Prostate Cancer Conference Report
Session 6: Public Health Research

In this session, the participants considered the role of the public health community in applied public health research in the areas of disease burden and risk, primary prevention, secondary prevention, and survivorship and quality of life. They summarized the following priorities:

Disease Burden and Risk

• Explore factors responsible for the increased risk and the disparity in disease burden among African American men and other racial and ethnic minority groups.
• Elucidate those risk factors.
• Assess variations in quality of care and access to care, particularly for underserved groups.

Primary Prevention

• Develop large cohorts of African-American men.
• Develop new methodologies to identify risk factors.
• Increase participation of minority and medically underserved men in research studies.

Secondary Prevention

• Develop tools to aid in decision making and conduct research on how these tools affect the decision making process.

Survivorship and Quality of Life

• Consider how the uncertainties of screening and treatment of prostate cancer can be communicated to health care providers and patients.
• Study modifiable risk factors for prostate cancer.

Disease Burden and Risk

Participants agreed that further research on risk factors is critically important. Most risk factors have not been explored prospectively in cohort studies. Trials should be designed with endpoints and inclusion criteria to allow extensive risk factor study. Research using human prostate cancer cells in mice should be conducted to evaluate protective and promotive substances.

Available information about risk factors is still inconclusive, and the public health community should give priority to determining the impact of these factors. Reliable information at the population level about disease trends and occurrence needed for the U.S., but it is currently unavailable outside of the SEER program areas. Known risk factors of race and family history should be examined in relation to disease trends. The CDC should develop a white paper—a detailed literature review—of what is and is not known about risk factors.

The group noted that research should include developing tools to study quality of life among cancer patient survivors and should examine how perceived quality of life changes over time and how it varies among population subgroups.

Participants suggested that the public health community could improve surveillance data quality, particularly with regard to disease stage and tumor grade. Public health should also be responsible for a consistent, high-quality expansion of cancer registries apart from SEER.

The group suggested that NCI and the CDC might cooperate on research, with CDC expanding public health research beyond program or service delivery. Although it might be inappropriate for CDC to concern itself with etiological research or basic science, a collaborative effort between CDC and NCI could take place with CDC translating basic science into community intervention.

Public health research is needed to determine reasons for the greater disease burden and risk among African Americans, whether this disparity is the result of biology, behavior, or a combination of several factors. Details about disease burden among other minority ethnic groups are not well known. Because migration studies provide opportunities to study changes in cancer incidence by generation, the group suggested adding birthplace to cancer registries to allow for such studies.

The participants also discussed the importance of developing more comprehensive assessments of disease burden in the U.S. population. Available data could be used to better quantify complications, quality of life, and the cost and use of resources.

The group discussed the need for greater understanding of the mechanisms associated with survivor disparities, such as access to health care or poverty. Participants noted that differences are clearly not limited to poverty; race does play a role, particularly for African American men.

To further the understanding of the relationship between ethnicity and disease burden, the group suggested that public health researchers conduct additional research on factors that affect men's decisions about screening and develop instruments to measure screening behaviors and factors that influence decisions about screening. These factors could include men's attitudes, access to screening and treatment, knowledge, and other health related behaviors. Currently, no standardized tools exist to determine whether prostate cancer testing was done for screening or for diagnosis. Such tools should be developed and tested for appropriateness in ethnically diverse populations and for literacy levels in specific populations.

Participants noted that screening tests can be evaluated through a case-control design only when use of the test has saturated the population and reached a steady-state level. A test that is relatively new cannot be evaluated because use in the control population will not be properly measured in relation to use in the case population. The rate of PSA testing is increasing so rapidly that more prostate cancer cases are being discovered, but use of the test has not saturated the overall population.

Evaluating whether screening with PSA affects the probability that men die of prostate cancer, is difficult, too, using medical records. Many more urologists use PSA screening to monitor benign prostatic hyperplasia (BPH), because a study showed an association between PSA levels and the progression of BPH.

The participants discussed the need to assess quality of care and whether variations in quality of care contribute to differences in disease burden. The group decided that burden should be examined in relation to financial and geographic access to care and to perceived access to care, including such factors as trust, communication, and language. Furthermore, they suggested that studies examine effects of differences in cancer culture, the extent to which men's knowledge and beliefs influence them to seek professional health advice when symptoms are present.

The group also suggested that because the diagnosis of prostate cancer has changed over time and throughout the world, international comparisons should be updated and revisited. For example, researchers in different countries should use the same specimens and records to compare pathology assessments made in the past with assessments made today. Studies should also compare assessments of the same sets of specimens and records by pathologists from the United States and from other countries.

The group discussed the perception that African men have lower incidence of prostate cancer than do African-American men in the United States. Recent data from Nigeria, South Africa, and Canada indicate that this perception may be inaccurate because earlier studies may not have been well designed. Some participants thought that public health researchers should collect specimens and records from pathology data banks in these countries. These data could include microscopic and premalignant lesions and clinically detected prostate cancer. Others disagreed and stated that such research would not be an appropriate use of time and money for prostate cancer research in the United States.

Participants suggested that tools to study quality of life should be developed to address the question of how perceived quality of life differs over time or among subpopulations. The quality of surveillance data should also be improved and the expansion of cancer registries should be consistent and of high quality.

The group discussed how to effectively change behavior through behavioral modification and by targeting high risk groups.

They suggested that public health researchers develop standardized tools that are easily understood and culturally appropriate for low-literacy populations in order to measure knowledge, attitudes, and behaviors.

Primary Prevention

The group discussed a recommendation that public health investigators identify questions in primary prevention that have not been previously addressed by research. More multiethnic cohorts and more diverse populations are needed in clinical trails such as the Prostate Cancer Prevention Trial. Because results of clinical trials strongly influence policies, investigators need to increase participation among African-American and Hispanic men and men with low SES in clinical studies. The Veterans Administration has begun studies on differences in participation in prostate cancer research between African American men and white men.

Participants discussed the need for clinical trials on the recruitment process to improve methods of recruitment. The majority of the trials focus on the disease rather than the recruitment issues. Millions of dollars have been spent on recruiting racial/ethnic minority groups. The University of Alabama study focused on African Americans in one study, but even with enormous effort, recruitment was very difficult, and the recruitment period was extended. Two related issues are retention and compliance with the study protocol.

Socioeconomic status, rather than skin color, could be the most important issue in study participation. If the studies address issues that are not perceived to be pertinent to the lives of the men, then their behavior will be different from that of men who regard issues as pertinent to them.

The group suggested that new strategies are needed to examine environmental risk factors for prostate cancer. Participants noted that newer, quicker methods are needed for laboratory studies using animals to test the effects of environmental exposures on the prostate. The criteria needed to start a large clinical trial with humans should be better evaluated and more clearly established, particularly criteria involving results from animal research.

Secondary Prevention and Treatment

The effectiveness of secondary prevention in reducing mortality is not clear. Participants noted that the PLCO trials will eventually determine the effectiveness of screening and secondary prevention in reducing mortality. The public health community should survey and monitor current community screening practices. Research on patients' perceptions of their treatment options and how they interpret the information given by health care providers about treatment options is also needed. The role of public health agencies in disseminating correct information to the public needs to be clarified.

Participants suggested that because physicians are unlikely to know how to communicate uncertainty about the effectiveness of prostate cancer screening, ways to communicate key issues of uncertainty should be developed. Tools are needed to help physicians convey this uncertainty to patients.

The group noted that ways of increasing recruitment of medically underserved men to randomized treatment trials are needed. Methods for explaining risk need to be improved by considering literacy, racial, or SES impediments, with a variety of messages tailored to selected subgroups.

Participants recommended creation of a core of credible information on prostate cancer screening from independent clinical sources. Public health agencies have a role in disseminating that core information. However, more research is needed on the most effective way to provide education on the uncertainty of the risks and benefits of screening.

Participants questioned how selected tools influence decision making processes. Clinical materials that help in decision making should be developed to help patients make informed decisions about screening and treatment. Patients do make complex medical decisions, but what prompts the final decision is not known. The decision making tool may not be the only factor in making a particular decision. Other sources of information, such as television or spokespersons and organizations that recommend screening, should be factored in. Research is needed on how specific tools affect the decision making process in the context of these other factors.

The group discussed issues related to survivorship including living with the disease as opposed to survival rate. It was suggested that survivorship means living with the diagnosis. It takes many years to determine the efficacy of prostate cancer treatment, making survivorship an important component. However, survival rates are also important as the classical measure of overall treatment outcome, although they do not help in assessing effectiveness of early detection because of lead-time bias. Mortality may be the only reasonable outcome measure for early detection or screening trials. Treatment research must include research or overall survival. Lead-time bias and length bias affect survival estimates when a disease is as susceptible to early diagnosis as prostate cancer. But while this complicates the interpretation of survival, it does not invalidate it. Lead-time and length bias also affect survival with breast cancer and a number of other malignancies, and survival rates are measures for these diseases.

Participants suggested that surveillance of survival rates should be maintained for research purposes. Overdiagnosis affects survival but it is not the same thing as lead time. The group also had other suggestions. The short- and long-term effects of treatment on prostate cancer patients and their families should be assessed. Biologic behaviors of tumor subgroups by histologic grade should be evaluated and therapies or strategies should be linked to subgroups at risk.

Survivorship and Quality of Life

The participants suggested that quality of life should be measured throughout the disease continuum of screening, diagnosis, treatment, and follow-up, and quality of life should be incorporated into existing surveillance systems, such as BRFSS, NHIS and NHANES. Decisions on what to measure, however, should be made in advance. The existing measures and their applicability, reliability, validity, and responsiveness should be assessed. The time periods that should be used for frequency and assessment also need to be determined.

Participants noted that research is needed to determine which changes in functional status and quality of life result from disease and which from the aging process. Baseline quality of life should be established for cancer patients and a comparable disease-free group. Quality of life should be measured in these two groups over time. Family members' quality of life should also be assessed, as well as patient preferences. Information is needed on what outcomes are important to individual patients. Changes in quality of life should be captured over time.

The group recommended that physicians be helped to discuss screening and treatment options with their patients, perhaps with an information summary by means of interactive technology that guides providers through the discussion. Public health agencies, such as CDC, should coordinate and disseminate the information. Uncertainty needs to be communicated in a useful manner, so patients can make more informed choices. An increase in knowledge can be measured, but the increase needs to be in concordance with the values involved in decision making. Side effects and negative treatment outcomes should be communicated more clearly, and the potential for harm should be reduced. After participating in an informed decision making process, older men with a lower likelihood of benefit, or even net harm, are less likely to choose PSA testing. The content of messages to be communicated to patients must be determined. The risk-benefit analysis related to family history and personal risks should be stated in an understandable manner.

In was suggested that messages need to translate science into common language. Visual analog scales can be used. Ongoing efforts to help men evaluate risk versus benefit should be continued to help men make decisions. The efforts may increase patient comprehension. Attitudes and behavior still need to be addressed.

Participants discussed the need to study lifestyle changes following diagnosis that may be related to quality of life. The impact of a prostate cancer diagnosis on other chronic diseases also needs evaluation. From a research standpoint, it is important not only to study prostate cancer incidence, but to examine factors that may affect recurrence of prostate cancer and survival. Patients should be helped to determine what their preference is so that they can have more autonomy in decision making.

The group listed these additional questions that should be addressed by public health research:

• When should intervention take place and what is its impact on quality of life?
• Is intervention at the first occurrence of elevated PSA levels preferable to later intervention?
• What are the treatment thresholds?
• What additional treatments should be evaluated?
• What information is shared with the patient?

Highlighted Suggestions

Disease Burden and Risk

1. Conduct research to describe disparities in disease risk and disease burden by race and ethnicity, particularly for African Americans. Consider the burden of illness in terms of costs and the use of resources for patients, the workforce, families, and communities.



2. Conduct research to increase knowledge of the factors that are responsible for increased risk and the disparity in burden, such as access to health care and poverty.



3. Develop a white paper—a detailed literature review of what is known about these two issues at this time.

Primary Prevention

1. Conduct research on a cohort that includes African Americans, Hispanics, and other underserved populations.



2. Develop new methods to measure risk factors (environmental exposures) and use new models (e.g., animal models).



3. Increase participation of members of race/ethnic minority groups, the poor, and other underserved men in research studies, including clinical trials (e.g., trials on the preventative effects of vitamin E or selenium).

Secondary Prevention and Treatment

1. Conduct research to develop tools to aid men, their families, and clinicians in decision making.



2. Perform research on how those tools affect the decision making process—how patients perceive their treatment options or interpret information that providers give and how patients understand and appreciate the uncertainties of screening tests.



3. Conduct research to determine how complex medical decisions are made and study how the incontinence, impotence, and treatment experiences play in decisions about prostate cancer screening and treatment.



4. Increase participation in screening trials of men from racial and ethnic minority groups, poor, and medically underserved men in screening trials.

Survivorship and Quality of Life

1. Conduct research on how to communicate uncertainty about screening and treatment to health care providers and patients.



2. Perform research on factors that may modify recurrence and survival.



3. Conduct research on thresholds for different treatments and on optimal treatments.

Session Participants

Robert Boer, Erasmus University
Rosalind Breslow, Centers for Disease Control and Prevention
Linda Burhansstipanov, Native American Cancer Corporation
John D. Carpten, National Human Genome Research Institute
June M. Chan, University of California, San Francisco, Medical School
Michel P. Coleman, London School of Hygiene and Tropical Medicine
Christopher M. Coley, Harvard University Health Services
Marvella E. Ford, Henry Ford Health System
John K. Gohagan, National Cancer Institute
Reginald C. Ho, University of Hawaii School of Medicine
Alan R. Kristal, Fred Hutchinson Cancer Research Center
Mark S. Litwin, University of California, Los Angeles
Matthew T. McKenna, Centers for Disease Control and Prevention
Judd Moul, Uniformed Services University
Walter Rayford, Louisiana State University Health Sciences Center
Wael Sakr, Wayne State University School of Medicine
Fritz H. Schröder, Erasmus Medical Center
Robert A. Smith, American Cancer Society
Fred L. Stallings, Centers for Disease Control and Prevention
Sally W. Vernon University of Texas, Houston
Richard Williams, University of Iowa
Facilitator—Amy Harris, Centers for Disease Control and Prevention
Writer and Editor—Brenda Kiser
Writer and Editor—Linda Burhansstipanov