|
 |
|
|||||||
|
|||||||||
© 2002 American Public Health Association
COMMENTARY |
David B. Thomas is with the Program in Epidemiology, Fred Hutchinson Cancer Research Center, Seattle, Wash.
Correspondence: Requests for reprints should be sent to David B. Thomas, MD, DrPH, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave North, MP-474, Seattle, WA 98109-1024 (e-mail: dbthomas{at}fhcrc.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
To support cancer control efforts, the National Program of Cancer Registries was established to enhance or develop cancer registries in every state. But there is an alternative.
State cancer control programs could be adequately planned and evaluated without a cancer registry, and federal support of cancer registration could be selectively provided to registries that provide data needed to monitor cancer incidence and survival for the nation or that serve as a resource for population-based etiologic and cancer control research. The funds saved could be redirected to support the continued collection of high-quality cancer incidence and survival data for the nation as such efforts become more costly and complex in the future, and to expand support of population-based cancer research efforts. (Am J Public Health. 2002;92:1064–1066)
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
IN A COMMENTARY PUBLISHED in the Journal in 2000, Izquierdo and Schoenbach1 argued that "a national system of population-based cancer registries is essential to monitor state- and local-based cancer patterns and trends and to orient cancer prevention and control activities," contending that "the main challenge is to continue nurturing the ongoing process: further expanding [cancer registration efforts] to cover all states. . . ." They predicted that "[w]ithin the next decade, the state cancer registry network and SEER [Surveillance, Epidemiology, and End Results] together will collect incidence data on approximately 97% of the US population, laying the foundation for a national system of cancer surveillance."
Although there is little doubt that some benefits would result from such a system, its development and maintenance would be very costly, and it therefore seems wise to consider alternatives. This article (1) questions the premise that cancer registries are essential for quality cancer control activities, (2) indicates the circumstances under which population-based cancer registries should be established and strongly supported, (3) argues against the establishment of a national system of universal cancer registration, and (4) urges that priorities be shifted from support of a nationwide system of cancer registries to more intensive support of registries meeting specific criteria on the basis of their usefulness for public health and research purposes.
|
ALTERNATIVES TO STATE CANCER REGISTRIES |
|---|
|
TOPABSTRACTINTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
The rationale for a network of cancer registries that covers representative populations of all major racial, ethnic, and socioeconomic groups in the United States is presented in the next section. With such a system in place, plus the availability of statewide mortality statistics and census data, states without cancer registries would not need them to plan or evaluate their cancer control activities. Consider the following activities for which a state cancer registry has been advocated and the alternatives.
1. Identification of high-risk groups.
Incidence rates for specific socioeconomic, racial, and ethnic groups in other areas could be applied to the comparable segments of the population in a state without a cancer registry to estimate numbers of incident cases in groups of interest with sufficient accuracy for planning purposes, and state mortality statistics would be available to enumerate cancer deaths. This information could then be used to establish priorities for primary or secondary cancer prevention efforts.
2. Evaluation of primary prevention efforts.
Primary prevention of cancer is affected by reducing or eliminating exposure to known carcinogenic substances or enhancing exposure to substances that prevent cancer. It is not the aim of state programs to determine whether a change in exposure to a substance will alter cancer incidence; this is a "given."
Furthermore, a decline in incidence would usually not be evident for many years after the reduction in exposure (notable exceptions are a decline in Kaposi's sarcoma following highly active antiretroviral therapy for HIV infection and a decline in endometrial cancer following reduction in use of high-dose estrogen replacement therapy). In addition, changes in incidence are subject to multiple interpretations; declines in exposure to one carcinogen could be offset by increases in another, obscuring the intervention's effect on incidence. The appropriate measure of the program's success would be the desired alteration in level of exposure in the target population. For example, the level of smoking among teenagers can be assessed by periodic surveys, and overall cigarette consumption can be estimated from data on sales and tax revenue from such sales.
3. Evaluation of secondary prevention efforts.
Secondary prevention is reduction of deaths by early detection followed by definitive treatment. Again, it is not the function of a state department of health to determine whether a particular screening modality for early detection is efficacious; this is assumed in planning a screening program. Cancer registries have been useful in showing whether there has been an increase in survival and a shift in tumor size and stage at diagnosis in response to a screening program, as well as in identifying cases that are missed or diagnosed between screenings (interval cancers). There are, however, alternative ways to assess success that cost less than routinely collecting data on all cancer cases in the population. Trends in mortality rates for the cancer of interest can be measured from death certificates and census data, cases not detected by screening can be identified by active follow-up of screened individuals, and information on tumor size and stage can be obtained through ad hoc studies of screen-detected, missed, and interval cancers. Most importantly, coverage of the target population can be estimated from the screening program data and estimates of the size of the target population from the census.
