Prostate Cancer Conference Report
Session 5: Surveillance and Monitoring

In this session, the participants defined surveillance as the routine and ongoing collection of health information about populations and the timely analysis, interpretation, and communication or dissemination of this information for the purpose of public health action. This action was further defined as being related to the prevention and control of prostate cancer without necessarily referring to governmental public health, but rather to activities at an organized community level to improve the health of a defined population rather than an individual.

Overall, the participants agreed that a major issue in surveillance is confidentiality of the data. They also discussed the need to inform the public about the collection of those data. Once biomarkers and genetic testing are available, confidentiality will become even more important. The availability of PSA testing means that the public needs to be told the extent of testing, as well as why it is done. The public will need to be educated about the need to collect these confidential data and the uses to which they will be put, as well as reassured that the data will remain confidential. The public needs to appreciate that data are collected to increase understanding of variations in the disease and the population, to describe the disease burden, and to obtain guidance on the directions for future research. Once data are collected, the public health community will have information on which to base its efforts for research in specific areas.

Participants noted that drawing public attention to the SEER and NPCR registries would be useful because these data show how cancer surveillance benefits people, which will help to ensure the continuation of the registries.

The more public health agencies can show the use of surveillance systems through different products and studies, the more likely the public will see the benefits of these systems and of continuing to fund these studies. Two types of studies are being conducted: those that are part of ongoing, routine surveillance systems and a special set of studies nested within surveillance systems. Additional surveillance is needed regarding prostate cancer treatment and its complications.

Disease Burden and Risk

A goal of understanding disease burden is to be able to prioritize resources for intervention. To understand disease burden, the population needs to be monitored through either a registry or another type of survey system. The availability of PSA testing causes an increase in disease burden.

The group noted the following key points:

• Surveillance needs to be more specific to ethnic and geographic subpopulations.
• Data on SES and social class would be useful.
• Surveillance of PSA testing is needed to determine what men are being tested and why.
• Improving data quality, completeness, and timeliness should be an ongoing goal.
• Local and public health practitioners should be able to access data more readily.

Some participants suggested that BRFSS might collect additional information that is specific to topics related to prostate cancer. Risk factors other than age, family history, and ethnicity might be identifiable by BRFSS, which surveys 160,000 people each year. For example, more information on dietary practices or physical activity might prove useful in identifying risk factors. One drawback to current BRFSS studies is the low number of Native Americans with telephones and the undersampling of this population. The public health community might consider funding a separate BRFSS-type survey that is focused on prostate cancer, because the current BRFSS survey may not provide adequate data.

One participant suggested that since NCI is studying ethnic differences in incidence, CDC does not need to duplicate this effort. Another participant disagreed, arguing that this is an appropriate issue for the public health community to address. The role of CDC is to facilitate research in specific areas and identify why it wants specific types of data collected. The goal is to determine how best to allocate resources.

Primary Prevention

The group noted additional possible activities related to risk factors. One intention of surveillance is to identify people at high risk. To date, surveillance has shown that race, age, family history, and geographic location appear to be the primary risk factors for prostate cancer. An area in which further surveillance is needed is lifestyle factors. To whatever extent possible, surveillance on key lifestyle factors should be conducted for prostate cancer and for other cancers as well. These factors should include diet, physical activity, and micronutrients.

Fat intake is a possible dietary risk factor, because incidence of prostate cancer changes when racial groups migrate. The genes have not changed, so presumably lifestyle factors affect incidence. For example, prostate cancer rates increase among Asians who have immigrated from China and Japan to Hawaii. This increase may correlate with changes in diet. Additionally, the incidence of prostate cancer among first generation Vietnamese in the United States is low. As part of a study of risk factors, the public health community might conduct surveys to determine whether the incidence rises with succeeding generations.

Participants said that making recommendations in the area of primary prevention is difficult because it is not yet known how to prevent prostate cancer. Although PSA screening is now available, it merely detects prostate cancer without predicting or preventing it. Better biological markers are needed to predict which men are more likely to develop prostate cancer, so that these individuals can be monitored.

The group discussed the difficulty of truly assessing changes in risk of death from prostate cancer. As deaths decrease from other causes, such as cardiovascular disease, deaths from prostate cancer rise because men are living longer. In the same vein, the availability of PSA testing has now given more men a diagnosis of prostate cancer. When they die, they may be more likely to have prostate cancer indicated as the cause of death, even though death may have resulted from another condition. This may make it appear that more men are dying from prostate cancer than has been the case historically.

Secondary Prevention and Treatment

Secondary prevention was defined as early detection. Some participants stated that for early detection of cancer, public health officials need to identify those at higher risk and ensure that these men are screened. SES can be a barrier to screening.

The group discussed several adaptations to surveillance systems to make them more useful for research. Because biomarkers may be the next step in understanding prostate cancer progression, these markers should be included in surveillance systems. Also, surveillance should include a search for geographic variation in incidence and mortality to identify possible risk factors.

Participants noted that linking existing surveillance systems, particularly cancer registries, with other data sources holds great promise because it aids in interpreting outcomes. Special studies on patterns of care, following the SEER model, should be instituted.

The group discussed the possibility that the public health community examine why biopsies are being done and evaluate their positive predictive value. Although this may be more a research than a surveillance question, members of the group agreed that this is an important issue. They noted that surveillance data might also address whether or not a PSA test led to the prostate cancer diagnosis and if so, why the test was performed. With current surveillance of PSA testing, it is not possible to know which cancer patients were detected with PSA screening. Unfortunately, no consensus exists about how to include PSA information in a cancer surveillance system. The group also discussed the usefulness of including information about whether or not a DRE was performed and what the results showed.

