not mutually exclusive of whites, African Americans, Asians/Pacific Islanders and American Indians/Alaska Native Source: Horner, Ries L, Krapcho et al. (eds). SEER Cancer Statistics Review, 1975-2006, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2006/, based on November 2008 SEER data submission, posted to the SEER web site, 2009.

Whites have consistently higher rates of survival than African Americans. (See table 2.2) Five-year survival rates for white males range from 4-27% higher than those for African-

11 | P a g e

12 | P a g e

15 | P a g e

American males. Five-year survival rates for white females range from 16-26% higher than those for African-American females. Disparities in survival rates are due in part to racial differences in stage of diagnosis and treatment (ACS, 2009). However, the survival advantage for whites may occur even in the context of randomized controlled therapeutic trials featuring uniform stage, treatment and follow up. Albain et al. (2009) report that whites’ advantage in breast, cervical and prostate cancer persists after adjusting for prognostic factors.

Table 2.2 5-Year Relative Survival Rates for All Cancers and Three Most Common Cancer Sites by Race and Gender, 1999-2005 White African American Rate Ratios Men All 67.0 60.6 1.11 Lung 13.7 10.8 1.27 Colorectal 66.3 55.5 1.19 Prostate 99.9 96.5 1.04 Women All 66.9 55.2 1.21 Lung 18.3 14.5 1.26 Colorectal 65.9 56.7 1.16 Breast 90.3 77.9 1.16 Source: Horner, Ries, Krapcho et al. (eds). SEER Cancer Statistics Review, 1975-2006, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2006/, based on November 2008 SEER data submission, posted to the SEER web site, 2009.

Race disparities in cancer incidence and mortality are due in part to differences in: 1)

exposures to cancer causing agents; 2) access to screening and cancer care after the onset of symptoms; 3) access to appropriate cancer treatment; 4) contextual and environmental factors that affect cancer survival; and 5) genetic factors (Hackbarth, 2001; Wong et al., 2009). These disparities in cancer risk factors stem from cultural and language differences and societal inequities related to discrimination, provider bias, patient mistrust, poor patient-provider communication, poor adherence and low quality care (ACS, 2008). While a few studies have reported genetic differences by race, these differences typically make only a modest impact on cancer mortality.

There is little evidence to indicate that cancer disparities are fundamentally due to genetic differences among race groups. The most often cited example of the influence of genetics in cancer disparities is documented for prostate and breast cancers (ACS, 2008). There are several reports of racial differences in the frequency of functional polymorphisms that have been implicated in carcinogenesis and/or cancer survival. However, the magnitude of the contribution of such genetic differences to the observed racial disparities has been consistently minimal or absent (Collins, 2004). While this remains a legitimate area of scientific investigation, the available evidence suggests differences in socioeconomic status, differences in exposure to

13 | P a g e

16 | P a g e

carcinogens, clinical factors, and disparities in cancer treatment as primary drivers of racial disparities (ACS, 2009).

Since race and ethnicity are related to socioeconomic status, some of the disparities between African Americans and whites are associated with race differences in income, education and health insurance coverage. (See table 2.3) African Americans have higher poverty rates, lower educational attainment and are more likely to be uninsured or covered by Medicaid compared to whites (DeNavas-Walt, Proctor, and Smith, 2008). However, it should be noted that despite having lower socioeconomic status, Hispanics do not have a higher cancer burden than whites. Asians have the lowest cancer burden. In comparison to whites, Asians have higher median income and education attainment but higher poverty rates, percentage uninsured and covered by Medicaid. This paradoxical disparity in cancer burden among whites, Hispanics and Asians, despite higher rates of poverty for the latter two groups, suggests that other factors in addition to poverty status account for African-American/white disparities in cancer mortality and morbidity. The apparent disadvantage of whites relative to Hispanics and Asians is understudied and not well understood.

Table 2.3 Poverty Rate, Educational Attainment* and Health Insurance Coverage by Race and Hispanic Origin, 2008. White, Non

Hispanic African American

Asian & Pacific Islander

Hispanic

Poverty Rate 8.6 24.7 11.8 23.2 Less the HS 9.4 18.1 11.2 37.0 High School Graduate

31.2 34.9 19.1 30.1

Uninsured 10.8 19.1 17.6 30.7 Medicaid 9.5 25.4 11.6 24.3 Medicare 17.0 11.9 9.5 6.8 Privately Insured

75.0 52.2 68.2 43.8

Military Health Care

4.3 4.1 2.2 1.9

Source: DeNavas-Walt, Carmen, Proctor and Smith, US Census Bureau, Current Population Reports, P60-236, Income, Poverty and Health Insurance Coverage in the United States: 2008, US Government Printing Office, Washington, DC 2009. US Census Bureau, Current Population Survey, 2008 Annual Social and Economic Supplement Tables 3-6. Internet Release Date: April 2009. * Computed for Persons 18 years of age and older. * Health insurance totals sum to more than 100 because individuals can be covered by more than one type of health insurance.

17 | P a g e

Socioeconomic Disparities in Cancer Burden Prior studies consistently show that cancer burden increases as socioeconomic status

(SES) declines (ACS, 2008). However, studying the association between SES and cancer burden is complicated because cancer registries do not routinely collect data on SES. Consequently, researchers must link area-level SES data from the Census or individual-level data from other sources. From the Census, researchers have relied on median household income, poverty rates and the percentage of adults with high school or college education within a zip code or census tract.

Researchers have also used Medicaid enrollment as a measure of SES. Area level studies usually show some association between cancer burden and SES (Du et al., 2005). Xiao et al. (2007) found a negative association between median income and the percentage of college graduates in a Census tract and the likelihood of being diagnosed with late-stage prostate cancer. Du et al. (2006) found that racial disparities in prostate cancer mortality were related to area-level SES. The relationship between cancer health and area-level SES appears to be independent of racial disparities in cancer (Albain et al., 2009); however, it is important to stress that data limitations hamper efforts to fully characterize the SES-race-cancer interaction. In studies that have been able to measure SES with individual-level data (i.e., the SES of the individual rather the community where they live), the association between SES and cancer burden is stronger. This seems to explain a substantial portion of the racial disparity in cancer burden (Albano et al., 2007). Persons with higher educational attainment have lower mortality rates for cancer. This association was consistent across sites, race and gender groups (Albano et al., 2007). Persons with low educational attainment have greater rates of cancer incidence, lower survival rates and higher mortality rates. Among Medicare beneficiaries, cancer mortality rates decline with income and educational attainment. Persons with private insurance have lower mortality rates compared to persons who are uninsured or covered by Medicaid. Socioeconomic status influences cancer burden in four ways (ACS, 2008):

1) Lower socioeconomic status increases exposure to cancer risk factors such as higher rates of smoking, heavy drinking, obesity, physical inactivity, and exposure to environmental carcinogens.

2) Lower socioeconomic status reduces the likelihood of cancer screening and early detection.

3) Lower socioeconomic status reduces the likelihood of timely treatment. 4) Lower socioeconomic status is associated with less effective contextual support for

cancer patients and greater contextual stress for cancer survivors. The data on the relationship between cancer burden and socioeconomic status are not as rich as the data for racial and ethnic differences, primarily because cancer registries do not include socioeconomic measures. This presents a significant impediment to effective surveillance of cancer health disparities.

Cancer mortality is inversely related to health insurance coverage. A Michigan-based

study combined cancer registry data, Medicaid enrollment files, and death certificate data and reported that Medicaid enrollees were a disproportionate share of the state cancer patients (Bradley et al., 2001). Among non-elderly women, Medicaid enrollees had a higher incidence of breast, lung, and colorectal cancer. Among older adults, both male and female Medicaid enrollees had a higher incidence of lung and colorectal cancer. Similarly, Medicaid enrollees younger than 65 were more likely to be diagnosed at advanced disease stages, i.e., stage III or

14 | P a g e

IV. The adjusted odds ratios, comparing Medicaid to non-Medicaid enrollees for lung, colorectal and breast cancers, were 1.57, 1.59 and 1.77, respectively. Table 2.4 Relative Risk of Cancer by Site Between Medicaid and Non Medicaid Enrollees in Michigan. Men (Ages 25-64) Men (Ages 65 and older) Lung 5.25 1.94 Colorectal 2.48 1.31 Prostate 1.02 0.95 Women (Ages 25-64) Women (Ages 65 and older) Lung 4.48 1.91 Colorectal 1.86 1.71 Breast 1.37 0.99 Source: Based on calculations from data published in Bradley, Givens and Roberts (2001). Disparities in Lung Cancer

Lung cancer is the leading cause of cancer death. African-American men and low-income persons bear the greatest lung cancer burden. There are major gender differences in the pattern of lung cancer disparities. Cancer incidence and mortality for African-American males is substantially greater than for white males, while the lung cancer mortality experience for African-American females is closer to that of white females. Cigarette smoking is the leading risk factor for lung cancer. Disparities exist in the prevalence of smoking, comparing African Americans, Asian and Pacific Islanders, Hispanics/Latinos, and whites. Compared to whites, African Americans and Hispanics are less likely to smoke in adolescence (Hernandez et al., 2010). However, both Hispanics and African Americans tend to take up smoking at a later age so that by mid-life both groups have higher smoking prevalence compared to non-Hispanic whites (LaVeist, 2005). Hispanics and African Americans are less likely to quit compared with whites, and both minority groups are more likely to smoke cigarettes that are high in tar and nicotine compared to whites. This, at least in part, explains higher lung cancer mortality among African Americans compared to whites. While there is no recommended evidence-based screening method for lung cancer, racial disparities in lung cancer survival are well documented. Some of the disparity in survival for non-small cell lung cancer can be attributed to disparities in treatment (Bach, 1999). Among patients who undergo similar treatment, survival rates were similar. However, African Americans were 16% less likely to undergo surgery. Shugarman et al. (2009) found significant racial differences in treatment among Medicare lung cancer patients. Compared to whites, African Americans were 66% less likely to receive timely and appropriate treatment and were 50% less likely to receive resection. The reasons for the relatively lower rate of surgical resection are not completely understood. Among stage III cancer patients, African Americans were 34% less likely to receive timely surgery, chemotherapy, or radiation compared to whites. African Americans were 51% less likely to receive chemotherapy in a timely fashion for stage IV cancer compared to whites (Shugarman et al., 2009).

Bryant and Cerfolio (2008) found that African Americans’ survival disadvantage was attributable to higher rates of smoking, lower income, greater treatment delays and lower use of

15 | P a g e

16 | P a g e

19 | P a g e

neo-adjuvant therapy in a sample of 930 non-small cell lung cancer patients. African Americans had lower five-year survival rates than whites for stage I (RR=0.93), stage II (RR=0.85), and stage III (RR=0.63) cancers. However, among persons diagnosed with stages I and II cancer, this difference disappears after controlling for socioeconomic status and smoking. The difference also disappears for patients diagnosed with stage III lung cancer after controlling for socioeconomic status, smoking and receipt of neo-adjuvant chemoradiotherapy. Some studies reported no racial disparities in survival when patients had comparable functional status and access to treatment (Blackstock et al., 2002; Mulligan et al., 2006). Blackstock and colleagues report that race differences in lung cancer patients can be explained by race differences in pre-disease performance status (i.e., the ability to do strenuous physical activity). There were no race differences in lung cancer survival and stage of disease among patients using the military health care system (Mulligan et al., 2006). (On a population basis, it is unclear to what extent racial differences in the prevalence of comorbid conditions contribute to the observed patterns of lung cancer disparities.) Disparities in Colorectal Cancer

Colorectal cancer represents about 10% of new cancer cases and 9% of cancer deaths. Because risk factors for colorectal cancer are not well defined, there currently is no recommended strategy for primary prevention of colorectal cancer for the population at average risk. However, available screening modalities detect colorectal cancer at early stages and are effective to prevent mortality. African Americans are at greater risk of being diagnosed with colorectal cancer and have a higher rate of mortality than any other racial or ethnic group. Mortality differences ranging from 45-90% persisted even when patients were stratified by educational attainment (Albano, 2007). Despite the disparities in colorectal cancer burden, African Americans are less likely to receive screening. Using data from the Medicare Current Beneficiary Survey for 2000, 2003, and 2005, Doubeni and colleagues found that a higher proportion of non-Hispanic whites, compared to African Americans and Hispanics, had undergone lower-gastrointestinal endoscopy within five years and/or home fecal occult blood test within the past year. During the study period, Hispanic/white differences persisted; however, African-American/white differences narrowed in 2003 and increased in 2005 (Doubeni et al., 2010). In addition, following flexible sigmoidoscopy, African Americans have a lower follow-up rate for screen-detected abnormalities than whites, with little difference in the yield of colorectal neoplasia (Laiyemo et al., 2010). Moreover, once diagnosed with colorectal cancer, African Americans are less likely than whites to receive standard therapy. White et al. (2008) estimated that African Americans were 16% less likely to receive standard therapy. Among Medicare patients, African Americans had lower rates of resection, sphincter preservation and adjuvant radiotherapy for patients with resectable rectal cancer (Morris et al., 2004). Esnola et al. (2008) reported that among more representative population of colorectal cancer patients, African Americans were less like to receive surgical resection and more likely to receive local excision versus segmental resection. There are also treatment disparities by socioeconomic status. Persons living in zip codes with low educational attainment and persons covered by Medicaid or who are uninsured were also less likely to receive surgical resection (Esnola et al., 2008).

17 | P a g e

20 | P a g e

Disparities in Prostate Cancer Prior studies show that African Americans were at greater risk of having prostate cancer

compared to whites. This race disparity persisted regardless of socioeconomic status and age group. Using data from the California SEER, Cheng and colleagues estimated odds ratios for age and SES groups (Cheng et al., 2009). These estimates ranged from 1.30 to 2.68, depending on the age group and SES (Cheng et al., 2009). Race disparities were greater for younger and more affluent groups. African Americans were more likely to be diagnosed with later-stage and higher-grade prostate cancer (Klassen et al., 2004; Xiao et al., 2007; Hoffman et al., 2001). The odds of being diagnosed with a late stage were 1.36 to 2.26 times higher for African-American men compared to white men. Compared to whites, the odds of being diagnosed with a higher grade of prostate cancer for African Americans were 1.32-1.44 to 1.