4. Assessment of outcome of treatment.
Assessing the efficiency of new treatment modalities is best accomplished by randomized clinical trails and is generally not a function of a population-based cancer registry. Observational studies of differences in survival are complex and the results subject to multiple interpretations. They are best conducted in a research setting, not as a routine activity of a state health department. The American College of Surgeons, with the support of the American Cancer Society, has long advocated registration of all cases by hospital tumor registrars, and annual follow-up of all cases, as part of its accreditation procedure. It is not the purpose of this article to judge the usefulness of this activity. It is relevant, however, to point out that this system preceded the development of population-based cancer registries and could continue in the absence of such registries with no cost to state health departments.
5. Search for unusual occurrences of cancer.
True local clusters of cancers are extremely rare, and since monitoring populations for clusters rarely leads to public health action, this is not a strong justification for a state cancer registry. Reports of possible clusters to which health departments must respond can be investigated on an ad hoc basis. Although the existence of a cancer registry is useful in such situations, it is not essential, and the need to respond to such inquiries is not a strong argument for support of a registry.
It has been argued that with nationwide cancer registration, patterns of cancer occurrence could be assessed and unusual occurrences identified. This has already been done with mortality statistics.2 Although some additional national cancer patterns of interest could well be identified from incidence data, mortality and incidence rates are highly correlated for most cancers (Hodgkin's disease, testicular cancer, and many childhood cancers being exceptions), and the gain in information would quite likely not be sufficient to justify the expense of registering all cases in the country.
In summary, some use can certainly be made of data from a state cancer registry to support cancer control activities. However, there are enough good alternatives to registry data that funding for a statewide registry is difficult to justify based only on its value for cancer control purposes. In the state of Washington in 1995, a plan for primary or secondary prevention of 5 major cancers was developed through the use of state mortality files, risk factor information from the state Behavioral Risk Factor Surveillance System, and cancer incidence from a local SEER registry that covers only a portion of the state.3
|
NATIONAL SUPPORT OF SELECTED REGISTRIES |
|---|
|
TOPABSTRACTINTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
Although national support of all state cancer registries is not justifiable solely to support local cancer control activities, 2 valuable functions argue for national support of selected population-based cancer registries.
1. Monitor cancer incidence and survival for the nation.
This function has been reasonably well served for over 25 years by the SEER program, which, until its recent expansion, collected high-quality information on cancer incidence and survival from almost 14% of the population.4 The major weakness of the program has been underrepresentation of certain populations of interest. This deficiency has been recognized, and attempts have been made to include additional populations. The registry in Atlanta was expanded in 1978 to cover 10 counties in Georgia with large rural Black populations. In 1992, the Cancer Surveillance Program in Los Angeles was brought into the SEER program to expand coverage of Asian and Hispanic populations. More recently, funds have been made available to provisionally include the states of Kentucky (for poor Whites of Appalachia), Louisiana (rural Blacks), New Jersey (Puerto Ricans), and California (multiple ethnic groups).5 If this process were to continue, either by expansion of the SEER program or through the National Program of Cancer Registries (NPCR) of the Centers for Disease Control and Prevention (CDC), a network of registries would exist that could provide cancer incidence and survival rates for all major racial, ethnic, and socioeconomic groups in the country.
This network would be sufficient to monitor temporal trends in incidence and survival across all major segments of the US population and to identify changes and disparities among groups that should be investigated and remedied. The numbers of rare cancers—and of all major histological types of the more common tumors—registered in this network should be large enough to provide stable incidence rates and adequate statistical power for research studies. This network should receive adequate, long-term stable funding. It would be in keeping with the 1999 recommendations of an ad hoc surveillance implementation group that an expanded surveillance system be developed from the SEER program.6 In addition to expanded coverage, this would include giving the populations covered by the registries high priority for inclusion in future risk factor and medical care surveys to enhance their usefulness in interpreting observed patterns of cancer incidence and survival.6
2. Serve as a resource for cancer research.
Population-based cancer registries have served as a resource for many studies of cancer etiology and cancer care. Some of these studies involve record linkage, collection of pathology specimens, or patient contact to obtain interviews and biologic specimens. Their value for such purposes is indisputable. In many instances, the research is initiated and conducted by investigators associated with, or having access to, a single local cancer registry. In other instances, collaborative studies involving multiple investigators associated with different cancer registries or governmental agencies have been conducted to enhance sample size, include multiple racial and ethnic groups, reduce the time for case accrual, and enhance statistical power.