The group recommended that CDC and other public health organizations should implement the recommendations from the Institute of Medicine's recently released report, Enhancing Data Systems to Improve the Quality of Cancer Care.

The participants discussed the need to pay more attention to the complications of treatment and recurrence of the disease. At a minimum, the public health community should undertake pilot studies to examine how additional data elements could be added to surveillance in the area of treatment. An improved treatment surveillance system might include information on which segments of the population receive what kinds of treatment. Neither the SEER nor the NPCR registries catalog all the chemotherapeutic modalities used, and even less surveillance is available on hormone and other types of therapies. Watchful waiting and complications of other therapies should be added to surveillance systems. Electronic data sharing between hospitals may facilitate gathering of this information. However, this type of system would miss treatments conducted at physicians' offices unless they were linked with the communication system. A goal of treatment surveillance might be to determine which therapies are most effective at different stages of cancer; both complications and mortality should be examined.

A hindrance to the collection of data on the side effects of therapy might be that patients are reluctant to admit in a telephone survey that they are experiencing some of the common side effects of therapy, such as incontinence and impotence. The public health community might conduct research on which complications are important to patients and determine how they perceive the severity and impact of these side effects.

Survivorship and Quality of Life

Participants noted that because survival and quality of life are separate issues, the public health community should incorporate and improve surveillance of subgroups. Survival should be monitored to determine whether longer survival is due to earlier diagnosis, better treatment, or improved control of comorbid conditions.

Improving existing measures and understanding the patient's perceptions of what is important are paramount in quality of life issues. For example, a treatment side effect or complication that one patient finds intolerable might not bother another patient. It will be important to include a method of measuring patients' perceptions of severity.

Participants suggested that the CDC's NPCR program should use the SEER outcome studies as a baseline and model for conducting special studies to establish quality of life surveillance systems. Existing systems could be used to obtain greater detail on survival and quality of life. The group spent some time discussing the problem of patients moving out of the study area and how this affects data collection.

The participants discussed the possibility of developing a survey to assess quality of life after treatment, given that treatment carries the risk of severe side effects. Telephone surveys should be conducted on a regular basis to track functional disabilities, such as impotence and urinary incontinence, and relate changes to treatments. Functional status may also change over time. Patterns of functional status could be associated with different stages of disease, treatment modalities, and comorbid conditions. These studies could be conducted on a cohort rather than the general population.

The group discussed SES and how this variable might account for some of the variability in cancer incidence in certain racial groups. Studies show that among men with prostate cancer, white men survive longer than black men do, but this finding may be a result of socioeconomic rather than racial characteristics. For example, the wealthier the man, the more likely he is to receive a radical prostatectomy, regardless of age or symptoms. Thus, better information is needed to determine how race and SES affect survival.

The group also suggested that the public health community should consider measuring the recurrence of prostate cancer. One difficulty with this lies in obtaining records from physicians' offices. Public health investigators do not have access to these records and because submission of these records would be voluntary, and collection of such data could be erratic. Some participants thought recurrence data were very important.

Health Communication

Participants suggested that the public health community should be very specific and clear in its rationale for surveillance and should communicate this rationale widely, because these programs depend upon political systems, which in turn depend upon acceptance by the public. They noted that if the value of surveillance systems is not provided to this broader audience, funding for surveillance will not continue.

The group recommended that the needs and benefits of surveillance should be explained to the public to maintain public trust in public health organizations' stewardship of confidential data. Members of the public need to be assured that the collection of personal information is not a threat to them, their security, or their future health. It is essential that the public health organizations cultivate public trust in surveillance and that the public understand the value of collecting, maintaining, and using surveillance information.

Other Comments About Surveillance

Current surveillance activities include the SEER and NPCR registries. A common feature of cancer registries is identification of incident cases and information about demographic characteristics, place of residence, characteristics of the cancer, and first course of treatment. Registries differ in extent and type of follow-up, which can be either active or passive. Participants suggested that an initial survey could be used to define additional needs for data. Surveillance systems could then be modified to collect more focused or in-depth information.

In addition to cancer registries, other relevant public health surveillance systems exist, including NHIS and NHANES.

The group noted that because PSA screening was introduced before its efficacy in reducing mortality being thoroughly established, surveillance data have been used to examine the efficacy of this screening. A similar scenario may occur for other biologic markers. Public health should anticipate future introduction of new biologic markers for cancer risk and prognosis and consider how they might be incorporated into surveillance systems.

Highlighted Suggestions

Disease Burden and Risk

1. Conduct surveillance of incidence among ethnic and geographic subpopulations and search for variations in incidence to identify possible risk factors.



2. Improve quality, completeness, and timeliness of data.

Primary Prevention

1. Conduct surveillance of key potentially prostate cancer-related, lifestyle factors identified as risk factors for other cancers, such as diet and physical activity.



2. Prepare to collect information on biological markers.

Secondary Prevention and Treatment

1. Conduct surveillance on PSA testing, treatment, complications of the disease and treatment, and recurrence. For treatments and comorbidities, consider links to other data sources.



2. Develop mechanisms for studies of patterns of care for prostate cancer.



3. Develop BRFSS and other surveys to collect data on use of the PSA test and on other tests developed in the future.



4. Develop surveillance on biopsies—determine the reasons biopsies are performed and their positive predictive value.

Survivorship and Quality of Life

1. Conduct research to improve existing measures of quality of life by testing reliability and validity and by ensuring that the measures include information about patients' perceptions of what is important to them.



2. Conduct population-based survival studies for subgroups.



3. Use SEER outcome data as a baseline, and conduct studies to establish quality of life surveillance systems.



4. Disentangle reasons differences in survival using existing surveillance systems.

Session Participants