Between ages 50 and 64, African Americans are more likely to report prostate cancer screening within the last year than whites. However, between ages 65 and 79, the screening prevalence is greater for whites, and it is comparable for both groups beginning around the age of 80. Similarly, the screening rates are higher for Hispanics than for whites between ages 50 and 59, but the rates for whites are higher between ages 60 and 79 and beginning at age 80, screening rates for Hispanics become appreciably greater than those for whites. While screening for prostate cancer is widely available through digital rectal exam and the test for prostate-specific antigen (PSA), PSA testing is controversial. Prostate cancer screening has not been shown to reduce mortality, and there is growing concern about a possible unfavorable balance of risks vs. benefits of PSA testing (Jerant et al., 2004). Under these circumstances, the effect of differences in screening patterns on race differences in prostate cancer mortality remains an open question.

Once diagnosed, African Americans were at a greater risk of mortality from prostate cancer (Albano et al., 2007; Albain et al., 2009; Thompson et al., 2001; Du et al., 2006). The disparity in mortality risk was inversely related to educational attainment. The relative risk of death was 3.32 for persons with a high school diploma or less and 2.19 for persons with more than a high school education (Albano et al., 2007). Hazard ratios ranged from 1.17 to 1.39 for cancer survival (Albain et al., 2009; Thompson et al., 2001; Du et al., 2006). Differences in survival can be explained by race variations in receipt of primary therapies (Holmes et al., 2009). Hazard ratios declined from 1.26 to 0.98 when use of primary therapies was considered. African Americans were more likely to receive watchful waiting alone and less likely to receive radiation plus androgen deprivation therapy (Holmes et al., 2009; Shavers et al., 2004). Race disparity in survival disappears in small geographic units, suggesting modifiable societal factors were responsible for African-American/white differences. Shavers and colleagues reported that African Americans and Hispanics were less likely to have a urology visit or PSA test but had a higher probability of receiving a bone scan. Compared to whites, African Americans had fewer primary care visits and were more likely to receive watchful waiting (OR = 1.4). Disparities in Breast Cancer

Despite lower incidence rates of breast cancer, African American women have lower survival rates and higher mortality rates than white women. Depending on the study, African-American women have hazard rates that are 10-80% greater than those for non-Hispanic white women. Disparities vary by age, socioeconomic status and region of the country. Race disparities are greatest among younger women (Swanson, Haslam and Azzouz, 2003). Compared to their white peers, younger African-American women are diagnosed at a more advanced stage of breast cancer, tend to have larger tumors and are more likely to have lymph

18 | P a g e

21 | P a g e

node positive tumors. They are also more likely to have triple negative tumors, an especially aggressive form of breast cancer.

After controlling for socioeconomic status, race disparities diminish but do not disappear. Albano reports that race disparity in the relative risk of dying was 1.43 for women with less than or equal to a high school education compared to 1.68 for women with more than a high school education. After controlling for treatment and socioeconomic status, Du, Fang and Meyer (2008) report that the African American to non-Hispanic white hazard ratios decline from 1.83 to 1.21. In addition to race disparities, there are significant disparities by socioeconomic status. Bradley, Given and Roberts (2001) report that women enrolled in Medicaid not only had a higher incidence rate, but they were diagnosed at a later stage and had a higher risk of death than women with other insurance or who lacked insurance. Cost of Cancer

Numerous studies have assessed the cost of cancer. The cost of cancer can be divided into three main parts: the cost of cancer treatment, the indirect cost of cancer due to lower productivity for cancer patients, and the cost of premature death due to cancers (Taylor, 2005; Yabroff et al., 2007; NHLBI, 2009). The cost of cancer treatment is the direct medical care expenditures attributed to cancer care and additional care provided patients for other conditions due to complications associated with cancer. The indirect cost of cancer is the value of the lost days of work and the lower wages incurred by cancer patients due to their illness. The cost of premature death is the value of the years of life lost due to cancer mortality. Presumably, persons whose lives are cut short by cancer would have contributed to their families, communities and society.

Yabroff and colleagues (2007) conducted a comprehensive review of 60 studies published from 1995 through 2006 on the cost of cancer treatment. Researchers have computed the costs of cancer prevalence and the costs of the initial and last year of life phases of care for the most common cancer sites. They found some variation in the cost estimates based on cancer site and stage, the demographic profile of the population, and the type of care rendered. Despite the variation, some patterns emerge. The annual and lifetime costs of cancer are expensive, ranging from several thousand dollars to tens of thousands of dollars. Lung and colorectal cancers were more expensive than breast and prostate cancers. The lifetime costs, regardless of cancer site, were u-shaped, i.e., higher at both the initial and last-year-of- life phases of care. Using Medicare administrative claims, Penberthy and colleagues (1999) found that the average costs during the initial treatment period were $24,910 for colorectal cancer, $21,351 for lung cancer, $14,361 for prostate cancer and $12,141 for breast cancer. Lung cancer appears to be the most expensive cancer site. Kutikova and colleagues (2005) found that incremental cost attributable to lung cancer was $42,990 over a two-year period. Studies identified by Yabroff and colleagues (2007) report a range of $29,178 in 1990 dollars over 1.7 years to $47,941 in 1992 dollars over 2 years. During the initial treatment phase (about 6 months), total costs per month were $11,496, and during the terminal treatment phase (the last 6 months), costs were $9,399 per month (Kutikova, 2005). Colorectal cancer appears to be the second most frequently studied. Yabroff and colleagues (2007) reviewed nine studies. Cost estimates for initial phase of treatment ranged from $16,197 for patients aged 35 and older to $29,384 for patients over 50. For the last 6 months of life, cost estimates ranged from $16,062 for elderly patients to $23,000 for patients of

22 | P a g e

all ages.1 Lifetime medical costs for long-term colorectal cancer survivors were almost $20,000 compared to Medicare patients without the disease. Using NCI SEER data linked to Medicare claims, Wright and colleagues estimated the 16 month net estimated cost of colorectal cancer care was $33,451. While there were racial differences in unadjusted colorectal cancer care costs, these differences disappear after adjusting for covariates (Wright et al., 2007). The variation in the cost of care estimates is due to the variation in data sources, the age of the population, the methodology for identifying colorectal cancer patients, and the methods used to estimate longitudinal costs (Yabroff et al., 2009). Most of the cost-of-treatment studies focus on breast cancer. A review of 29 U.S. cost-of-illness studies for breast cancer found that the lifetime per-patient costs of breast cancer ranged from $20,000 to $100,000 (Campbell et al., 2008). A study of women in five managed care plans showed adjusted per member per month costs of 2.28 times higher for women with breast cancer compared to those who do not have breast cancer (Barron et al., 2008). The estimated annual costs for a breast cancer patient were at least $12,828 higher than those for women without breast cancer (Barron et al., 2008). Like the other cancer sites, the cost of treating prostate cancer is u-shaped. The cost of treating the initial phase of prostate cancer ranged from $6,375 in 1996 dollars in men aged 65 years and older to $17,226. The cost of treating prostate cancer in the last year of life ranged from $12,061 to $24,260 (Yabroff et al., 2007).

The National Institutes of Health estimate the economic cost of cancer was $259.6 billion in 2009: $99 billion in direct medical costs, $19.6 billion in lost productivity and $124.8 billion in premature death (NHLBI, 2009). Using a willing-to-pay approach and life tables from the Berkeley Mortality Database, Yabroff and colleagues (2008) estimated the cost of premature death due to cancer in the year 2000 was $960.6 billion. Lung cancer deaths were the most costly at $270.8 billion, or 28% of the total cost, followed by colorectal at $90.9 billion (9.5%), female breast cancer at $85.3 billion (8.9%) and prostate cancer at $34.8 billion. These costs are expected to increase dramatically, even if mortality rates remain constant, because the population is aging. Using a human capital approach to value premature death suggests that the cost of cancer mortality was approximately $115.8 billion in 2000 (Bradley et al., 2008).

1 For comparison purposes, costs cited from the Yabroff review were converted into 2002 dollars.

19 | P a g e

III. Current Trends in Disparities in Cancer Incidence and Mortality Rates

Recent data on cancer disparities show that racial and ethnic disparities in cancer persist. (See table 3.1) African-American males are at greatest risk of cancer, followed by white males and Hispanic males. Asian/Pacific Islander (API) and American Indian/Alaska Native (AI/AN) males have the lowest risk of cancer. Compared to white males, relative risks of cancer are 0.57 for APIs and 0.61 for AI/ANs. The cancer risks for Asian/Pacific Islander and American Indian/Alaska Native males were below the cancer risks for white and African-American females. White women had the highest risk of cancer, followed by African-American women, then Hispanic women. API and AI/AN women have the lowest cancer risks among women. Based on this data, it is clear that eliminating disparities in cancer incidence between African Americans and whites would be a modest goal because these two groups tend to have the highest levels of cancer incidence. A more ambitious goal would be to reduce the cancer incidence for all race/ethnic groups to levels experienced by APIs and AI/ANs. Such a goal would have a substantial impact on number of persons diagnosed with cancer. Not only would such a campaign impact the cancer risks for African Americans, but also for whites and Hispanics. Table 3.1 Age-Adjusted Cancer Incidence Rates by Race/Ethnicity and Gender, 2002 - 2006 Males White African

American Asian/Pacific Islander

American Indian/Alaskan Native

Hispanic

All-cause 548.9 621.8 332.6 313.3 429.9 Lung 86.2 104.8 50.2 57.1 49.6 Colorectal 58.2 67.9 43.8 37.4 50.0 Prostate 145.3 229.3 81.7 81.3 130.4 Females White African



Hispanic

All-cause 419.9 390.3 274.8 262.2 327.2 Lung 57.3 51.5 27.5 40.6 26.6 Colorectal 42.6 51.6 33.0 30.3 35.1 Breast 123.2 113.0 81.0 66.7 90.2 Source: U.S. Cancer Statistics Working Group. United States Cancer Statistics: 1999–2006 Incidence and Mortality Web-based Report. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2010. Available at: www.cdc.gov/uscs.

Racial/ethnic disparities in cancer mortality follow a pattern somewhat similar to cancer incidence rates. (See table 3.2.) African-American males were at greatest risk, followed by white

20 | P a g e

24 | P a g e

males, then African-American females and then white females. API and AI/AN males had lower cancer mortality than African Americans and whites, regardless of gender. Unlike cancer incidence rates, African-American women have higher mortality rates than white women with the exception of lung cancer, where the mortality rate is relatively close. Similar to cancer incidence rates, equalizing cancer mortality between African Americans and whites would have a modest impact; however, reducing cancer mortality to the level experienced by APIs and AI/ANs would reduce the cancer burden substantially. Table 3.2 Age-Adjusted Cancer Mortality Rates by Race/Ethnicity and Gender, 2002 - 2006 Males White African



Hispanic

All-cause 226.7 304.2 135.4 146.4 154.5 Lung 69.9 90.1 36.9 41.2 33.8 Colorectal 21.4 31.4 13.8 15.3 16.1 Prostate 23.6 56.3 10.6 16.8 19.6 Females White African



Hispanic

All-cause 157.3 183.7 95.1 110.1 103.8 Lung 41.9 40.0 18.2 28.3 14.4 Colorectal 14.9 21.6 10.0 10.2 10.7 Breast 23.9 33.0 12.5 14.3 15.5 Source: U.S. Cancer Statistics Working Group. United States Cancer Statistics: 1999–2006 Incidence and Mortality Web-based Report. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2010. Available at: www.cdc.gov/uscs.

Socioeconomic disparities in cancer incidence rates do not pattern uniformly, so that persons living in poor communities are consistently disadvantaged. For all cancers, persons in more affluent counties and counties with the lowest percentage of high school drop-outs had the highest cancer incidence rates. (See table 3.3.) This was also true for breast and prostate cancer, but was not true for lung and colorectal cancer. Cancer incidence rates for lung and colorectal cancer followed the traditional health-wealth gradient whereas poverty increases cancer incidence rates increase.

21 | P a g e

25 | P a g e

Table 3.3 Age-Adjusted Cancer Incidence Rates by County Level Socioeconomic Status, 2002 - 2006 Poverty Status % below the poverty line 0%-6.23% 6.24%-

8.31% 8.32%-10.72%

10.73%-14.60%

14.61%-55.74%

All-cause 479.6 463.4 458.1 448.5 455.8 Lung 61.8 61.6 63.6 61.7 74.4 Colorectal 49.5 49.6 46.4 49.0 51.7 Breast 131.9 125.8 124.2 116.9 111.5 Prostate 165.6 157.7 152.5 159.9 151.0 Education Status % less than HS grad 3.04%-

15.04% 15.05%-18.90%

18.91%-23.48%

23.49%-30.48%

30.49%-65.30%

All-cause 475.0 469.8 474.4 435.3 456.5 Lung 60.5 63.9 69.2 56.4 76.8 Colorectal 48.3 49.5 51.0 47.8 50.7 Breast 132.0 127.3 122.7 115.7 109.6 Prostate 167.6 158.1 162.2 154.4 144.3 Source: Cancer incidence data for 2002-2006 from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program.

Cancer mortality rates do follow the traditional health-wealth gradient. For all cancers, and the four most common cancers, persons living in high poverty counties and counties with high percentages of high school drop-outs had higher cancer mortality rates. (See table 3.4.) This is surprising given the higher incidence rates for more affluent and better educated counties. However, this is probably an indication that persons living in poorer and less educated counties have less access to cancer treatments that may prolong survival. Also, the lower incidence rates may be due to a higher percentage of undiagnosed cancers in the poorer and less educated counties.

22 | P a g e

26 | P a g e

Table 3.4. Age-Adjusted Cancer Mortality Rates by County Level Socioeconomic Status, 2002 - 2006 % below the poverty line 0%-6.23% 6.24%-

8.31% 8.32%-10.72%

10.73%-14.60%

14.61%-55.74%

All-cause 747.7 783.2 840.4 851.3 911.2 Lung 49.5 53.2 57.1 53.6 56.7 Colorectal 17.6 17.7 18.6 18.5 19.8 Breast 24.4 23.6 25.0 24.7 25.9 Prostate 24.9 23.9 26.1 26.3 30.1 % less than HS grad 3.04%-

15.04% 15.05%-18.90%

18.91%-23.48%

23.49%-30.48%

30.49%-65.30%

All-cause 752.0 790.5 847.6 849.8 891.9 Lung 49.9 53.3 56.2 54.7 55.8 Colorectal 17.2 17.8 19.3 18.6 19.1 Breast 24.3 24.3 25.0 24.7 24.5 Prostate 25.3 24.6 26.3 26.1 28.0 Source: Cancer mortality data for 2002-2006 from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program.

The association between race/ethnicity and cancer burden persists even after controlling for the level of poverty and education in the county. (See tables 3.5 and 3.6.) Among counties with similar socioeconomic status, African-American males had the highest rates of cancer incidence and mortality, followed by white males, white females and African-American females. APIs and AI/ANs have the lowest rates of cancer incidence and mortality within each socioeconomic group.

23 | P a g e

27 | P a g e

Table 3.5 Race and Gender Specific, Age-Adjusted Cancer Incidence and Mortality Rates, By Poverty Status of Counties, 2002-2006 0%-

6.23% 6.24%-8.31%

8.32%-10.72%

10.73%-14.60%

14.61%-55.74%

Incidence Rates Male White 579.4 569.8 566.2 568.8 576.2 African American 627.1 636.8 642.0 680.6 672.5 AI/AN 304.7 423.3 280.9 243.4 261.5 API 334.7 382.4 334.9 359.6 304.2 Hispanic 454.3 417.4 415.9 398.3 383.7 Female White 450.3 440.0 436.8 439.3 418.1 African American 400.8 403.8 406.5 409.2 396.0 AI/AN 312.9 397.1 249.3 192.1 219.1 API 271.9 306.7 282.4 295.2 246.1 Hispanic 354.7 321.5 313.3 299.2 295.9 Mortality Rates Male

White 60.0 60.1 57.2 60.8 64.4 African American 67.8 67.8 69.0 74.3 76.4 AI/AN 33.8 66.4 32.1 23.0 30.9 API 42.2 52.1 43.1 52.2 42.8 Hispanic 52.5 50.8 47.2 43.6 46.0

Female White 45.0 44.4 42.1 44.6 45.5 African American 51.2 53.9 55.2 56.7 56.4 AI/AN 34.9 65.8 32.5 18.5 20.6 API 30.9 37.5 32.5 37.4 29.5 Hispanic 36.7 35.4 31.6 30.8 30.9

Source: Cancer incidence and mortality data for 2002-2006 from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program.

24 | P a g e

28 | P a g e

Table 3.6 Race and Gender Specific, Age-Adjusted Cancer Incidence and Mortality Rates, By Education Status of Counties, 2002-2006 3.04%-

15.04% 15.05%-18.90%

18.91%-23.48%

23.49%-30.48%

30.49%-65.30%

Incidence Rate Male



Mortality Rate Male



Source: Cancer incidence and mortality data for 2002-2006 from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program.

25 | P a g e

29 | P a g e

Factors Associated with Cancer Prevalence The association between cancer prevalence and race, income, insurance, and education

was estimated using the Medical Expenditure Panel Survey (MEPS) data from 1996-2007. The goal of this analysis was to provide some insight into how variations in race and socioeconomic factors influence the prevalence of cancer. This analysis was conducted for any cancer and for the four most common cancer sites (lung, colorectal, breast and prostate). Initially, the model was estimated using MEPS data from 2002-2007 to include body mass index, exercise and smoking behavior as predictors in the model. This data is not available in the MEPS for the prior years. We compared the results of our model with and without these variables and concluded that they did not confound the associations between cancer prevalence and race, income, insurance, and education. Therefore, we dropped these variables from the analysis and were able to expand our data to include the years 1996-2001. This increased the number of persons with cancer in our analysis and was important for our analysis of the four most common cancer sites. Results for Any Cancer

Four different model specifications were used to compare the results for robustness and to address potential co-linearity between education, income, and insurance status. (See tables 3.7 and 3.8.) Minority patients, including African-Americans and Hispanics, are less likely to have been diagnosed with any types of cancer compared to white patients in all four models after controlling for demographic, social-economic, access, and health behaviors variables (P<0.001). To compute the odds ratio relative to whites, the coefficients for African-Americans and Hispanics were exponentiated (i.e., e-0.45=0.64 and e -.50= 0.61). African-Americans had 36% and Hispanics had 39% lower odds of having cancer compared to whites. In models that controlled for different income levels but not for education, any cancer prevalence was negatively associated with low-income patients in families with income less than 200% of the Federal Poverty Level (FPL) compared to those with income over 400% of FPL (P<0.10). However, there was no significant association between any cancer prevalence and different income levels in models that controlled for different levels of education, including individuals with less than high school degrees, high school degrees, and college graduates. In fact, there were significant negative associations between any cancer prevalence and lower levels of education such as no high school education (up to 8th grade) and some high school (9th-11th grades) compared to those with a high school diploma or GED (P<0.001). There was no statistical difference between high school graduates and persons with college experience. The odds of having cancer were 35% lower for persons with no high school training and 27% lower with some high school. Yet, the sizes of the coefficients for lower levels of education were reduced but remained significant in full models that controlled for health insurance status and different income levels. Finally, any cancer prevalence was negatively associated with ‘un-insurance’ status compared to patients with private insurance (P<0.001) even after controlling for education and income. The odds of an uninsured person having cancer were 25% lower compared to a privately insured person.

26 | P a g e

30 | P a g e

Table 3.7 Logistic Regression Analysis – Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 2002-2007 Base

Model Base Model Plus Education

Base Model Plus Poverty

Base Model Plus Insurance

Full Model

Female -0.15** -0.13* -0.14** -0.15** -0.13** (0.07) (0.07) (0.07) (0.07) (0.07) Age 35-44 0.77*** 0.73*** 0.76*** 0.75*** 0.72*** (0.15) (0.15) (0.15) (0.15) (0.15) Age 45-54 1.06*** 1.03*** 1.05*** 1.05*** 1.01*** (0.12) (0.12) (0.12) (0.12) (0.12) Age 55-64 1.47*** 1.43*** 1.45*** 1.45*** 1.41*** (0.13) (0.13) (0.13) (0.13) (0.13) Age 65-74 1.81*** 1.80*** 1.82*** 1.77*** 1.75*** (0.13) (0.13) (0.13) (0.13) (0.14) Age 75+ 1.67*** 1.69*** 1.69*** 1.62*** 1.63*** (0.14) (0.14) (0.14) (0.15) (0.15) African American -0.49*** -0.45*** -0.49*** -0.49*** -0.45*** (0.11) (0.11) (0.11) (0.11) (0.11) Hispanic -0.60*** -0.52*** -0.60*** -0.57*** -0.50*** (0.12) (0.12) (0.12) (0.12) (0.12) Asian -0.19 -0.21 -0.2 -0.2 -0.22 (0.18) (0.17) (0.18) (0.18) (0.18) Other Race -0.81*** -0.77*** -0.79*** -0.80*** -0.77*** (0.25) (0.25) (0.25) (0.26) (0.26) Smoking status -0.18* -0.12 -0.16* -0.15 -0.11 (0.1) (0.09) (0.1) (0.09) (0.09) Exercise -0.06 -0.08 -0.06 -0.06 -0.08 (0.07) (0.07) (0.07) (0.07) (0.07) Underweight 0.85*** 0.87*** 0.85*** 0.85*** 0.87*** (0.11) (0.11) (0.11) (0.11) (0.11) Overweight -0.18** -0.17** -0.18** -0.18** -0.17** (0.08) (0.08) (0.08) (0.08) (0.08) Obesity -0.36*** -0.33*** -0.35*** -0.36*** -0.33*** (0.1) (0.1) (0.1) (0.1) (0.1) No High School -0.46*** -0.43*** (0.13) (0.13) Some High School -0.36*** -0.32*** (0.12) (0.12) College -0.21 -0.2

27 | P a g e

28 | P a g e

31 | P a g e

Table 3.7 Logistic Regression Analysis – Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 2002-2007 Base


Base Model Plus Poverty

Base Model Plus Insurance

Full Model

(0.13) (0.13) Low Income -0.12 -0.05 (0.09) (0.1) Middle Income -0.12 -0.07 (0.07) (0.07) MSA 0.11 0.12 0.08 (0.08) (0.08) (0.08) Public Insurance 0.01 0.06 (0.09) (0.09) Uninsured -0.35*** -0.29** (0.13) (0.13) Constant -4.69*** -4.39*** -4.74*** -4.75*** -4.42*** (0.12) (0.17) (0.14) (0.15) (0.19) Observations 72806 72806 72806 72806 72806 Source: Based on authors’ analysis of the Medical Expenditure Panel Survey 2002-2007 Coefficients with standard errors in parentheses. Exponentiate the coefficient to compute the odds ratio relative to the reference group, e.g., e -0.13 = 0.88. Stars denote level of statistical significance: * - p < 0.05, ** - p < 0.01, and *** - p < 0.001.

32 | P a g e

Table 3.8 Logistic Regression Analysis – Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education

Base Model Plus Income and Insurance

Full Model

Female -0.04 -0.02 -0.04 -0.03 (0.06) (0.06) (0.07) (0.07) Age 35-44 0.50*** 0.47*** 0.48*** 0.46*** (0.15) (0.15) (0.15) (0.15) Age 45-54 0.89*** 0.86*** 0.85*** 0.84*** (0.12) (0.12) (0.12) (0.12) Age 55-64 1.37*** 1.34*** 1.33*** 1.31*** (0.12) (0.12) (0.12) (0.12) Age 65-74 1.82*** 1.81*** 1.78*** 1.77*** (0.13) (0.13) (0.13) (0.13) Age 75+ 1.80*** 1.82*** 1.79*** 1.79*** (0.13) (0.13) (0.14) (0.14) African American -0.43*** -0.40*** -0.40*** -0.38*** (0.11) (0.11) (0.11) (0.11) Hispanic -0.72*** -0.67*** -0.64*** -0.61*** (0.12) (0.12) (0.12) (0.12) Asian -0.09 -0.09 -0.09 -0.09 (0.1) (0.1) (0.1) (0.1) Other Race -0.92*** -0.89*** -0.88*** -0.87*** (0.29) (0.29) (0.29) (0.29) No High School -0.38*** -0.32*** (0.12) (0.12) Some High School -0.37*** -0.33*** (0.11) (0.11) College -0.2 -0.19 (0.13) (0.13) Low Income -0.18* -0.13 (0.1) (0.1) Middle Income -0.05 -0.01 (0.07) (0.07) Public Insurance -0.05 -0.02 (0.09) (0.09) Uninsured -0.39*** -0.36*** (0.13) (0.13) MSA 0.09 0.07

29 | P a g e

Table 3.8 Logistic Regression Analysis – Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education


Full Model

(0.08) (0.08) Constant -4.86*** -4.55*** -4.83*** -4.57*** (0.11) (0.15) (0.13) (0.17) Observations 82889 82889 82889 82889 Source: Based on authors’ analysis of the Medical Expenditure Panel Survey, 1996-2007. Coefficients with standard errors in parentheses. Exponentiate the coefficient to compute the odds ratio relative to the reference group, e.g., e -0.13 = 0.88. Stars denote level of statistical significance: * - p < 0.1, ** - p < 0.05, and *** - p < 0.01.

Running additional models that included behavioral measures such as BMI, smoking, and exercise using data from the 2002-2007 MEPS further tested the robustness of our results. These variables were not used in previous models because they were not consistently reported in the MEPS before 2002. Similar results to those reported above using MEPS data from 1996-2007 were found in all five models, except for individuals with family income that is below 200% of the FPL, whereas the marginal significance found above disappeared. Additionally, any cancer prevalence was positively associated with underweight (BMI<18.5, OR= 2.33, P<0.001) and negatively associated with overweight (BMI between 25 and 29.9 kg/m2, OR= 0.83, P<0.05), and obesity (BMI>29.9, OR=0.70, P<0.001) compared to patients with normal weight (BMI between 18.5 and 24.9). Compared to normal weight persons, the odds of underweight persons having cancer were 2.3 times greater, while the odds for overweight and obese persons were 16% and 30% lower respectively. This may be an indication that persons with cancer tend to lose weight rather than underweight persons being at greater cancer risk. Colon, Breast, Lung, and Prostate Cancer

The characteristics associated with cancer prevalence were further explored by stratifying the sample by common types of cancer such as colon, breast, lung, and prostate cancer. (See tables 3.9 through 3.12.) MEPS data from 1996-2007 were used for the stratified analysis because of sample size. Colon cancer was positively associated with African Americans compared to white patients in all four models (P<0.05). Breast cancer was negatively associated with African-American females (significance levels range from P<0.10 to p<0.05) compared to white patients, but the significance disappeared in full models that controlled for education, income, and insurance status at the same time. Breast cancer was also negatively associated with public insurance and the uninsured (p<0.10) compared to private/employer insurance, but the significance level disappeared when controlling for different levels of education as well.

Lung cancer was positively associated with African-Americans compared to white patients, but the significance disappeared when controlling for education, income, and insurance status. Furthermore, lung cancer was positively associated with public insurance (p<0.001) compared to patients with private insurance, even after controlling for income and education. Lower income individuals with annual family income less than 200% of the FPL were less likely to have prostate cancer compared to those making above 400% of the FPL (P<0.10). However,

30 | P a g e

31 | P a g e

34 | P a g e

the associations indicated above disappeared in all the models that combined colon, lung, and breast cancer for women as well as colon, lung, and prostate cancer for men.

Table 3.9 Logistic Regression Analysis – Lung Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education


Full Model

Female -0.18 -0.22 -0.25 -0.29 (0.31) (0.31) (0.32) (0.32) Age 35-44 0.99 1.04 0.98 1.01 (1.14) (1.14) (1.14) (1.13) Age 45-54 1.47 1.51* 1.45 1.47 (0.91) (0.92) (0.93) (0.93) Age 55-64 3.21*** 3.26*** 3.12*** 3.16*** (0.76) (0.76) (0.77) (0.77) Age 65-74 3.89*** 3.90*** 3.36*** 3.40*** (0.74) (0.75) (0.76) (0.76) Age 75+ 3.50*** 3.47*** 2.85*** 2.89*** (0.77) (0.77) (0.8) (0.8) African American 0.83** 0.75* 0.59 0.55 (0.39) (0.39) (0.41) (0.41) Hispanic -0.78 -0.9 -0.98 -1.01 (0.76) (0.8) (0.78) (0.82) Asian 0.3 0.28 0.31 0.31 (0.51) (0.51) (0.52) (0.52) No High School 1.56 1.33 (1.06) (1.03) High School 1.44 1.37 (1.03) (1.02) College 1.25 1.22 (1.1) (1.1) Low Income 0.27 0.2 (0.43) (0.41) Middle Income -0.36 -0.43 (0.45) (0.44) Public Insurance 1.07*** 1.02*** (0.41) (0.39) Uninsured -1.04 -1.09 (1.06) (1.05) MSA (0.02 0.05 (0.39) (0.39) Constant -9.90*** -11.26*** -9.83*** -11.08*** (0.74) (1.25) (0.83) (1.3) Observations 80054 80054 80054 80054

32 | P a g e

36 | P a g e

Table 3.9 Logistic Regression Analysis – Lung Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education


Full Model

Source: Based on authors’ analysis of the Medical Expenditure Panel Survey, 1996-2007. Coefficients with standard errors in parentheses. Exponentiate the coefficient to compute the odds ratio relative to the reference group, e.g., e -0.13 = 0.88. Stars denote level of statistical significance: * - p < 0.1, ** - p < 0.05, and *** - p < 0.01.

33 | P a g e

37 | P a g e

Table 3.10 Logistic Regression Analysis – Colon Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education Base Model Plus Income and Insurance

Full Model

Female -0.73** -0.74** -0.71** -0.70** (0.31) (0.32) (0.31) (0.31) Age 35-44 0.68 0.68 0.76 0.75 (1.24) (1.24) (1.24) (1.24) Age 45-54 3.20*** 3.21*** 3.31*** 3.32*** (1.08) (1.08) (1.1) (1.1) Age 55-64 3.48*** 3.49*** 3.62*** 3.63*** (1.06) (1.06) (1.07) (1.07) Age 65-74 4.66*** 4.68*** 5.01*** 5.02*** (1.04) (1.04) (1.09) (1.09) Age 75+ 4.99*** 5.02*** 5.36*** 5.38*** (1.04) (1.04) (1.07) (1.07) African American 0.91** 0.93** 0.94** 0.96** (0.37) (0.37) (0.4) (0.39) Hispanic 0.72 0.75 0.67 0.69 (0.44) (0.47) (0.47) (0.49) Asian 0.61 0.62 0.59 0.59 (0.46) (0.45) (0.46) (0.45) Other Race 0.79 0.81 0.81 0.81 (1.03) (1.01) (1.03) (1.02) No High School 0.22 0.19 (0.65) (0.67) Some High School 0.26 0.21 (0.68) (0.69) College 0.49 0.48 (0.75) (0.75) Low Income -0.16 -0.15 (0.41) (0.4) Middle Income 0.26 0.26 (0.34) (0.34) Public insurance -0.24 -0.23 (0.38) (0.39) Uninsured 0.81 0.81 (0.52) (0.52) MSA -0.01 -0.02 (0.4) (0.4) Constant -10.71*** -11.00*** -11.04*** -11.28*** (1.0) (1.13) (1.09) (1.19)

34 | P a g e

35 | P a g e

38 | P a g e

Table 3.10 Logistic Regression Analysis – Colon Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education Base Model Plus Income and Insurance

Full Model

Observations 81653 81653 81653 81653 Source: Based on authors’ analysis of the Medical Expenditure Panel Survey, 1996-2007. Coefficients with standard errors in parentheses. Exponentiate the coefficient to compute the odds ratio relative to the reference group, e.g., e -0.13 = 0.88. Stars denote level of statistical significance: * - p < 0.1, ** - p < 0.05, and *** - p < 0.01.

Table 3.11 Logistic Regression Analysis – Breast Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education


Full Model

Age 35-44 2.56*** 2.53*** 2.51*** 2.49*** (0.59) (0.59) (0.59) (0.59) Age 45-54 3.35*** 3.34*** 3.29*** 3.29*** (0.62) (0.62) (0.61) (0.62) Age 55-64 3.85*** 3.85*** 3.81*** 3.82*** (0.61) (0.61) (0.61) (0.62) Age 65-74 4.04*** 4.08*** 4.08*** 4.11*** (0.64) (0.64) (0.65) (0.65) Age 75+ 3.32*** 3.40*** 3.44*** 3.49*** (0.66) (0.65) (0.66) (0.66) African American -0.68** -0.62* -0.55* -0.52 (0.32) (0.32) (0.32) (0.32) Hispanic -0.37 -0.25 -0.15 -0.09 (0.3) (0.3) (0.3) (0.3) Asian -0.07 -0.05 -0.06 -0.05 (0.26) (0.26) (0.26) (0.26) Other Race -1.54 -1.51 -1.47 -1.46 (0.96) (0.96) (0.96) (0.96) No High School -0.54 -0.39 (0.33) (0.35) Some High School -0.32 -0.26 -0.32 (0.32) College -0.03 -0.02 -0.34 (0.34) Low Income -0.4 -0.34 (0.27) (0.28) Middle Income 0.1 0.15 (0.2) (0.21) Public insurance -0.44* -0.39 (0.25) (0.26) Uninsured -0.89* -0.84* (0.47) (0.47) MSA 0.0 -0.03 (0.23) (0.23) Constant -8.53*** -8.27*** -8.40*** -8.21*** (0.59) (0.64) (0.63) (0.67)

36 | P a g e

37 | P a g e

40 | P a g e

Table 3.11 Logistic Regression Analysis – Breast Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base Model Base Model

Plus Education


Full Model

Observations 44202 44202 44202 44202 Source: Based on authors’ analysis of the Medical Expenditure Panel Survey, 1996-2007. Coefficients with standard errors in parentheses. Exponentiate the coefficient to compute the odds ratio relative to the reference group, e.g., e -0.13 = 0.88. Stars denote level of statistical significance: * - p < 0.1, ** - p < 0.05, and *** - p < 0.01.

Table 3.12 Logistic Regression Analysis – Prostate Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base


Base Model Plus Income, Insurance

Full Model

Age 45-54 11.81*** 11.77 11.79 11.77*** (0.51 (.) (.) (0.52) Age 55-64 13.76 13.70*** 13.77*** 13.72 (.) (0.51) (0.52) (.) Age 65-74 14.38*** 14.32*** 14.48*** 14.44*** (0.3) (0.5) (0.54) (0.31) Age 75+ 15.14*** 15.09*** 15.32*** 15.27*** (0.27) (0.5) (0.55) (0.27) African American -0.01 0 0.07 0.04 (0.33) (0.33) (0.33) (0.33) Hispanic -0.59 -0.61 -0.54 -0.63 (0.38) (0.41) (0.4) (0.42) Asian 0.24 0.23 0.25 0.23 (0.31) (0.31) (0.31) (0.31) Other Race -0.61 -0.62 -0.47 -0.53 (0.69) (0.69) (0.71) (0.71) No high School -0.18 0.05 (0.34) (0.36 Some High School -0.47 -0.37 (0.36) (0.36) College -0.03 0.01 (0.4) (0.4) Low Income -0.58* -0.59* (0.34) (0.35) Middle Income 0.03 0.06 (0.27) (0.28) Public Insurance -0.25 -0.25 (0.26) (0.27) Uninsured 0.08 0.08 (0.6) (0.61) MSA 0.66** 0.67** (0.28) (0.28) Constant -18.94*** -18.64*** -19.47*** -19.30*** (0.22) (0.57) (0.52) (0.39) Observations 29953 29953 29953 29953 Source: Based on authors’ analysis of the Medical Expenditure Panel Survey, 1996-2007.

38 | P a g e

39 | P a g e

42 | P a g e

Table 3.12 Logistic Regression Analysis – Prostate Cancer Prevalence, Demographic, Health Behavior and Socioeconomic Factors, 1996-2007 Base


Base Model Plus Income, Insurance

Full Model

Coefficients with standard errors in parentheses. Exponentiate the coefficient to compute the odds ratio relative to the reference group, e.g., e -0.13 = 0.88. Stars denote level of statistical significance: * - p < 0.1, ** - p < 0.05, and *** - p < 0.01.

Discussion

An analysis of the Medical Expenditure Panel Survey (MEPS) explored patterns of cancer prevalence comparing various racial/ethnic groups, socioeconomic status and health insurance status using multivariate regression analytic techniques with appropriate controls, including age, gender, and obesity. The results revealed seemingly counterintuitive findings. Traditionally disadvantaged groups tended to report lower prevalence rates of cancer (cancer for any site) diagnosis. There was a lower prevalence of all-site cancer diagnoses among less well-educated persons compared with more highly educated persons. The analysis also found lower prevalence of all-site cancer diagnosis among persons without health insurance compared to privately insured persons. There was also a lower prevalence of all-site cancer diagnosis among African-Americans and Hispanics compared to whites. Analyses of specific cancers reveal a somewhat more complex set of patterns. The analysis found: higher colorectal cancer prevalence among African-Americans compared to whites; higher lung cancer prevalence among those with public health insurance compared with privately insured; and lower prevalence of prostate cancer among lower income men. We caution against interpreting the findings as evidence of lower cancer risk among traditionally socially disadvantaged groups. The results may be reflecting lower cancer survival rates among disadvantaged populations, which have been well documented (Horner et al., 2009). That is, if socially disadvantaged groups have lower cancer survival rates, then members of socially disadvantaged groups who remain alive to respond to surveys such as MEPS are selected disproportionately from those without cancer. Moreover, cancer survival varies by the type of cancer. For example, for lung cancer, which is usually associated with a relatively short survival, incidence has a proportionally greater effect (relative to survival) on prevalence, compared to breast and prostate cancer which are associated with more prolonged survival. Additionally, it has been well documented that persons from socially disadvantaged groups tend to be diagnosed at later stages of disease (Marlow et al., 2010). MEPS data do not include data on stage at diagnosis, a major determinant of survival. Differences in healthcare access and utilization among socially disadvantaged groups are well known (Laiyemo et al., 2010; Carpenter et al., 2010). Differences in healthcare access and utilization can also influence self-reported survey data if persons from socially disadvantaged groups are more likely to be unaware of their cancer status. Survey respondents cannot report a cancer diagnosis if they have not received a diagnosis as a result of underutilization of health services and screening, or lower quality care (Onega et al., 2010).

40 | P a g e

43 | P a g e

IV. The Societal and Economic Impact of Cancer Health Disparities

The economic impact of cancer disparities was costly to the United States. We estimated the costs of cancer disparities across racial/ethnic groups and across socioeconomic groups. The annual costs of racial/ethnic disparities were $193 billion for premature death, $2.3 billion for direct medical costs and $471.5 million for lost productivity. The annual costs for disparities between counties at the poverty level were $36.7 billion for premature death, $236.9 million for direct medical costs and $42.5 million for lost productivity. The annual costs for disparities between counties at the poverty level were $35.9 billion for premature death, $344.5 million for direct medical costs and $77 million for lost productivity. This report’s estimates compared to NIH’s 2006 estimate of the economic costs of cancer: $124.8 billion for premature death, $99 billion for direct costs, and $19.6 billion for lost productivity (National Heart, Lung, and Blood Institute, 2009). In 2006, there were 559,233 deaths attributable to cancers, which equates to an estimated 9.25 million years of life lost. The NIH used a value of lifetime earnings methodology to estimate the economic costs of cancer mortality at $124.8 billion. However, this methodology would tend to under-value the cost of cancer mortality because 72% of cancer deaths occur in persons ages 55-74. These are persons at the ends of their careers or who are retired; therefore, the value of their remaining lifetime earnings will be relatively low compared to younger persons. Consequently, we believe a willingness-to-pay measure is a better indicator of the value of their lives than their potential wages. Such an estimate is based on society’s willingness to pay for medical interventions to prevent death, which considers more than just a person’s value in the labor market but also their intrinsic value to family, friends and the community at large. Using the cost effectiveness estimates of $50,000 to $264,000 per quality life year (Braithwaite, 2008), we estimate the cost of cancer mortality to be $462 billion to $2.44 trillion. The Cost of Premature Death

If society were able to eliminate racial disparities in cancer mortality, this analysis estimates there would have been 243,160 fewer cancer deaths (or 3.87 million fewer years of life lost) in 2006. The elimination of cancer disparities would have reduced the economic costs of cancer deaths by $193 billion to over $1 trillion in 2006; and over a five-year period (2002-2006) by $955 billion to $5.5 trillion in constant dollars. (See Table 4.1) Most of the costs saving are attributable to a reduction in cancer mortality for whites (80%) and African Americans (17%), who tend to have the higher death rates relative to Asians and Hispanics.

41 | P a g e

44 | P a g e

Cancer burden also varies by socioeconomic status. The costs of socioeconomic cancer disparities were estimated two ways: 1) by reducing disparities between counties with different poverty rates; and 2) by reducing disparities between counties with different levels of educational attainment. Cancer disparities associated with poverty results in over 734.7 thousand life years lost, annually. (See table 4.2) Eliminating cancer disparities between counties with different levels of poverty would result in a cost savings ranging from $36.7 billion to $194 billion annually and from $183.7 billion to almost a trillion dollars over a five-year period. Lung cancer was the most expensive, ranging from $13.8 billion to $73.6 billion annually. Prostate cancer was least expensive, ranging from low estimate of $1.6 billion to $8.3 billion.

Table 4.1. Cost of Premature Cancer Deaths Due to Race Disparities in Cancer Mortality in Billions of Dollars Year Number

of Life Years Lost (000s)

$50,000 per life-year



2006 3876 193 709 1023 2005 3774 188 690 996 2004 3835 192 702 1012 2003 3775 189 691 997 2002 3859 193 706 1019 Total 19120 955 3498 5047 Source: Based on authors’ calculation using cancer mortality data from Centers for Disease Control and Prevention National Vital Statistics System and life expectancy data from Arias. United States life tables, 2004. National vital statistics reports; vol. 56 no 9. Hyattsville, MD: National Center for Health Statistics. 2007. Life-year values based on recommendation in Braithwaite et al. “What Does the Value of Modern Medicine Say About the $50,000 per Quality-Adjusted Life-Year Decision Rule?” Medical Care 2008; 46: 349–356.

42 | P a g e

45 | P a g e

Table 4.2 Cost of Premature Cancer Deaths Due to Disparities in Cancer Mortality Across Counties with Different Poverty Rates in Billions of Dollars for All Cancers, Lung, Colorectal, Breast and Prostate Number of Life

Years Lost (000s)




All Cancer Annual 734.7 36.7 134.4 194.0 5 Year Total (2002-2006) 3,673.3 183.7 672.2 969.7 Lung Cancer Annual 276.9 13.8 50.7 73.1 5 Year Total (2002-2006) 1,384.6 69.2 253.3 365.5 Colorectal Cancer Annual 74.3 3.7 13.5 16.6 5 Year Total (2002-2006) 371.4 18.6 68.0 98.1 Breast Cancer Annual 45.9 2.3 8.4 12.1 5 Year Total (2002-2006) 229.3 11.5 42.0 60.5 Prostate Cancer Annual 31.1 1.6 5.7 8.3 5 Year Total (2002-2006) 155.8 7.8 28.5 41.1 Source: Based on authors’ calculation using cancer mortality data from Centers for Disease Control and Prevention National Vital Statistics System and life expectancy data from Arias United States life tables, 2004. National vital statistics reports; vol. 56 no 9. Hyattsville, MD: National Center for Health Statistics. 2007. Life-year values based on recommendation in Braithwaite et al. “What Does the Value of Modern Medicine Say About the $50,000 per Quality-Adjusted Life-Year Decision Rule?” Medical Care 2008; 46: 349–356.

The costs of premature death associated with disparities in cancer mortality across counties with different levels of educational attainment ranged from $35.9 billion to $189.5 billion. These disparities were associated with almost 718 thousand years of life lost. The five-year costs (2002-2006) ranged from $179.5 billion to $947.7 billion. Lung cancer was most expensive, followed by colorectal, breast and prostate cancer. (See table 4.3.)

43 | P a g e

46 | P a g e

Table 4.3 Cost of Premature Cancer Deaths Due to Educational Attainment Disparities in Cancer Mortality in Billions of Dollars for All Cancers, and Lung, Colorectal, Breast and Prostate Cancers Reduce Mortality to Lowest Education Counties

Number of Life Years Lost (000s)




All Cancer Annual 718.0 35.9 131.4 189.5 5 Year Total (2002-2006) 3,589.8 179.5 656.9 947.7 Lung Cancer Annual 301.8 15.1 55.2 79.7 5 Year Total (2002-2006) 1509.1 75.4 276.2 398.4 Colorectal Cancer Annual 80.0 4.0 1460 21.1 5 Year Total (2002-2006) 400.0 20.0 73.2 105.6 Breast Cancer Annual 49.1 2.5 9.0 13.0 5 Year Total (2002-2006) 245.6 12.3 44.9 64.8 Prostate Cancer Annual 23.1 1.2 4.2 6.1 5 Year Total (2002-2006) 115.5 5.8 21.1 30.5 Source: Based on authors’ calculation using cancer mortality data from Centers for Disease Control and Prevention National Vital Statistics System and life expectancy data from Arias United States life tables, 2004. National vital statistics reports; vol. 56 no 9. Hyattsville, MD: National Center for Health Statistics. 2007. Life-year values based on recommendation in Braithwaite et al. “What Does the Value of Modern Medicine Say About the $50,000 per Quality-Adjusted Life-Year Decision Rule?” Medical Care 2008; 46: 349–356.

The Direct Cost of Cancer Disparities The actual and incremental costs of medical care were computed for cancer patients using

data from the 2002-2007 Medical Expenditure Panel Survey. (See table 4.4.) The actual average medical costs for persons with cancer ranged from $9,830 for persons with any type of cancer to $21,147 for persons with lung cancer. This is substantially higher than the average medical costs of $4,157 for persons without cancer. After adjusting for demographic characteristics, socioeconomic factors, and other health conditions, the incremental cost of cancer was $4,347. Cost estimates for the four most common cancers ranged from $1,784 for prostate cancer to $9,027 for lung cancer.

44 | P a g e

Table 4.4 The actual and incremental costs of cancer, for all cancers and selected cancer sites, 2002-2007 Average

Costs Incremental Costs

95% Confidence Interval

Any Cancer $ 9,830 $4,347 $4,270 - $4,424 Lung $21,147 $9,027 $8,859 - $9,194 Colorectal $15,243 $8,236 $8,082 - $8,389 Breast $13,657 $7,946 $7,801 - $8,091 Prostate $12,297 $1,784 $1,753 - $1,815 Source: Based on authors’ calculations using the MEPS 2002-2007.

To compute the direct costs of cancer disparities, the number of cases due to disparities in cancer incidence was estimated. (See table 4.5.) For 2002-2006, there were on average 1,334,544 new cancer cases a year. If racial disparities in cancer incidence were eliminated, the annual number of new cancer cases would decline by 531,983. This reduction would result in $2.3 billion savings on annual medical expenditures for persons with cancer. The annual cost savings associated with the elimination of racial disparities in cancer incidence for the four most common cancers ranged from $166.8 million for prostate cancer to $814 million for lung cancer. Table 4.5 The Reduction in Medical Care Associated with an Elimination of Racial/Ethnic Disparities in Cancer Incidence. Incremental

Cost Average Number of Cancer Cases

Reduction in Cancer Cases

Reduction in Medical Costs (millions)

All Cancer $ 4,347 1,334,544 531,983 $2,312.7 Lung $9,027 190,687 90,253 $814.7 Colorectal $8,236 142,839 47,666 $392.5 Breast $7,946 184201 84,838 $674.1 Prostate $1,784 189838 93,487 $166.7 Source: Based on Calculations by the Authors Using Data from the Medical Expenditure Panel Survey 2002-2007 and Cancer incidence data for 2002-2006 from the U.S. Cancer Statistics Working Group. United States Cancer Statistics: 1999–2006 Incidence and Mortality Web-based Report. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2010. Available at: www.cdc.gov/uscs

Incidence data from SEER/STAT 2002-2006 were used to calculate the reduction in new cancer cases to compute the cost savings associated with the elimination of socioeconomic cancer disparities. This incidence data is only for the 17 geographic areas covered by SEER. It represents 26.2% of the U.S. population. The cost savings for this population were computed and then extrapolated to the nation by multiplying by 3.817 (the reciprocal of 0.262). This analysis looked at: 1) the elimination of disparities between counties categorized by poverty

45 | P a g e

status; and 2) the elimination of disparities between counties categorized by educational attainment. (See table 4.6.) The elimination of poverty disparities in cancer incidence would reduce the number of new cancer cases by 14,278 in the SEER 17 geographic areas. This change in the annual number of cancer cases would lower health expenditures by $62.1 in the SEER 17 geographic areas and by $236 million nationally. Among the four most common cancers, the cost savings associated with the reductions in colorectal cancer were $228.7 million nationally. Eliminating educational disparities in cancer incidence would result in a reduction of 20,763 cancer cases in SEER 17. This decline in new cancer cases corresponds to a $90.3 million reduction in healthcare costs in SEER 17 and an estimated $344.5 million nationally. Table 4.6 The Reduction in Medical Care Costs Associated with an Elimination of Socioeconomic Disparities in Cancer Incidence. Incremental



Reduction in Medical Cost for SEER 17

Reduction in Medical Costs of US (millions)

Poverty Disparities All Cancer $4,347 338,210 14,278 $62.1 $236.9 Lung $9,027 45,277 1,980 $17.9 $68.2 Colorectal $8,236 35,673 7,278 $59.9 $228.7 Breast $7,946 49,551 5,023 $39.9 $152.3 Prostate $1,784 51,057 4,596 $8.2 $31.3 Education Disparities All Cancer $4,347 338,210 20,763 $90.3 $344.5 Lung $9,027 45,277 4,802 $43.3 $165.4 Colorectal $8,236 35,673 1,927 $15.9 $60.6 Breast $7,946 49,551 3,971 $31.6 $120.4 Prostate $1,784 51,057 6,066 $10.8 $41.3 Source: Based on Calculations by the Authors Using Data from the Medical Expenditure Panel Survey 2002-2007 and cancer incidence data for 2002-2006 from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program.

The Indirect Cost of Cancer Disparities

Elimination of disparities in cancer incidence would also result in cost savings attributable to improved labor market outcomes, i.e., fewer disability days, more hours worked and higher wages for working age adults. The 2002-2007 MEPS was used to compute the incremental indirect costs of cancer. These estimates are based on three regression models to predict disability days, hours worked and hourly wages. These models were then used to estimate the change in these labor market outcomes associated with having any cancer, and separately for lung, colorectal, breast and prostate cancers. The incremental indirect cost for any cancer was $1,972. The incremental indirect costs for the four most common cancers ranged from $2,291 for persons with prostate cancer to $3,947 for persons with colorectal cancer. The reduction in

46 | P a g e

49 | P a g e

productivity costs for any cancer was $471.5 million. Breast and lung cancer had the greatest reductions in productivity costs at about $116 million each. (See table 4.7.) Table 4.7 The Reduction in Indirect Care Associated with an Elimination of Racial/Ethnic Disparities in Cancer Incidence. Incremental



Reduction in Productivity Costs (millions)

All Cancer $1,972 574,865 239,089 $471.5 Lung $3,404 62,660 34,104 $116.1 Colorectal $3,947 50,334 14,668 $57.9 Breast $2,390 108,233 48,869 $116.8 Prostate $2,291 72,154 44,792 $102.6 Source: Based on Calculations by the Authors Using Data from the Medical Expenditure Panel Survey 2002-2007 and Cancer incidence data for 2002-2006 from the U.S. Cancer Statistics Working Group. United States Cancer Statistics: 1999–2006 Incidence and Mortality Web-based Report. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; 2010. Available at: www.cdc.gov/uscs

To estimate the potential cost saving associated with eliminating SES disparities in

cancer, we compared cancer rates and their associated costs between counties with high poverty rates against counties with low poverty rates. (See table 4.8.) Similar analysis was conducted for counties with high and low levels of educational attainment. An elimination of cancer disparities among counties with various level of poverty would reduce costs related to lost productivity by $42.5 million for any cancer and $24.4 million and $23.7 million for breast and prostate cancer respectively. The elimination of cancer disparities among counties with different percentages of adults with low levels of educational attainment would reduce productivity losses by $77 million for any cancer, and $35.1 million for prostate cancer, followed by $22.7 million for lung cancer.

47 | P a g e

50 | P a g e

Table 4.8 The Reduction in Indirect Cancer Costs Associated with an Elimination of Socioeconomic Disparities in Cancer Incidence Incremental

Cost Average Number of New Cancer Cases

Reduction in New Cancer Cases

Reduction in Medical Cost for SEER 17

Reduction in Medical Costs of US (millions)

Poverty Disparities All Cancer $1972 147,970 5,652 $11.1 $42.5 Lung $3404 14,380 1,078 $3.7 $14.0 Colorectal $3947 12,606 706 $2.8 $10.6 Breast $2390 29,058 2,678 $6.4 $24.4 Prostate $2291 19,553 2,705 $6.2 $23.7 Education Disparities All Cancer $1972 147,970 10,228 $20.2 $77.0 Lung $3404 14,380 1745 $5.9 $22.7 Colorectal $3947 12,606 840 $3.3 $12.7 Breast $2390 29,058 2314 $5.5 $21.1 Prostate $2291 19553 4017 $9.2 $35.1 Source: Based on Calculations by the Authors Using Data from the Medical Expenditure Panel Survey 2002-2007 and cancer incidence data for 2002-2006 from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program.

48 | P a g e

51 | P a g e

Recommendations Racial and ethnic disparities and socioeconomic disparities in cancer persist and are costly, not only in the monetary sense but also in losses that cannot be measured economically. What can be done? 1. Society needs to focus on cancer prevention and cancer treatment that result in long-

term survival. Programs designed to lower individual risk of cancer, if they are successful in reducing incidence and mortality among whites and African Americans to levels experienced by Asians and Hispanics, would yield substantial economic value to the nation. Similarly, if cancer mortality for persons with low socioeconomic status could be lowered to mortality rates experienced by more affluent and better educated persons, substantial costs due to premature death could be avoided.

2. The nation should invest in policies and programs designed to reduce cancer risks

associated with environmental factors and unhealthy behaviors. Regardless of socioeconomic status, individuals from racial and ethnic minority populations experience a myriad of stressors in the context of their social location. The effects of social location are mediated by lifetime exposure to social and environment factors, as well as individually modifiable behavioral risks. Such social and individual-level exposures translate into differential health care quality and access.

3. Policymakers, providers, and communities need to develop strategies to reduce racial

disparities in the quality of cancer-related care that also address the need for care coordination to promote rational utilization of preventive and therapeutic care, including management of comorbid conditions. The available evidence points to opportunities to intervene to reduce disparities in cancer-related care through health risk management, access to a usual source of care, access to primary and secondary treatment, and adjuvant therapies. However, the evidence also suggests that despite health care coverage, many members of racial and ethnic minority populations underutilize services for cancer detection, diagnosis and treatment.

4. Society should address modifiable risk factors such as physical inactivity, obesity, heavy

alcohol consumption, a diet high in red or processed meats, a diet that lacks fruits and vegetables, a diet high in animal fat, smoking, and environmental hazards such as radon, asbestos, air pollution, and certain metals (chromium, cadmium, arsenic) (ACS, 2009). Racial disparities in health care outcomes have many causes. They occur within a broader context and a social/political environment that produces systemic racial/ethnic inequalities. The fundamental social causes of such disparities (e.g., poverty, residential segregation) cannot be addressed absent the political will to implement appropriately targeted social policies. This case statement underscores the urgent need for intervention to reduce cancer disparities.

5. Community-level interventions are warranted. At the community level, such interventions

might address community norms and lifestyles that promote adverse health outcomes in racial/ethnic minority populations. Community-level interventions should also address

49 | P a g e

52 | P a g e

potential adverse effects of the physical and social environment, as well community differences in health care availability.

6. Strategies to reduce race disparities in the quality of cancer-related care must also address the need for care coordination to promote rational utilization of preventive and therapeutic care, including management of comorbid conditions. The available evidence points to opportunities to intervene to reduce disparities in cancer-related care, through health risk management, access to a usual source of care, and access to primary and secondary treatment, and adjuvant therapies. However, the evidence also suggests that despite health care coverage many members of racial and ethnic minority populations underutilize services for cancer detection, diagnosis and treatment.

7. Finally, we need better data. Cancer registries should collect socioeconomic data (income,

education and health insurance status). To make use of existing cancer registry data to compute survival and mortality rates, cancer registries should be linked to social security records. This will permit the computation of trends in cancer incidence, survival and mortality. Population surveys (e.g. the National Health Interview Survey) should develop cancer supplements to collect stage and survival information. Also, information on health behaviors could be collected using a national survey of new cancer patients based on a sample of new cancer registrants.

Any efforts to address the disparities in cancer health must address not only the monetary costs of such disparities, but also the costs that cannot be translated into dollars and cents.

50 | P a g e

53 | P a g e

References AHRQ National Healthcare Disparities Report, 2008 http://www.ahrq.gov/qual/nhdr08/Chap2.htm Albain, K.S., et al. (2009). Racial disparities in cancer survival among randomized clinical trials patients of the Southwest Oncology Group. J Gen Intern Med., 101(14), 984-92. Epub 2009 Jul 7. Albano, J.D. et al. (2007). Cancer mortality in the United States by education level and race. J Natl Cancer Inst., 99(18), 1384-94. Epub 2007 Sep 11. American Cancer Society. (2008). Cancer Facts and Figures 2008. Atlanta: American Cancer Society. American Cancer Society. (2009). Cancer Facts and Figures 2009. Atlanta: American Cancer Society. Arias, E. (2004). United States life tables, 2004. National vital statistics reports, 56(9). Hyattsville, MD: National Center for Health Statistics. 2007. Bach, P.B. et al. (1999). Racial differences in the treatment of early-stage lung cancer. N Engl J Med., 341(16):1198-205. Barron, J.J., Quimbo, R., Nikam, P.T., Amonkar, M.M. (2008). Assessing the economic burden of breast cancer in a US managed care population. Breast Cancer Res Treat., 109(2), 367-77. Epub 2007 Aug 3. Blackstock, A.W. et al. (2002). Outcomes among African-American/non-African-American patients with advanced non-small-cell lung carcinoma: report from the Cancer and Leukemia Group B. J Natl Cancer Inst., 94(4):284-90. Book Chapter: Disparities in Tobacco Use and Lung Cancer http://www.springerlink.com/content/gh8k745w78m73467/ Howard K. Koh2 Contact Information, Loris Elqura3 and Sarah Massin Short. Bradley, C.J. et al. (2001). Disparities in cancer diagnosis and survival. Cancer, 91(1),178-88. Bradley, C.J., Yabroff, K.R., Dahman, B., Feuer, E.J., Mariotto, A., Brown, M.L. (2008). Productivity costs of cancer mortality in the United States: 2000-2020. J Natl Cancer Inst., 100(24),1763-70. Epub 2008 Dec 9. Braithwaite, R.S., Meltzer, D.O., King, J.T., Leslie, D., Roberts, M.S. (2008). What Does the Value of Modern Medicine Say About the $50,000 per Quality-Adjusted Life-Year Decision Rule? Medical Care, 46, 349–356.

51 | P a g e

54 | P a g e

Bryant, A.S. et al. (2008). Impact of race on outcomes of patients with non-small cell lung cancer. J Thorac Oncol., 3(7), 711-5. Buntin, M.B., Zaslavsky, A.M. (2004). Too much ado about two-part models and transformation? Comparing methods of modeling health care costs. J Health Econ, 23(3), 525-542. Campbell, J.D., Ramsey, S.D. (2009). The costs of treating breast cancer in the US: a synthesis of published evidence. Pharmacoeconomics, 27(3),199-209. doi: 10.2165/00019053-200927030-00003. Review. Carpenter, W.R., Howard, D.L., Taylor, Y.J., Ross, L.E., Wobker, S.E., Godley, P.A. (2010). “Racial differences in PSA screening interval and stage at diagnosis.” Cancer Causes Control. Cheng, I. et al. (2009). Socioeconomic status and prostate cancer incidence and mortality rates among the diverse population of California. Cancer Causes Control. [Epub ahead of print] Clegg, L.X. et al. (2002). Cancer survival among US whites and minorities: a SEER (Surveillance, Epidemiology, and End Results) Program population-based study. Arch Intern Med., 162(17), 1985-93. Collins, F.S. (2004). What we do and don’t know about ‘race’, ‘ethnicity’, genetics and health at the dawn of the genome era. Nature Genetics Supplement, 36(11),S13-S15. Deb P, Manning W, Norton E. “Modeling Health Care Costs and Counts,” presentation at ASHE-Madison Conference, 2006. DeNavas-Walt, C., Proctor, B.D., Smith, J.C. (2008). US Census Bureau, Current Population Reports, P60-236, Income, Poverty and Health Insurance Coverage in the United States: 2008, US Government Printing Office, Washington, DC 2009. Doubeni, C.A., Laiyemo, A.O., Klabunde, C.N., Young, A.C., Field, T.S., Fletcher, R,H. (2010). Racial and ethnic trends of colorectal cancer screening among Medicare enrollees. Am J Prev Med., 38(2),184-91. Du, X.L. et al. (2006). Racial disparity and socioeconomic status in association with survival in older men with local/regional stage prostate carcinoma: findings from a large community-based cohort. Cancer, 106(6), 1276-85. Du, X.L., Fang, S., Vernon, S.W., El-Serag, H., Shih, Y.T, Davila, J., Rasmus, M.L. (2007). Racial disparities and socioeconomic status in association with survival in a large population-based cohort of elderly patients with colon cancer. Cancer, 110(3), 660-9. Du XL, Fang S. Meyer TE. (2008). Impact of treatment and socioeconomic status on racial disparities in survival among older women with breast cancer. Am J Clin Oncol. Apr;31(2):125-32.

52 | P a g e

55 | P a g e

Esnola, .NF. et al. (2008). Age-, race-, and ethnicity-related differences in the treatment of nonmetastatic rectal cancer: a patterns of care study from the national cancer data base. Ann Surg Oncol., 15(11), 3036-47. Epub 2008 Aug 19. Hackbarth, D.P. et al. (2001). Collaborative research and action to control the geographic placement of outdoor advertising of alcohol and tobacco products in Chicago. Public Health Rep., 116(6), 558–567. Hernandez, L., Easton, C.A, Fairlee, A.M., Chun, T.H., Spirito, A. (2010). Ethnic group differences in substance use, depression, peer relationships, and parenting among adolescents receiving brief alcohol counseling. J Ethn Subst Abuse,9(1),14-27. Hirth, R.A., Chernew, M.E., Miller, E. et al. Willingness to pay for a quality-adjusted life year: in search of a standard. Med Decis Making, 20, 332–342. Hoffman, R.M. et al. (2001). Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst., 93(5), 388-95. Holmes, L. Jr. et al. (2001). Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst., 93(5), 388-95. Horner, M.J., Ries, L.A.G., Krapcho, M. et al: SEER Cancer Statistics Review, 1975–2006, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov.ezproxy.welch.jhmi.edu/csr/1975_2006/ based on November 2008 SEER data submission, posted to the SEER website 2009. Jerant, A.F., Franks, P., Jackson, J.E., Doescher, M.P. (2004). Age-related disparities in cancer screening: analysis of 2001 Behavioral Risk Factor Surveillance System data. Ann Fam Med., 2(5), 481-7. Kutikova, L., Bowman, L., Chang, S., Long, S.R., Obasaju, C., Crown, W.H. (2005). The economic burden of lung cancer and the associated costs of treatment failure in the United States. Lung Cancer 50(2),143-54. Epub 2005 Aug 19. Laiyemo, A.O., Doubeni, C., Pinsky, P.F., Doria-Rose, V.P., Bresalier, R., Lamerato, L.E., Crawford, E.D., Kvale, P., Fouad, M., Hickey, T., Riley, T., Weissfeld, J., Schoen, R.E., Marcus, P.M., Prorok, P.C., Berg, C.D. (2010). Race and Colorectal Cancer Disparities: Health-Care Utilization vs. Different Cancer Susceptibilities. J Natl Cancer Inst., Apr 6. Manning, W.G., Mullahy, J. (2001). Estimating Log Models: To Transform or Not to Transform? J Health Econ 2001, 20(4), 461-494. Marlow, N.M., Halpern, M.T., Pavluck, A.L., Ward, E.M., Chen, A.Y. (2010). Disparities associated with advanced prostate cancer stage at diagnosis. J Health Care Poor Underserved, 21(1),112-31.

53 | P a g e

56 | P a g e

Morris, A.M. et al. (2009). Residual treatment disparities after oncology referral for rectal cancer. J Natl Cancer Inst., 100(10), 738-44. Epub 2008 May 13. Mulligan, C.R. et al. (2004). Racial disparities in rectal cancer treatment: a population-based analysis. Arch Surg., 139(2),151-5, discussion 156. National Heart, Lung, and Blood Institute, Fact Book Fiscal Year 2008. (2009). National Institutes of Health. Neumann, P.J., Sandberg, E.A., Bell, C.M., et al. Are pharmaceuticals cost effective? A review of the evidence. (2000). Health Aff (Millwood), 19, 92–109. Onega, T., Duell, E.J., Shi, X., Demidenko, E., Goodman, D.C. (2010). Race versus place of service in mortality among Medicare beneficiaries with cancer. Cancer. 2010 Mar 22. [Epub ahead of print] Penberthy, L., Retchin, S.M., McDonald, M.K., McClish, D.K., Desch, C.E., Riley, G.F., Smith, T.J., Hillner, BE., Newschaffer, C.J. (1999). Predictors of Medicare costs in elderly beneficiaries with breast, colorectal, lung, or prostate cancer. Health Care Manag Sci, 2(3), 149-60. Shavers, V.L. et al. (2004). Racial influence on biochemical disease-free survival in men treated with external-beam radiotherapy for localized prostate cancer. J Natl Med Assoc., 96(7), 939-44. Shavers, V.L., Brown, M., Klabunde, C.N., Potosky, A.L., Davis, W., Moul, J., Fahe, A. (2004). Race/ethnicity and the intensity of medical monitoring under 'watchful waiting' for prostate cancer. Med Care., 42(3), 239-50. Shugarman, L.R., Mack, K., Sorbero, M.E., Tian, H., Jain, A.K., Ashwood, J.S., Asch, S.M. (2009). Race and sex differences in the receipt of timely and appropriate lung cancer treatment. Med Care, 47(7), 774-81. Swanson, M. et al. (2009). Cancer epidemiology in the United States: racial, social, and economic factors. Methods Mol Biol., 471,65-83. Taylor, E.A. (2005). Economic Costs of Cancer Health Disparities Summary of Meeting Proceedings, Center to Reduce Cancer Health Disparities. US Department of Health and Human Services National Institutes of Health. Thompson, I. et al. (2003). Colorectal cancer screening among African Americans: the importance of physician recommendation. J Natl Med Assoc., 95(9), 806-12. White, A. et al. (2002). Prostate cancer education in the Washington, DC, area. J Natl Med Assoc., 94(11), 963-70. Wright, G.E., Barlow, W.E., Green, P., Baldwin, L.M., Taplin, S.H. (2007). Differences among the elderly in the treatment costs of colorectal cancer: how important is race? Med Care, 45(5), 420-30.

54 | P a g e

57 | P a g e

Xiao, H. et al. (2007). Analysis of prostate cancer incidence using geographic information system and multilevel modeling. J Natl Med Assoc., 99(3), 218-25. Yabroff, K.R., Warren, J.L., Brown, M.L. (2007). Costs of cancer care in the USA: a descriptive review. Nat Clin Pract Oncol., 4(11), 643-56. Yabroff, K.R., Warren, J.L., Schrag, D., Mariotto, A., Meekins, A., Topor, M., Brown, M.L. (2009). Comparison of approaches for estimating incidence costs of care for colorectal cancer patients. Med Care 47(7 Suppl 1), S56-63.

55 | P a g e

58 | P a g e

Cancer Disparities Publications from 2000-2010 Adams, S.A. et al. (2006). Breast cancer disparities in South Carolina: early detection, special programs, and descriptive epidemiology. J S C Med Assoc., 102(7), 231-9. Advani, A.S. et al. (2003). Barriers to the participation of African-American patients with cancer in clinical trials: a pilot study. Cancer, 97(6), 1499-506. Albain, K.S., et al. (2009). Racial disparities in cancer survival among randomized clinical trials patients of the Southwest Oncology Group. J Gen Intern Med., 101(14), 984-92. Epub 2009 Jul 7. Albano, J.D. et al. (2007). Cancer mortality in the United States by education level and race. J Natl Cancer Inst., 99(18),1384-94. Epub 2007 Sep 11. Alderman, A.K. et al. (2009). Racial and ethnic disparities in the use of postmastectomy breast reconstruction: results from a population- based study. J Clin Oncol., 27(32), 5325-30. Epub 2009 Oct 5. Arozullah, A.M., Calhoun, E.A., Wolf, M., Finley, D.K., Fitzner, K.A., Heckinger, E.A., Gorby, N.S., Schumock, G.T., Bennett, C.L. (2004). The financial burden of cancer: estimates from a study of insured women with breast cancer. J Support Oncol., 2(3), 271-8. Ashing-Giwa, K. et al. (1999). Quality of life of African-American and white long term breast carcinoma survivors. Cancer, 85(2), 418-26. Ashing-Giwa, K.T. et al. (2004). Understanding the breast cancer experience of women: a qualitative study of African American, Asian American, Latina and Caucasian cancer survivors. Psychooncology, 13(6),408-28. Bach, .PB. et al. (1999). Racial differences in the treatment of early-stage lung cancer. N Engl J Med., 341(16),1198-205. Baquet, C.R. et al. (2009). Analysis of Maryland Cancer Patient Participation in National Cancer Institute–Supported Cancer Treatment Clinical Trials. Journal of Health Care for the Poor and Underserved, 20, 120–134. Baquet, C.R. et al. (2000). Esophageal cancer epidemiology in blacks and whites: racial and gender disparities in incidence, mortality, survival rates and histology. Cancer, 88(S5), 1256 – 1264. Baquet, C.R. et al. (2000). Socioeconomic factors and breast carcinoma in multicultural women. Cancer, 88(5 Suppl),1256-64. Barlow, W.E. (2009). Overview of methods to estimate the medical costs of cancer. Med Care, 47(7 Suppl 1),S33-6.

56 | P a g e

59 | P a g e

Barron, J.J., Quimbo, R., Nikam, P.T., Amonkar, M,M. (2008). Assessing the economic burden of breast cancer in a US managed care population. Breast Cancer Res Treat., 109(2), 367-77. Epub 2007 Aug 3. Behbakht, K. et al. (2001). Characteristics and survival of cervical cancer patients managed at adjacent urban public and university medical centers. Gynecol Oncol., 81(1),40-6. Berry, J., Caplan, L., Davis, S., Minor, P., Counts-Spriggs, M., Glover, R., Ogunlade, V., Bumpers, K., Kauh, J., Brawley, O.W., Flowers, C. (2009). A black-white comparison of the quality of stage-specific colon cancer treatment. Cancer, Nov 30. [Epub ahead of print] Blackstock, A.W. et al. (2002). Outcomes among African-American/non-African-American patients with advanced non-small-cell lung carcinoma: report from the Cancer and Leukemia Group B. J Natl Cancer Inst., 94(4), 284-90. Bonham, V.L., Knerr, S. (2008). Social and ethical implications of genomics, race, ethnicity, and health inequities. Semin Oncol Nurs., 24(4), 254-61. Bradley, C.J., Neumark, D., Oberst, K., Luo, Z., Brennan, S., Schenk, M. (2005). Combining registry, primary, and secondary data sources to identify the impact of cancer on labor market outcomes. Med Decis Making, 25(5), 534-47. Bradley, C.J. et al. (2001). Disparities in cancer diagnosis and survival. Cancer, 91(1), 178-88. Bradley, C.J., Yabroff, K.R., Dahman, B., Feuer, E.J., Mariotto, A., Brown, M.L. (2008). Productivity costs of cancer mortality in the United States: 2000-2020. J Natl Cancer Inst., 100(24),1763-70. Epub 2008 Dec 9. Breslin, T.M. et al. (2009). Hospital factors and racial disparities in mortality after surgery for breast and colon cancer. J Clin Oncol., 27(24), 3945-50. Epub 2009 May 26. Brookfield, K.F. et al. (2009). Survival disparities among African American women with invasive bladder cancer in Florida. Cancer, 115(18), 4196-209. Brown, M.L., Riley, G.F., Potosky, A.L., Etzioni, R.D. (1999). Obtaining long-term disease specific costs of care: application to Medicare enrollees diagnosed with colorectal cancer. Med Care, 37(12),1249-59. Bryant, A.S. et al. (2008). Impact of race on outcomes of patients with non-small cell lung cancer. J Thorac Oncol., 3(7), 711-5. Byers, T.E. et al. (2008). 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, 113(3), 582-91. Byers, T.E. (2010). Two Decades of Declining Cancer Mortality: Progress with Disparity. Annu Rev Public Health. 31:121-32.

57 | P a g e

Campbell, J.D., Ramsey, S.D. (2009). The costs of treating breast cancer in the US: a synthesis of published evidence. Pharmacoeconomics, 27(3), 199-209. doi: 10.2165/00019053-200927030-00003. Carpenter, W.R., Howard, D.L., Taylor, Y.J., Ross, L.E., Wobker, S.E., Godley, P.A. (2010). Racial differences in PSA screening interval and stage at diagnosis. Cancer Causes Control, Mar 24. Chen, L.M. et al. (2009). Matched-pair analysis of race or ethnicity in outcomes of head and neck cancer patients receiving similar multidisciplinary care. Cancer Prev Res, 2(9), 782-91. Epub 2009 Sep 8. Cheng, I. et al. (2009). Socioeconomic status and prostate cancer incidence and mortality rates among the diverse population of California. Cancer Causes Control, Jun 13. [Epub ahead of print] Christie, J. et al. (2005). Predictors of endoscopy in minority women. J Natl Med Assoc., 97(10), 1361-8. Chu, Q.D. et al. (2009). Race/Ethnicity has no effect on outcome for breast cancer patients treated at an academic center with a public hospital. Cancer Epidemiol Biomarkers Prev., 18(8), 2157-61. Epub 2009 Jul 21. Clegg, L.X. et al. (2002). Cancer survival among US whites and minorities: a SEER (Surveillance, Epidemiology, and End Results) Program population-based study. Arch Intern Med., 162(17), 1985-93. Clegg, L.X., Reichman, M.E., Miller, B.A., Hankey, B.F., Singh, G.K., Lin, Y.D., Goodman, M.T., Lynch, C.F., Schwartz, S.M., Chen, V.W., Bernstein, L., Gomez, S.L., Graff, J.J., Lin, C.C., Johnson, N.J., Edwards, B.K. (2009). 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, 20(4), 417-35. Epub 2008 Nov 12. Cohen, J.W., Krauss, N.A. (2003). Spending and Service Use among People with the Fifteen Most Costly Medical Conditions, 1997, Health Affairs, 22(2), 129–138. Crawley, L.M. et al. (2008). Perceived medical discrimination and cancer screening behaviors of racial and ethnic minority adults. Cancer Epidemiol Biomarkers Prev., 17(8), 1937-44. Epub 2008 Aug 6. DeLancey, J.O. et al. (2008). Recent trends in Black-White disparities in cancer mortality. Cancer Epidemiol Biomarkers Prev., 17(11), 2908-12. Dimou, A. et al. (2009). Disparities in colorectal cancer in African-Americans vs. Whites: before and after diagnosis. World J Gastroenterol., 15(30), 3734-43.

58 | P a g e

61 | P a g e

Doubeni, C.A., Laiyemo, A.O., Klabunde, C.N., Young, A.C., Field, T.S., Fletcher, R.H. (2010). Racial and ethnic trends of colorectal cancer screening among Medicare enrollees. Am J Prev Med., 38(2), 184-91. Du, X.L. et al. (2006). Racial disparity and socioeconomic status in association with survival in older men with local/regional stage prostate carcinoma: findings from a large community-based cohort. Cancer, 106(6), 1276-85. Du, X.L., Fang, S., Vernon, S.W., El-Serag, H., Shih, Y.T., Davila, J., Rasmus, M.L. (2007). Racial disparities and socioeconomic status in association with survival in a large population-based cohort of elderly patients with colon cancer. Cancer,110(3), 660-9. Eisert, S.L. et al. (2008). Can America's urban safety net systems be a solution to unequal treatment? J Urban Health, 85(5), 766-78. Epub 2008 Jun 14. Ell, K. et al. (2008). Economic stress among low-income women with cancer: effects on quality of life. Cancer, 112(3), 616-25. Esnola, N.F. et al. (2008). Age-, race-, and ethnicity-related differences in the treatment of nonmetastatic rectal cancer: a patterns of care study from the national cancer data base. Ann Surg Oncol., 15(11), 3036-47. Epub 2008 Aug 19. Fuller, S.M., et al. (1992). Breast cancer beliefs of women participating in a television-promoted mammography screening project. Public Health Rep., 107(6), 682-90. Goff, B.A. et al. (2007). Predictors of comprehensive surgical treatment in patients with ovarian cancer. Cancer, 109(10), 2031-42. Gordon, L., Scuffham, P., Hayes, S., Newman, B. (2007). Exploring the economic impact of breast cancers during the 18 months following diagnosis. Psychooncology, 16(12), 1130-9. Grann, V. et al. (2006). Regional and racial disparities in breast cancer-specific mortality. Soc Sci Med. , 62(2), 337-47. Epub 2005 Jul 26. Greenberg, D., Earle, C., Fang, C.H., Eldar-Lissai, A., Neumann, P.J. (2010). When is Cancer Care Cost-Effective? A Systematic Overview of Cost-Utility Analyses in Oncology. J Natl Cancer Inst. Jan 7. [Epub ahead of print] Guidry, J.J. et al. (2003). Barriers to breast cancer control for African-American women: the interdependence of culture and psychosocial issues. Cancer, 97(1 Suppl), 318-23. Hackbarth, D.P. et al. (2001). Collaborative research and action to control the geographic placement of outdoor advertising of alcohol and tobacco products in Chicago. Public Health Rep., 116(6), 558–567.

59 | P a g e

62 | P a g e

Hao, Y., Landrine, H., Jemal, A., Ward, K.C., Bayakly, A.R., Young, J.L. Jr, Flanders, W.D., Ward, E.M. (2009). Race, Neighborhood Characteristics, and Disparities in Chemotherapy for Colorectal Cancer. J Epidemiol Community Health, Dec 3. [Epub ahead of print] Harlan, L.C., et al. (2003). Trends in surgery and chemotherapy for women diagnosed with ovarian cancer in the United States. J Clin Oncol., 21(18), 3488-94. Hickey, M. et al. (2009). Breast cancer in young women and its impact on reproductive function. Hum Reprod Update, 15(3), 323-39. Epub 2009 Jan 27. Higgins, R.S.D. et al. (2003). Quality of life concerns in patients with breast cancer: evidence for disparity of outcomes and experiences in pain management and palliative care among African-American women. Cancer, 97(1 Suppl), 311-7. Hofmann, L.J., Lee, S., Waddell, B., Davis, K.G. (2010). Effect of race on colon cancer treatment and outcomes in the Department of Defense healthcare system. Dis Colon Rectum, 53(1), 9-15. Hoffman, R.M. et al. (2001). Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst., 93(5), 388-95. Holmes, L. Jr. et al. (2001). Racial and ethnic differences in advanced-stage prostate cancer: the Prostate Cancer Outcomes Study. J Natl Cancer Inst., 93(5), 388-95. Horner, M.J., Ries, L.A.G., Krapcho, M. et al. (2009). SEER Cancer Statistics Review, 1975–2006, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov.ezproxy.welch.jhmi.edu/csr/1975_2006/ Howard, D.H., Molinari, N..A, Thorpe, K.E. (2004). National estimates of medical costs incurred by nonelderly cancer patients. Cancer, 100(5), 883-91. Howerton, M.W. et al. (2009). Impact of androgen deprivation therapy on racial/ethnic disparities in the survival of older men treated for locoregional prostate cancer. Cancer Control, 16(2), 176-85. Hunter, C.P. et al. (2007). Provider roles in the recruitment of underrepresented populations to cancer clinical trials. Cancer, 109(3), 465-76. Jayadevappa, R., Chhatre, S., Weiner, M., Bloom, B.S., Malkowicz, S.B. Medical care cost of patients with prostate cancer. (2005). Urol Oncol., 23(3), 155-62. Jerant, A.F. et al. (2000). Epidemiology, stage at diagnosis, and tumor biology of breast carcinoma in multiracial and multiethnic populations. Cancer, 88(5 Suppl), 1193-202. Jerant, A.F., Franks, P., Jackson, J.E., Doescher, M.P. (2004). Age-related disparities in cancer screening: analysis of 2001 Behavioral Risk Factor Surveillance System data. Ann Fam Med., 2(5), 481-7.

60 | P a g e

Kattlove, H., Liberati, A., Keeler, E., Brook, R.H. (1995). Benefits and costs of screening and treatment for early breast cancer. Development of a basic benefit package. JAMA, 273(2), 142-8. Keating, N.L., Herrinton, L.J., Zaslavsky, A.M., Liu, L., Ayanian, J.Z. (2006). Variations in hospice use among cancer patients. J Natl Cancer Inst., 98(15), 1053-9. Khankari, K. et al. (2008). Determinants of racial/ethnic colorectal cancer screening disparities. Arch Intern Med., 168(12), 1317-24. Kim, P. (2007). Cost of cancer care: the patient perspective. J Clin Oncol., 25(2), 228-32. Klassen, A.C. et al. (2007). Improving colorectal cancer screening among the medically underserved: a pilot study within a federally qualified health center. J Gen Intern Med., 22(10), 1410-4. Epub 2007 Jul 26. Kutikova, L., Bowman, L., Chang, S., Long, S.R., Obasaju, C., Crown, W.H. (2005). The economic burden of lung cancer and the associated costs of treatment failure in the United States. Lung Cancer, 50(2), 143-54. Epub 2005 Aug 19. Lang, K., Lines, L.M., Lee, D.W., Korn, J.R., Earle, C.C., Menzin, J. (2009). Trends in healthcare utilization among older Americans with colorectal cancer: a retrospective database analysis. BMC Health Serv Res., 9, 227 Laiyemo, A.O., Doubeni, C., Pinsky, P.F., Doria-Rose, V.P., Bresalier, R., Lamerato, L.E., Crawford, E.D., Kvale, P., Fouad, M., Hickey, T., Riley, T., Weissfeld, J., Schoen, R.E., Marcus, P.M., Prorok, P.C., Berg, C.D. (2010). Race and Colorectal Cancer Disparities: Health-Care Utilization vs. Different Cancer Susceptibilities. J Natl Cancer Inst. Apr 6. [Epub ahead of print]) Lauzier, S., Maunsell, E., De Koninck, M., Drolet, M., Hébert-Croteau, N., Robert, J. (2005). Conceptualization and sources of costs from breast cancer: findings from patient and caregiver focus groups. Psychooncology, 14(5):351-60. Li, B.D. et al. (2004). The role of area-level influences on prostate cancer grade and stage at diagnosis. Prev Med., 39(3), 441-8. Lipscomb, J. (2008). Estimating the cost of cancer care in the United States: a work very much in progress. J Natl Cancer Inst. 100(9), 607-10. Epub 2008 Apr 29. Lythcott, N. et al. (2000). Patient compliance is critical for equivalent clinical outcomes for breast cancer treated by breast-conservation therapy. Ann Surg., 231(6), 883-9. Mandelblatt, J.S., Schechter, C.B., Yabroff, K.R., Lawrence, W., Dignam, J., Muennig, P., Chavez, Y., Cullen, J., Fahs, M. (2004). Benefits and costs of interventions to improve breast cancer outcomes in African American women. J Clin Oncol., 22(13):2554-66. Epub 2004 Jun 1. Mandelblatt, J.S., Schechter, C.B., Yabroff, K.R., Lawrence, W., Dignam, J., Extermann, M., Fox, S., Orosz, G., Silliman, R., Cullen, J., Balducci, L. (2005). Breast Cancer in Older Women

61 | P a g e

64 | P a g e

Research Consortium. Toward optimal screening strategies for older women. Costs, benefits, and harms of breast cancer screening by age, biology, and health status. J Gen Intern Med., 20(6), 487-96. Marlow, N.M., Halpern, M.T., Pavluck, A.L., Ward, E.M., Chen, A.Y. (2010). Disparities associated with advanced prostate cancer stage at diagnosis. J Health Care Poor Underserved, 21(1), 112-31. Marly, R.C. et al. (2003). The perspective of African-American breast cancer survivor-advocates. Cancer, 97(1 Suppl), 324-8. Max, W., Sung, H.Y., Stark, B. (2008). The economic burden of breast cancer in California. Breast Cancer Res Treat., 116(1), 201-7. Epub 2008 Aug 6. Max, W., Rice, D.P., Sung, H.Y., Michel, M., Breuer, W., Zhang, X. (2002). The economic burden of prostate cancer, California, 1998. Cancer, 94(11), 2906-13. McNeill, L.H. et al. (2008). Racial/ethnic differences in breast cancer outcomes among older patients: effects of physician communication and patient empowerment. Health Psychol., 27(6), 728-36. Morris, A.M. et al. (2009). Breast and prostate cancer survival in Michigan: can geographic analyses assist in understanding racial disparities? Cancer, 115(10), 2212-21. Morris, A.M. et al. (2009). Residual treatment disparities after oncology referral for rectal cancer. J Natl Cancer Inst., 100(10), 738-44. Epub 2008 May 13. Mulligan, C.R. et al. (2004). Racial disparities in rectal cancer treatment: a population-based analysis. Arch Surg., 139(2), 151-5. Mullins, C.D., Cooke, J.L. Jr., Wang, J., Shaya, F.T., Hsu, D.V., Brooks, S. (2004). Disparities in prevalence rates for lung, colorectal, breast, and prostate cancers in Medicaid. J Natl Med Assoc., 96(6), 809-16. Mullins, C.D., Snyder, S.E., Wang, J., Cooke, J.L., Baquet, C. (2004). Economic disparities in treatment costs among ambulatory Medicaid cancer patients. J Natl Med Assoc., 96(12), 1565-74. Mullins, C.D. et al. (2006). Unlimited access to care: effect on racial disparity and prognostic factors in lung cancer. Cancer Epidemiol Biomarkers Prev., 15(1), 25-31. Murthy, V.H. et al. (2004). Disparities in prevalence rates for lung, colorectal, breast, and prostate cancers in Medicaid. J Natl Med Assoc., 96(6), 809-16. Nicholson, R.A. et al. (2004). Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA, 291(22), 2720-6.

P a g e | 62

65 | P a g e

Ogedegbe, G. et al. (2008). Unintended effects of emphasizing disparities in cancer communication to African-Americans. Cancer Epidemiol Biomarkers Prev., 17(11), 2946-53. O'Malley, A.S. et al. (2005). Perceptions of barriers and facilitators of cancer early detection among low-income minority women in community health centers. J Natl Med Assoc., 97(2), 162-70. Onega, T., Duell, E.J., Shi, X., Demidenko, E., Goodman, D.C. (2010). Race versus place of service in mortality among Medicare beneficiaries with cancer. Cancer, Mar 22. [Epub ahead of print] Paterniti, D.A. et al. (2005). Disparities despite coverage: gaps in colorectal cancer screening among Medicare beneficiaries. Arch Intern Med., 165(18), 2129-35. Payne, R. et al. (2005). Asian Americans and cancer clinical trials: a mixed-methods approach to understanding awareness and experience. Cancer, 104(12 Suppl), 3015-24. Penberthy, L., Retchin, S.M., McDonald, M.K., McClish, D.K., Desch, C.E., Riley, G.F., Smith, T.J., Hillner, B.E., Newschaffer, C.J. (1999). Predictors of Medicare costs in elderly beneficiaries with breast, colorectal, lung, or prostate cancer. Health Care Manag Sci., 2(3), 149-60. Ramsey, S.D., Howlader, N., Etzioni, R.D., Donato, B. (2004). Chemotherapy use, outcomes, and costs for older persons with advanced non-small-cell lung cancer: evidence from surveillance, epidemiology and end results-Medicare. J Clin Oncol., 22(24), 4971-8. Ramsey, S.D., Berry, K., Etzioni, R. (2002). Lifetime cancer-attributable cost of care for long term survivors of colorectal cancer. Am J Gastroenterol.,97(2), 440-5. Rosser, C.J. et al. (2003). Understanding Disparities in Outcomes in Cardiovascular Medicine and Thoracic Oncology in African-American Patients. Ann Thorac Surg., 76(4), S1346-7. Schwartz, K., Powell, I.J., Underwood, W. 3rd, George, J., Yee, C., Banerjee, M. (2009). Interplay of race, socioeconomic status, and treatment on survival of patients with prostate cancer. Urology. 74(6), 1296-302. Shavers, V.L. et al. (2004). Racial influence on biochemical disease-free survival in men treated with external-beam radiotherapy for localized prostate cancer. J Natl Med Assoc., 96(7), 939-44. Shavers, V.L. Brown, M., Klabunde, C.N., Potosky, A.L., Davis, W., Moul, J., Fahey, A. (2004). Race/ethnicity and the intensity of medical monitoring under 'watchful waiting' for prostate cancer. Med Care, 42(3), 239-50. Shinagawa, S.M. et al. (2004). Race/ethnicity and the receipt of watchful waiting for the initial management of prostate cancer. J Gen Intern Med., 19(2), 146-55.

63 | P a g e

66 | P a g e

Short, L.J., Fisher, M.D., Wahl, P.M., Kelly, M.B., Lawless, G.D., White, S., Rodriguez, N.A., Willey, V.J., Brawley, O.W. (2009). Disparities in medical care among commercially insured patients with newly diagnosed breast cancer: opportunities for intervention. Cancer, 116(1), 193-202. [Epub ahead of print] Shugarman, L.R., Mack, K., Sorbero, M.E., Tian, H., Jain, A.K., Ashwood, J.S., Asch, S.M. (2009). Race and sex differences in the receipt of timely and appropriate lung cancer treatment. Med Care, 47(7),774-81. Sloane, D. et al. (2000). The excess burden of breast carcinoma in minority and medically underserved communities: application, research, and redressing institutional racism. Cancer, 88(5 Suppl), 1217-23. Smith, A.K., Earle, C.C., McCarthy, E.P. (2009). Racial and ethnic differences in end-of-life care in fee-for-service Medicare beneficiaries with advanced cancer. J Am Geriatr Soc., 57(1), 153-8. Epub 2008 Nov 21. Swanson, M. et al. (2009). Cancer epidemiology in the United States: racial, social, and economic factors. Methods Mol Biol., 471, 65-83. Szucs, T.D., Dedes, K.J.. Is cancer prevention ever going to be profitable? Recent Results Cancer Res., 181, 41-7. Taylor, V. et al. (2003). Breast cancer among young African-American women: a summary of data and literature and of issues discussed during the Summit Meeting on Breast Cancer Among African American Women, Washington, DC, September 8-10, 2000. Cancer, 97(1 Suppl), 273-9. Thompson, I. et al. (2003). Colorectal cancer screening among African Americans: the importance of physician recommendation. J Natl Med Assoc., 95(9), 806-12. Warren, J.L., Yabroff, K.R., Meekins, A., Topor, M., Lamont, E.B., Brown, M.L. (2008). Evaluation of trends in the cost of initial cancer treatment. J Natl Cancer Inst., 100(12), 888-97. Epub 2008 Jun 10. Warrick, C. et al. (2001). Association of African-American ethnic background with survival in men with metastatic prostate cancer. J Natl Cancer Inst., 93(3), 219-25. White, A. et al. (2002). Prostate cancer education in the Washington, DC, area. J Natl Med Assoc., 94(11), 963-70. Wolsko, P.M. et al. (2008). Racial disparities and treatment trends in a large cohort of elderly African Americans and Caucasians with colorectal cancer, 1991 to 2002. Cancer, 113(12), 3400-9. Wong, M.D. et al. (2002). Insurance coverage, medical conditions, and visits to alternative medicine providers: results of a national survey. Arch Intern Med., 162(3), 281-7.

P a g e | 64

67 | P a g e

Wright, G.E., Barlow, W.E., Green, P., Baldwin, L.M., Taplin, S.H. (2007). Differences among the elderly in the treatment costs of colorectal cancer: how important is race? Med Care, 45(5), 420-30. Xiao, H. et al. (2009). The contribution of cancer incidence, stage at diagnosis and survival to racial differences in years of life expectancy. J Gen Intern Med., 24(4), 475-81. Epub 2009 Feb 3. Xiao, H. et al. (2007). Analysis of prostate cancer incidence using geographic information system and multilevel modeling. J Natl Med Assoc., 99(3), 218-25. Yabroff, K.R., Warren, J.L., Brown, M.L. (2007). Costs of cancer care in the USA: a descriptive review. Nat Clin Pract Oncol., 4(11), 643-56. Yabroff, K.R., Davis ,W.W., Lamont, E.B., Fahey, A., Topor, M., Brown, M.L., Warren, J.L. (2007). Patient time costs associated with cancer care. J Natl Cancer Inst, 99(1), 14-23. Yabroff, K.R., Warren, J.L., Schrag, D., Mariotto, A., Meekins, A., Topor, M., Brown, M.L. (2009). Comparison of approaches for estimating incidence costs of care for colorectal cancer patients. Med Care, 47(7 Suppl 1), S56-63. Yabroff, K.R., Warren, J.L., Knopf, K., Davis, W.W., Brown, M.L. (2005). Estimating patient time costs associated with colorectal cancer care. Med Care, 43(7), 640-8. Yabroff, K.R., Mariotto, A.B., Feuer, E., Brown, M.L. (2008). Projections of the costs associated with colorectal cancer care in the United States, 2000-2020. Health Econ., 17(8), 947-59. Yabroff, K.R., Bradley, C.J., Mariotto, A.B., Brown, M.L., Feuer, E.J. (2008). Estimates and projections of value of life lost from cancer deaths in the United States. J Natl Cancer Inst., 100(24), 1755-62. Epub 2008 Dec 9. Yabroff, K.R., Bradley, C.J., Mariotto, A.B., Brown, M.L., Feuer, E.J. (2008). Estimates and projections of value of life lost from cancer deaths in the United States. J Natl Cancer Inst., 100(24), 1755-62. Epub 2008 Dec 9. Yabroff, K.R., Kim, Y. (2009). Time costs associated with informal caregiving for cancer survivors. Cancer, 115(18 Suppl), 4362-73.

the societal and economic impact of cancer health disparities · an g eir ge lopm te cancer a...

Documents