If the 2 major functions of federally funded, population-based cancer registries are to monitor cancer occurrence and survival for the nation and to serve as a resource for cancer research, then it logically follows that priority should be given to national support of registries that meet one or both of the following criteria: (a) cover one or more of the major racial, ethnic, or socioeconomic groups of interest; (b) have an association with a strong epidemiological research program, staffed with investigators willing to participate in collaborative efforts.
|
ARGUMENTS FOR SELECTED REGISTRIES ONLY |
|---|
|
TOPABSTRACTINTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
There are several reasons not to establish a cancer registry in every state but instead to support carefully selected registries.
1. A state registry is not needed for cancer control efforts.
As indicated above, cancer control activities at the state and local level can be planned and evaluated in the absence of a state cancer registry. Therefore, national efforts should be in support of selected registries with a broader function.
2. It is unlikely that state or local governments will provide long-term support of cancer registries.
State and local health agencies have responsibility for a wide range of services, of which cancer control is a relatively small part. Given this, plus the limited role cancer registries need to play in cancer control efforts, it is reasonable for persons responsible for allocating state tax dollars to give low priority to a statewide population-based cancer registry. Although this argument against generous local funding may not be fully appreciated or accepted, the fact of underfunding at the state level is not in dispute. The NPCR was specifically established to provide support for state cancer registries because local funds were generally not available or were insufficient. In order to sustain many state cancer registries, either the CDC will have to provide support through the NPCR indefinitely or state legislatures will have to be convinced that the large sums of money needed are warranted. If neither of these occurs, registries currently being established or upgraded through the NPCR will be disbanded and large sums of money will have been wasted.
3. Cancer registration is likely to become more complex in the future.
The computerization of pathology, hospital, and other medical records poses new challenges to cancer registries. Although these changes provide opportunities for enhancing the efficiency—and possibly reducing the costs—of cancer registration, the development of software and acquisition of hardware needed for automating some procedures is expensive, as is the maintenance of computerized systems. In addition, increased concerns over confidentiality may require more complex and costly procedures to protect privacy. Furthermore, the current revolution in molecular biology will undoubtedly lead to new ways of classifying tumors, based not on their site of origin and morphology but on their underlying genetic profiles or etiology, and cancer registries will have to change to accommodate these new classification procedures. Cancer registrars will have to be trained in new classification schemes, and software that supports registry operations will have to be modified. These changes will require a national effort to standardize the new procedures, train local registry personnel, and defray the costs. Focusing such efforts on selected registries meeting the above-mentioned criteria would be much more efficient, and less costly, than a more diffuse effort in all 50 states.
|
ROLE OF NATIONAL ORGANIZATIONS |
|---|
|
TOPABSTRACTINTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
If enhanced coverage of selected populations, rather than universal coverage of the nation, is accepted as reasonable policy, then the role of the relevant national organizations should change. The NPCR, SEER, and the North American Association of Central Cancer Registries are cooperating to standardize definitions and procedures for cancer registration in the country and to monitor data for quality and completeness. These are important activities that will have to be intensified in the future and should be vigorously supported and given priority over broadening the areas of coverage. The North American Association of Central Cancer Registries,7 through its annual meetings and various working committees, provides an important forum for standardization and exchange of information on registration procedures and uses of cancer registries, and the value of these activities would not be diminished by less than complete coverage of the US population. The 2 funding agencies, SEER and the NPCR, should intensify support of registries meeting the advocated criteria and reevaluate continued support of registries not meeting those criteria. In this way, the finite funds available for cancer registration can be directed away from universal registration and toward support of a more rational program for monitoring cancer in the US population and conducting peer-reviewed, population-based etiologic and cancer control research.
|
Footnotes |
|---|
Peer Reviewed
Accepted for publication October 12, 2001.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONALTERNATIVES TO STATE CANCER...NATIONAL SUPPORT OF SELECTED...ARGUMENTS FOR SELECTED...ROLE OF NATIONAL ORGANIZATIONSReferences |
|---|
1. Izquierdo JN, Schoenbach VJ. The potential and limitations of data from population-based state cancer registries. Am J Public Health.2000;90:695–698.
2. Atlas of Cancer Mortality in the United States: 1950–94. Bethesda, Md: National Cancer Institute; 1999.
3. Opportunities for Prevention: County Profiles for Five Major Cancers. Olympia: Washington State Dept of Health; 1995.
4. Ries LAG, Eisner MP, Kosary CL, et al., eds. SEER Cancer Statistics Review, 1973–1997. Bethesda, Md: National Cancer Institute; 2000.
5.
Pillitterre D. SEER to collect data on underrepresented populations. J Natl Cancer Inst.2000;92:871.
6. Cancer Surveillance Research Implementation Plan. Bethesda, Md: National Cancer Institute; 1999.
7. Chen VW, Wu XC, Andrews PA, eds. Cancer in North America, 1990–1994; vol 1: Incidence. Sacramento, Calif: North American Association of Central Cancer Registries; 1998.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |