Session: Cancer Prevention and Control in the US: Learning from the Past and Moving into the Future
Session Title: Cancer Prevention and Control in the US: Learning from the Past and Moving into the Future
Time: Tuesday 8:30 a.m.-9:45 a.m.
Room Bayside 105
Session Type: Abstract Driven Session
Chair:Jane Hall (University of Technology-Sydney)
Organizer (1): Nancy Breen (National Cancer Institute - US)
Session Description (393 words)
The hard work of public health officials, physicians, and disease advocacy groups to educate Americans about the importance of cancer screening has resulted in uptake of screening tests at levels equivalent to or higher than in countries with organized cancer screening programs. Studies have shown higher survival rates from cancer in the United States than in other industrialized countries. However, cancer survival, especially in screenable cancers, can be a misleading outcome; per capita costs of health care for Americans in the US are higher than per capita costs in other industrialized countries; and, perhaps most important, many Americans do not have access to the finest cancer care the US can provide. This panel highlights findings from studies that explore cancer care utilization, costs, and access among vulnerable populations. Vulnerable populations are those who are least likely to obtain continuous high-quality cancer care including routine screening and test results, and timely, high quality treatment for cancer. Often they are poor, uninsured, from racial-ethnic minorities, or all three. The US largely relies on market-based for-profit health service delivery for cancer services which does not yield public data sources and data are not widely available. However, cancer can be studied more readily than some other diseases. These researchers have linked various sources of cost, utilization and quality data from large population-based health surveys, cancer registries, hospitals, and low-cost programs. Their results about quality and outcomes in cancer care may be generalized beyond cancer and their methods used to study other conditions. For example, Dr. Barry found that many of the health system variables, including primary care physicians that had been elusive in previous studies became significant when controlling for whether an individual woman had seen a doctor and received a screening recommendation within the past year. Dr. Ponce tested several measures of self-reported discrimination and found more proximal measures were associated with cancer screening. Dr. Bradley found that safety net hospitals play an important role in reducing rates of emergency surgery and recommends continuing to fund them under the Affordable Care Act (Obama care). Dr. Tankga analyzed cost data according to patient risk to obtain valuable estimates on what a proposed program to provide colorectal cancer screening to low-income patients is likely to cost. In doing so, they identify opportunities for organizational change to improve access and quality, and promote cost effective delivery of cancer services.
Theme: Health care reform
Key Terms: primary care, cancer screening, cancer treatment,
Presenter: Janis Barry (Fordham University. Economics)
Abstract (409 words)Most women in the U.S. receive a cancer-screen referral from a primary care physician during a routine, preventive health care visit. This presentation investigates whether the supply of primary care doctors and other health system resources is associated with increased breast or cervical cancer-screening use in the U. S. A distinguishing characteristic of the research presented is the linking of 2005 contextual data on physician characteristics, area poverty levels, population densities, and geographic capacity for screening at the Primary Care Service Area (PCSA), to the nationally representative 2005 National Health Interview Survey (NHIS). All health system resource supply variables were derived from within the 6, 542 bounded medical markets in the U.S. known as PCSAs. Individual responses from the NHIS demographic and health variables (i.e. age, race/ethnicity, health status, insurance coverage, education level) were merged with measures such as the number of OB-GYN’s, international medical graduates in clinical practice, and travel distance to mammography providers. The dependent variable indicates whether a woman received a recent mammogram or Pap test. Each age group of women was examined separately. The empirical findings indicate that proximity to greater numbers of primary care doctors, or full-time equivalent primary care providers encourages more women to receive Pap and mammography screening. Controls for region of the country, residence in urban or rural areas, or in highly populated PCSAs allowed for identification of health care resources that matter for receipt of preventive services. Yet, not all measures of physician supply were equally significant for understanding variations in cancer screening use. The most surprising finding was that characteristics measured at the PCSA-level were significantly modified by the inclusion of whether the respondent “saw Dr. within past year and Dr. recommend test.” This measure of past use of physician services out distanced in significance other key individual or contextual determinants, which affect cancer-screening outcomes. The results reveal that it is the past visit/consultation/recommendation in association with measures of area-level primary care physician supply that best explains the impact of health system resource supply on mammography and Pap screening probabilities. It is only after controlling for whether an individual woman had seen a doctor and received a screening recommendation within the past year, that many of the health system variables became significantly important. Recognition of the need for this intermediate control variable in empirical tests, helps to clarify why in previous studies, the number of physicians in an area has not always been positively associated with increased screening use.
Janis Barry is Associate Professor of economics at Fordham University. A graduate of the New School for Social Research, she specializes in labor economics with research interests in U.S. and international health policy. Her publications focus on gender, race, and class differences in employment, wage rates, hours worked, health-care outcomes, and social capital networks. She is currently investigating the on-going development of two-tier health systems and the increased reliance on market signals to determine whose health is profitable for investment, and whose is not.
Key Terms: physician supply, PCSA, cancer screening
Author (1): Janis Barry (Fordham University. Economics)
Presenter: Ninez Ponce (University of California-Los Angeles. Health Policy and Manag)
Abstract (467 words)Experience of racial/ethnic discrimination in healthcare and other service delivery settings raises equity concerns and imposes differential transaction costs on optimal healthcare. Racial/ethnic discrimination has been implicated as a root cause of health disparities, but fewer studies have documented its negative effect on access to healthcare. While testable, the empirical challenge is the measurement of self-reported discrimination by the person seeking health care. In this study, we empirically test whether self-reported breast and colorectal cancer screening yield different results when using discrimination measures: (1) of lifetime versus recent experiences and (2) one of which asks, in a single question, about racial/ethnic discrimination and the other of which first asks about unfair treatment in general, then ask why/attribution. We pooled experimental data from the California Health Interview Survey (CHIS) Discrimination Module (DM) collected in 2007 and 2009. One experimental arm administered questions that directly attributed race/ethnicity to the experience of discrimination; the other arm first asked about experience of discrimination in general, followed by a question about the reason. Sample sizes for both arms, were 4471 adults, aged 50-75 years old in the colorectal cancer screening analysis and 2927 women aged 40-75 in the breast cancer screening analysis. While the study samples are large, they may not be large enough to detect differences in relatively small population subgroups. We used multivariate logit regression to estimate the predicted margins of screening for each experimental group in models with discrimination, gender and race-ethnicity interactions. Taylor Series approximation was used to compute 95% confidence intervals. Estimates were weighted and included survey year, gender, race/ethnicity, age, education, poverty level, insurance coverage, health status and usual source of care. Overall, self-reports of cancer screening, for both colorectal or breast cancer, exhibited either chiefly null or inconsistent associations – in models that controlled for socioeconomic position, access to health care, and age -- with both lifetime and recent self-reported experiences of racial discrimination. Among the three observed statistically significant (p < 0.05) associations, two indicated less screening among persons who self-reported any versus no racial/ethnic discrimination (for black women: recent discrimination assessed by the 1-stage question, 66.9 vs. 89.0% for colorectal screening; for white men: recent discrimination assessed by the 2-stage question, 37.6% vs. 66.8% for colorectal screening), and one indicated higher breast cancer screening among black women reporting racial/ethnic lifetime discrimination assessed by the 1-stage question (71.1% vs. 58.1%). In this study, the different approaches to measuring lifetime and recent self-reported experiences of racial discrimination did not markedly differ in predicting the likelihood of cancer screening. Given the sparse literature on whether or not discrimination levies transaction costs on use of health services, especially cancer screening, future research is needed to explore whether similar or different findings would result from use of questions that focused specifically on discrimination encountered in health care settings.
Ninez A. Ponce, PhD is an associate professor in the Department of Health Services, UCLA School of Public Health and a senior research scientist at the UCLA Center for Health Policy Research, and a member of the UCLA Jonsson Comprehensive Cancer Center. As Co-Principal Investigator of the California Health Interview Survey (CHIS) 2001 and 2003, she led the efforts in the measurement of race/ethnicity, acculturation, physician-patient communication, and discrimination. In addition, Dr. Ponce conceptualized the rationale and implementation of CHIS 2001's Asian ethnic oversamples and the cultural and linguistic adaptation of the survey.
Key Terms: racial-ethnic discrimination, cancer screening, self-reported discrimination
Authors (4): Nancy Breen (National Cancer Institute - US. Cancer Control and Population Sciences) , Nancy Krieger (Harvard University. Society, Human Development, and Health) , Benmei Liu (National Cancer Institute - US. Cancer Control and Population Sciences) and Melissa Gatchell (University of California. Health Policy and Management)
Presenter: Cathy Bradley (Virginia Commonwealth University. Healthcare Policy and Research)
Abstract (457 words)Safety net hospitals are U.S. institutions that by legal mandate or explicit mission offer access to services regardless of patients’ ability to pay. Consequently, most patients are uninsured, Medicaid insured, or belong to other vulnerable groups. Because these institutions have a mission to serve uninsured and Medicaid insured patients they may provide better access to timely and appropriate care than these patients receive outside the safety net. We examine whether safety net hospitals reduce the likelihood of emergency colorectal cancer (CRC) surgery in uninsured and Medicaid insured patients. If these patients have access to safety net providers, they should be less likely to undergo emergency resection relative to similar patients at non-safety net hospitals. We extracted inpatient admission data from the Virginia Health Information discharge database. Cases were matched to the Virginia Cancer Registry for patients aged 21 to 64 years that underwent a CRC resection between January 1, 1999 and December 31, 2005 (n=7425). We differentiate between medically defined emergencies and those that originate in the emergency department (ED). First, we combined all CRC resection patients admitted through the ED with patients that had an emergency defined as bowel perforation, peritonitis, obstruction or hemorrhage. The second definition restricts the sample to all CRC patients admitted through the ED and the third definition restricts the sample to those with an ED admission, but no ICD-9 emergency code. By considering ED admissions without an emergency diagnosis code, we identify admissions related to access, but without a condition that requires immediate medical attention. These resections reflect an emerging condition that was assessed in the ED. The analysis of these resections provides a cleaner test of the hypothesis that patients may have used the ED as a portal for symptom appraisal, reflecting the fact that they have no other source of health care. The fourth definition restricts the sample to patients with an emergency ICD-9 code. A diagnosis with an ICD-9 emergency code reflects an immediate medical emergency. For each definition of emergency surgery, we estimate linear probability models for ease of interpretation and to avoid challenges associated with the interpretation of interaction terms in nonlinear models. Our results suggest that safety net institutions play a role in reducing rates of emergency surgery, particularly those stemming from ED admissions, among uninsured and Medicaid patients with CRC. The Affordable Care Act (ACA) will cut support for safety net hospitals if fully implemented, including DSH payments, in anticipation of expanded insurance coverage. Yet, a recent study from Massachusetts indicates that safety net hospitals faced increased demand after reform because they continued to be heavily utilized by low-income populations. Combined with evidence that safety net hospitals reduce emergency resections for CRC patients, this suggests that continued support for these institutions may be warranted.
Cathy J. Bradley, Ph.D. chairs the Department of Healthcare Policy and Research at Virginia Commonwealth University Medical Center. Her research focus includes the impact of uninsurance, public insurance (such as
Medicaid) and the safety net on access to care and the health of low-income and underserved populations. She is also a leading expert in labor market outcomes of cancer survivors and, more broadly, the effects of employment-contingent health insurance on employment and health insurance status.
Formerly Professor of Health Economics in the Health Services Research Division in the Department of Medicine, Michigan State University, Dr. Bradley received her Ph.D. in health policy and administration with a minor in economics from the University of North Carolina at Chapel Hill in 1994, a Masters in Public Administration from the University of North Carolina, and a B.A. from Old Dominion University. She was previously Section Head of Worldwide Pharmacoeconomics at Procter & Gamble.
Key Terms: safety net hospitals, vulnerable populations, Medicaid, medically defined emergencies
Presenter: Maggie Cole Beebe (RTI International)
Discussant (1): jane Hall (University of Technology, Sydney. Centre for Health Economics Research and Evaluation (CHERE))
Abstract (484 words)Background: Colorectal cancer (CRC) is the second-leading cause of cancer-related deaths among adults in the United States, and CRC mortality and survival rates vary substantially by race. Although there is strong scientific evidence that regular screening decreases CRC incidence and mortality rates, only about two-thirds of U.S. residents for whom screening is recommended are screened at intervals recommended by U.S. Preventive Services Task Force guidelines. The rate of compliance with these screening guidelines is even lower among uninsured U.S. residents. Screening programs that specifically target the uninsured were proposed to help reduce disparities along racial, ethnic, and socioeconomic lines in CRC screening, incidence, and mortality rates. Objective: In 2005, the Centers for Disease Control and Prevention designed and established a 4-year Colorectal Cancer Screening Demonstration Program (CRCSDP) in five sites in the United States to explore the feasibility of establishing a large-scale CRC screening program for low-income, under or uninsured men and women aged 50-64. In this study, we assess the costs of screening colonoscopy for patients at average risk and those at increased risk for CRC due to family history and identify factors that may affect screening costs. Data Sources: We used cost data extracted by participating sites from provider billing data, as well as clinical data on each client screened through the program for 2006–2009. Study Design: We utilize cost data linked with clinical screening data to assess the difference in cost between average risk and increased risk individuals undergoing screening colonoscopy. We performed multivariate analysis to identify factors affecting clinical costs of colonoscopy for CRC screening, controlling for patient, provider, and grantee characteristics. Principal Findings: The patient’s risk level influenced costs. Increased risk patients were on average more than 6% more expensive than average risk patients. Screening a patient at increased risk of developing CRC was associated with higher screening costs because they were more likely to have a polyp or cancer detected during the screening colonoscopy. Overall total screening costs varied significantly across the five program sites and differed for each risk level by provider type and clinical service site. Screening costs were lower when a colonoscopy was performed by a provider other than a gastroenterologist. In particular, screenings performed by colorectal surgeons, general surgeons, and other providers all cost less than those performed by a gastroenterologist. Colonoscopy screenings in the hospital setting cost more than those performed at other clinical service sites (for example, an ambulatory/endoscopy service center) – this is expected given the higher Medicare reimbursement rates for hospital settings. Conclusion: The findings in this presentation can assist future CRC screening programs understand the resources required in screening individuals at average and increased risk levels for developing CRC and the type of providers and clinical service sites to partner with in an effort to efficiently use CRC screening resources and identify more people with precancerous polyps or early stage cancer, when treatment is possible and cost is low.
Key Terms: colonoscopy, colon cancer screening, demonstration program, screening cost
Authors (5): Florence Tangka (Centers for Disease Control and Prevention. Cancer Prevention and Control) , Sujha Subramanian (RTI International) , Maggie Cole-Beebe (RTI International) , Sonja Hoover (RTI International) and Laura Seeff (Centers for Disease Control and Prevention. Cancer Prevention and Control)