ISSUE REPORT IMPROVING CANCER TRACKING TODAY SAVES LIVES TOMORROW: Do States Make The Grade? SEPTEMBER 2003 TRUST FOR AMERICA’S HEALTH IS A NON-PROFIT, NON-PARTISAN ORGANIZATION DEDICATED TO SAVING LIVES BY PROTECTING THE HEALTH OF EVERY COMMUNITY AND WORKING TO MAKE DISEASE PREVENTION A NATIONAL PRIORITY. ACKNOWLEDGEMENTS: This report is supported by grants from The Robert Wood Johnson Foundation and The Pew Charitable Trusts and from individual champions for improved public health, including Joy A. Horwitz. The opinions expressed in this report are those of Trust for America’s Health and do not necessarily reflect the views of the funders. SEPTEMBER 2003 Introduction M ore than 30 years after the launch of the national War on Cancer, the disease remains the top health concern facing Americans today.1 The American Cancer Society (ACS) estimates that more than half a million people will die of cancer in the U.S. this year.2 Cancer costs our nation more than $180 billion in health care spending and lost productivity from illness and death.3 Millions of Americans are living with the disease every day. The statistics are alarming: I In 2003, about 1.3 million new cancer generation of cancer tracking, control and cases are expected to be diagnosed.4 prevention programs. The report specifi- cally examines the efforts of health agencies I Approximately 9,000 new cases of child- in the states to track information about hood cancer are expected among chil- cancer and their use of this information to dren from birth to age 14 in 2003. While prevent and control the disease. First, it childhood cancer is rare, it is the chief describes the importance of cancer tracking cause of death from disease in children and some of the successes in the fight between ages 1 and 14.5 against cancer that tracking has helped to I An estimated 211,300 new cases of inva- achieve. Second, it evaluates how well states sive breast cancer will occur among are doing in their efforts to track, control women in 2003. Breast cancer is the and prevent cancer, and awards grades on a most frequently diagnosed non-skin can- state-by-state basis. The report concludes cer in women. with recommendations for how cancer tracking programs should be modernized. I Approximately 220,900 new prostate can- Increased efforts made to improve and cer cases are expected during 2003.6 expand prevention programs could help sig- Prostate cancer incidence rates for nificantly reduce the burden of cancer on African-American men are 66 percent families, communities and the nation. higher than for white men.7 I Other cancer disparities exist: The aver- Tracking: A Key To Winning The age age-adjusted cervical cancer mortality War On Cancer rate for white women from 1996-2000 was State health agencies have all developed 2.7 per 100,000. For African-American programs to track, control and prevent can- women, the rate was more than double at cer. Cancer registries are the cornerstone of 5.9 per 100,000. 8 efforts to track the disease. They collect Despite advances in diagnosis and treat- valuable information about cancer trends ment, cancer is still responsible for one of that is crucial to informing treatment and every four deaths in the U.S. ACS estimates, control efforts. Cancer registries compile however, that one-third of cancer deaths cancer data for several purposes: to advance could be prevented.9 cancer research, to develop public health programs to reduce cancer rates and to Trust for America’s Health undertook this improve treatment for cancer patients. report to help create a vision for the next 1 The battle against cancer involves many research community has been a national people: health care providers, physicians, pioneer in recognizing the importance of researchers, epidemiologists, public health disease tracking. Today, all 50 states and the officials, legislators and communities. All of District of Columbia have cancer registry these individuals rely on registry data to programs. develop and evaluate prevention and treat- Many of America’s cancer registries trace ment strategies.10 their origins to the Connecticut Tumor Cancer registries in 40 states and the Registry, which was established in 1935. District of Columbia are housed in public Other states, as well as some cities and health departments. In the other 10 states, counties, gradually followed course, with cancer registries are housed in academic many funding their registries through state institutions. Both the federal and most or local public health departments. Rather state governments provide funding for can- than functioning in concert with public cer registries. health programs, cancer registries histori- cally collaborated more closely with sur- “Health tracking” is the monitoring of dis- geons and other physicians who treat can- ease rates. The information compiled is cer patients. As a result, cancer registries essential for understanding and identifying traditionally have focused on a clinical potentially preventable causes for diseases, approach to battling cancer, which is based learning who is at risk and developing pre- on treatment of patients once they have vention strategies. Currently, there is no been diagnosed. nationwide system in place to track cases of most chronic diseases, like diabetes, asthma In 1972, the National Cancer Institute or Parkinson’s. However, the cancer (NCI) launched the Sur veillance, HEALTH TRACKING: U.S. NEEDS A NATIONWIDE SYSTEM NOW Through “health tracking,” cancer control experts provide. This is crippling our ability to reduce collect and analyze information about cancer and prevent chronic disease and help Americans cases, including whom the disease strikes, the type live longer, healthier lives. of cancer and severity at the time of diagnosis. The U.S. needs a Nationwide Health This information is vital to fighting the disease with Tracking Network to improve our ability to: prevention programs, early detection such as screenings and treatment. While cancer tracking • Guide intervention and prevention strategies, in this country can be improved, it has helped save including lifestyle improvements, lives and reduce health care treatment costs. • Identify, reduce and prevent harmful risks, Unfortunately there is no nationwide health • Improve public health policymaking, and tracking network in place for other chronic • Track progress toward achieving a diseases, such as diabetes and heart disease. healthier nation. Yet the national cost of chronic disease is staggering: four out of five deaths each year The Nationwide Health Tracking Network would and cost $750 billion in annual health care be administered by Centers for Disease Control spending and lost productivity. While the and Prevention (CDC). Congress appropriated U.S. health care system excels at treating $17.5 million in FY 2002 and $28 million in FY disease, we lack an understanding of chronic 2003 to begin such a network. diseases that a national tracking system could 2 Epidemiology and End Results (SEER) grown from $16.8 million in 1994 to $45.6 program under the National Cancer Act. million in 2003. This law mandated the collection, analysis According to CDC, “cancer registries col- and dissemination of data useful for the lect information about the occurrence prevention, diagnosis and treatment of (incidence) of cancer, the types of cancers cancer. The SEER program currently that occur and their locations within the comprises nine statewide and six regional body, the extent of cancer at the time of cancer registries. For three decades, diagnosis (disease stage), and the kinds of SEER has served as a valuable source for treatment that patients receive.”12 CDC also national cancer incidence and survival states that the registries they fund should be trends and an important resource for designed to: researching the causes and most effective treatments for cancer.11 I Monitor cancer trends over time, To further expand the scope of the reg- I Determine cancer patterns in various istries, including adding an increased focus populations, on prevention activities, Congress passed I Guide planning and evaluation of cancer the Cancer Registries Amendment Act in control programs (e.g., determine whether 1992. The law was later reauthorized in prevention, screening and treatment efforts 1998 and is expected to be reauthorized are making a difference), again in 2003. It provided significant feder- al funds to the CDC’s National Program of I Help set priorities for allocating health Cancer Registries (NPCR) to support exist- resources, ing state cancer registries and to help estab- I Advance clinical, epidemiologic, and lish registries in states without them. NPCR health services research, and also required standardized data elements and reporting guidelines that registries were I Provide information for a national data- required to follow in order to receive feder- base of cancer incidence. al funds. Funding for the program has FISCAL PRESSURE ON STATES: WILL IT COST LIVES? Most states are facing the worst fiscal crises However, it did suffer nearly a 10 percent in decades, with a collective shortfall of up to budget cut for 2004-2005. This cut is further $85 billion. Cancer registries have been impacted straining the state’s ability to track cancer and by these budget woes and often do not have its effort to obtain a nationally certified cancer the resources required to do their jobs in the registry. For instance, Texas has a current back- most effective way possible. When states make log of 25,000 cancer case records that must be spending decisions, the importance of registries processed. The budget cut will make it even for controlling and preventing cancer is often harder to reduce the backlog. In addition, with overlooked. less funding, plans to enhance data collection from outpatient facilities and physician offices In the state of Texas, for example, the cancer will be severely impeded. The Texas registry registry narrowly escaped being “zeroed-out” in needs additional staff and resources to address the state budget for 2004-2005, as the legisla- the backlog and continue its efforts to enhance ture struggled to reduce its $10 billion deficit. data collection. Fortunately, the cancer registry was preserved. 3 In February 2000, CDC and NCI signed a report contains federal government cancer Memorandum of Understanding (MOU) to statistics for more than one million invasive establish formal collaboration between NCI’s cancer cases diagnosed during 1999 among surveillance research programs and CDC’s residents of 37 states, six metropolitan areas NPCR. The MOU seeks to better coordinate and the District of Columbia.13 national cancer surveillance efforts by The information collected by cancer reg- improving the availability of high-quality data, istries can be used by a range of offices and advancing the capacity for surveillance within state health agencies and by other research. For the first time ever in 2002, CDC researchers to develop public health initia- and NCI published a national report, United tives to reduce cancer rates. States Cancer Statistics: 1999 Incidence. The SOME SUCCESSES The cancer trends revealed by tracking the dis- L The Kentucky registry was able to help save lives ease have resulted in numerous advancements and millions of dollars in cancer treatment costs. in treatment and prevention, such as cancer In the early 1990s, 35 percent of women diag- detection and screening programs. Finding the nosed with breast cancer in that state were in disease at the earliest stages provides more the late-stage of the disease, for which the sur- opportunities for effective treatment options vival rate is low. Registry data was used to iden- and increases patients’ chances for survival. tify areas of the state that had high rates of late- Some examples of how registries have made a stage and low rates of early-stage breast cancer. difference include: With additional funding from the CDC, the state expanded mammography outreach activities in L The Connecticut Tumor Registry identified these communities. In 1996, the percentage of an epidemic of lung cancer in women in women in the state diagnosed with late-stage 1977.14 These findings provided important breast cancer had declined to 30 percent. In data about the connection between addition to the potential lives saved, detecting smoking and lung cancer. This resulted in these cancers earlier also saved an estimated increased efforts to raise awareness $4.7 million in treatment expenditures.16 through public health programs about the health impacts of smoking. L Cancer registries provide subjects for research studies that lead to primary L In Canada registry data informed the devel- prevention efforts. State cancer registries in opment of more effective cervical cancer Maine, New Hampshire, Massachusetts and screening guidelines, including improved Wisconsin participated in a study published recommendations about the age and fre- in 1995 that found increased risk of breast quency at which women should be cancer with increasing lifetime consumption screened. These updated screening guide- of alcohol.17 This and similar studies revealed lines help doctors detect cervical cancer that reducing excessive alcohol consumption earlier, when it can be more successfully is an important factor in reducing a woman’s treated. As a result, death rates due to cer- risk of developing breast cancer. vical cancer have decreased steadily.15 4 How Well Are States Tracking Cancer? S tate cancer tracking efforts traditionally have been evaluated by the North American Association of Central Cancer Registries (NAACCR). NAACCR, established in 1987, is a collaborative umbrella organization for can- cer registries, governmental agencies, professional associations and private groups in North America interested in enhancing the quality and use of cancer registry data. The organization annually reviews member registries for their ability to produce complete, accurate and timely data. The registry certification program then recognizes those registries meeting the highest standards of data quality with “gold” or “silver” award certificates for each data year. Good data is critical, but ensuring the data ing data available for research to deter- is well used will save lives. In this report, mine whether specific behavioral and Trust for America’s Health (TFAH) focused lifestyle factors, such as diet and exercise on examining not just states’ performance habits, or environmental factors such as in compiling quality data, but also on how second-hand smoke, certain viruses and well the states use and apply this data to pollutants, affect people’s risk of develop- improve their cancer prevention and con- ing specific cancers. trol programs. This added a new dimension 3. Community-Level Answers– how well the to existing quality standards by analyzing if state provides data at the neighborhood and how the information compiled helps level to answer community questions inform prevention strategies and initiatives. about cancer rates, while protecting the TFAH developed a grading system to deter- confidentiality of cancer patients. mine how well states performed on criteria 4. State Legislation– the state legislative and in four categories. NAACCR’s and NPCR’s regulatory efforts to create and maintain data-quality criteria were used in conjunc- effective cancer registries. tion with additional criteria to evaluate how the data is being used and applied to better This survey of state health agencies was con- understand risk factors and possible causes ducted from November 2002 through of cancer. The categories include: February 2003. The cancer registries were the point of contact for the states’ response. 1. Data Quality– the accuracy, completeness The survey asked states to report their pro- and timeliness of data cancer registries col- ficiency on 36 standards.18 These standards lect on individuals diagnosed with cancer. were grouped into the four categories 2. Data Linkage and Data Availability– described above. Designed to measure state whether states are using the data from performance, the standards drew on three cancer registries to link to other available sources: 1) NAACCR, 2) NPCR and 3) health information sources. The criteria TFAH. (For a description of grading crite- also measure states’ performance in mak- ria, see Appendix A). 5 States were awarded points based on their STATE GRADES: Category Weights performance on the 36 standards. The total State points a state received in each categor y Legislation were weighted, and then overall grades were 10% calculated. Data Quality, including accuracy, completeness, and timeliness comprised Data Linkage and Availability 50% of the total grade. This was weighted Data Quality 20% most heavily because it is critical to ensure 50% that information is accurate before it is dis- Community-Level seminated and used as the basis for public Answers health initiatives. The Data Linkage and 20% Data Availability and Community-Level Answers categories were each weighted as 20% of the grade. State Legislation, was weighted the Percentage scores were translated into letter lowest at 10% since adequate regulations grades as follows: 90-100% = A; 80-89% = B; and laws already exist in most states 70-79% = C; 60-69% = D; and below 60% = F. CASE STUDY NEW JERSEY: Cancer registry data can (ATSDR) to begin an epidemiological help determine whether a community has study in 1997 to examine the potential elevated rates of cancer and suggest the exposures associated with the elevated reasons for this increase. In 1995, the New childhood cancer rates. Jersey Department of Health and Senior The investigation found that girls from Services (NJDHSS) responded to citizens’ birth to age 19 were five times more likely concerns that there was an elevated rate of to have leukemia if they were exposed in childhood cancer in Dover Township, utero to well water which had been con- located in Ocean County, NJ. Using data taminated with the solvent trichloroethyl- from the New Jersey cancer registry, health ene and other chemicals leaking from a officials discovered that the incidence of nearby hazardous waste site. Children childhood cancer in Dover Township was who drank well water from contaminated significantly higher than would normally groundwater areas or who lived near a be expected when compared with the rest waste pipe from Ciba-Geigy Corporation’s of the state. Leukemia and brain and cen- now closed Toms River manufacturing tral nervous system cancer rates were par- plant were far more likely to develop ticularly elevated and the excess seemed to leukemia. This is a strong example of cluster in a particular section of Dover how registry data can be used to address Township called Toms River. This prompt- public concerns about the health of a ed the NJDHSS and the U.S. Agency for community.34 Toxic Substances and Disease Registry 6 State-By-State Grades T hirty-four states and the District of Columbia participated in TFAH’s survey, representing 71 percent of the U.S. population. Sixteen states either declined to participate or provided insufficient information to be eval- uated. Of the participants, 12 received an A, 16 earned a B, and three received a C. Three states earned a D and one state received an F. (See Map, State-by-State Grades). The top five performing states overall would help public health officials identify were California, Colorado, Massachusetts, and develop initiatives focusing on risk Washington and Wyoming. factors and possible causes of cancer. The report found that, while most states are I Second, state performance in providing performing well, particularly in maintaining information about cancer rates at the high-quality cancer data, there is generally neighborhood level is mixed. Providing room for improvement when it comes to this information to communities is impor- using the information to improve preven- tant when it comes to answering the pub- tion efforts in two specific areas: lic’s questions about possible disparities in cancer rates among different localities I First, many states do not conduct ade- and ethnic groups. quate linkages to other sources of avail- able information, such as occupational These two areas will be addressed further in and behavioral studies. This information the next section of the report. CASE STUDY NEW YORK: As part of New York’s throughout the state. The CSII also is plan- Cancer Surveillance Improvement Initiative ning to use maps to display the distribution (CSII), the New York State Cancer Registry of cancer risk factors by zip code through- is improving the way information about can- out the state. These tools allow the public to cer is conveyed and disseminated to the begin to answer questions and concerns public. Using cancer registry data along about cancer in their communities using with state-of-the-art mapping techniques, readily available online information from the New York State Cancer Registry’s the cancer registry.31 Web site maps cancer rates by zip code 7 STATE CANCER TRACKING GRADES A B C D F N/A California Alaska Alabama Maine Mississippi Arizona Colorado Arkansas Montana North Dakota Connecticut Idaho Delaware Virginia Tennessee19 Hawaii Illinois District of Columbia Iowa Maryland Florida Kentucky Massachusetts Georgia Louisiana Michigan Indiana Minnesota Missouri Kansas New Jersey Oregon Nebraska New Mexico Pennsylvania Nevada New York Washington New Hampshire North Carolina Wyoming Oklahoma Ohio Rhode Island South Carolina Texas South Dakota20 West Virginia Utah Wisconsin Vermont Life-Saving Links Not Maximized Of the 34 states and the District of Columbia health-related behavioral/lifestyle, occupa- that responded to TFAH’s survey, 30 have at tional and environmental. some time linked cancer registry data with When cancer registries conduct “data link- other sources of information to better under- ages” – combining cancer data with other stand risk factors associated with cancer rates databases – the combination identifies oppor- within a specific population. However, only tunities for cancer risk reduction. Twenty- eleven states – Arkansas, California, seven cancer registries surveyed report that Colorado, Illinois, Massachusetts, Oklahoma, institutional constraints, including inadequate Texas, Washington, West Virginia, Wisconsin funding and staffing levels, limit their ability to and Wyoming — have ever linked their can- perform data linkages on a routine basis. cer data with all of the following types of data: There are a number of basic data linkage stud- 8 ies that could be conducted more routinely to tors such as certain viral agents, or pollu- help enhance cancer tracking. Improved link- tants, can help better target cancer preven- age efforts could positively impact: tion efforts. For instance, environmental I Enhancing Data Quality and Treatment factors, including second-hand smoke, Options. According to the CDC, linking some pollutants, and viral agents such as registry information to other sources can certain types of human papillomavirus help to enhance the data collected. For (HPV), have been linked to cancer. example, most registries do not routinely I Development of Prevention Initiatives. collect information about other health For example, several states have linked issues a patient might have, such as heart information about low rates of mammog- disease, which can affect treatment raphy screening with later-stage cancer options, or details about a treatment diagnosis rates. Through such a study, course, such as type of chemotherapy. Kentucky was able to target mammogra- Linkages with Medicare and health insur- phy screening outreach efforts and reduce ance claims databases can provide this the rates of late-stage breast cancer.21 important information. I Reducing Health Disparities. Linkages I Identifying Risk Factors. Conducting data allow for the development of strategies to linkages also can help identify opportuni- help target and reduce health disparities, ties for cancer-risk reduction. Linking reg- such as differences in cancer rates within istry data with information on lifestyle communities, socio-economic status, occu- behaviors like diet, or environmental fac- pations, age ranges and ethnic groups. PROTECTING PATIENT CONFIDENTIALITY Medical information is among the most sensitive At the same time, the rule also recognizes the and personal information collected. Privacy is also importance of public health surveillance, includ- central to the doctor-patient relationship. To ing cancer tracking conducted by cancer reg- protect medical privacy, the U.S. Department of istries. Consequently, the Privacy Rule general- Health and Human Services (HHS) issued regu- ly exempts cancer registries from the informed lations known as the “Privacy Rule.” Congress consent requirement. Health organizations gave HHS the authority to issue the rule in the covered under the rule must, however, maintain Health Insurance Portability and Accountability a log of private health information disclosures to Act of 1996 (HIPAA). The Privacy Rule defines cancer registries. This requirement preserves administrative policies and procedures to safe- the patient’s right to review an accounting of guard patients’ personal, private health informa- the disclosures of their private health informa- tion. It went into effect for most health organi- tion over a period of time, generally six years.27 zations in April 2003. All states have enacted laws requiring the confi- The Privacy Rule aims to empower patients by dentiality of cancer registry operations. The laws guaranteeing them access to their medical protect the identity of cancer patients and the records, giving them more control over the use sources reporting to the registries. CDC’s and disclosure of their health information, and National Program of Cancer Registries (NPCR) providing recourse if their medical privacy is also requires the protection of patient confiden- compromised. A central feature of the rule is tiality. Together, state laws, NPCR requirements the informed consent provision, which specifies and the federal Privacy Rule strive to achieve a that a health organization cannot disclose pri- balance between safeguarding patient confiden- vate health information without the written, tiality and enabling cancer tracking to prevent and informed consent of the patient. control disease. 9 I Identifying At-Risk Communities. For I Nevada linked cancer registry data to instance, linking water-quality data to emergency room data to help deter- cancer rates in neighborhoods may mine the cost of cancer treatment and yield insights about the potential for to evaluate the financial burden of can- water contamination to be associated cer in Nevada. with elevated cancer rates and the pos- I Oregon linked cancer registry data to sible value of reducing exposures in Medicaid data to study cancer in the dis- future generations. abled Medicaid population, including I Informing Policy Priorities for Treatment the degree to which inadequate cancer and Control Efforts. Systematically link- control services among people with dis- ing cancer data with Medicare and abilities results in higher risk for pre- Medicaid could provide important infor- ventable cancers and for diagnosis at mation for establishing priorities for can- later stages. cer treatment and control policies and I Colorado linked cancer registry data with resource allocation. Medicare data to evaluate quality of care received by the elderly with breast and/or According to the TFAH survey, a number of colorectal cancer. states have shown how data linkages can be I Seventeen states, including Maryland, used to break new ground to better under- New Hampshire, Pennsylvania and stand the obstacles to cancer reduction and Wyoming, have linked cancer registry treatment: data with the Behavioral Risk Factor I Illinois linked cancer data to hospital dis- Surveillance Survey (BRFSS) to under- charge data to explore whether or not stand issues such as how the use of insurance status is associated with stage of screening services varied by different cancer at time of diagnosis. types of cancer.22 THE KEY TO LOWERING DISPARITIES Some ethnic minority groups and individuals liv- According to the Institute of Medicine, individuals ing in poverty are more likely to become ill and living in poverty often lack health insurance or die from cancer at rates that are higher than access to high-quality cancer care and typically the rest of the population. For example: experience high cancer incidence and mortality L The number of new cases of invasive cancer rates and low rates of survival.29 per year from 1996-2000 was 521.7 for Cancer registry data can help public health offi- African-Americans and 479.8 for whites per cials create targeted programs that seek to 100,000 people. reduce the number of cancer cases and cancer L African-American men experience the high- deaths. For example, analysis of cancer registry est incidence rates of lung and bronchial can- data can increase understanding of the factors cer, 120.4 per year per 100,000 males versus that contribute to breast cancer severity among 79.4 for whites. African-American women at the time of detec- L Asian/Pacific Islanders have a liver cancer tion, which can lead to improved survival rates. rate that is nearly three times as high as Understanding these factors can help public whites and a stomach cancer rate that is health officials design outreach programs for more than double that of whites. African-American women that encourage them to seek mammography screening earlier, when L Hispanic/Latina females experience the highest treatment options may be more effective. incidence rates for cervical cancer, 16.8 per year per 100,000 females compared to 9.2 for whites.28 10 I Twenty-one states, including Idaho, highlight possible exposures to cancer- Massachusetts, Texas, Florida, and causing agents in the work environment. Georgia, have linked registry data with environmental databases to explore asso- Data linkage studies can involve a wide ciations between cancer cases and air pol- range of potential analyses with varying lution, water contamination and potential amounts of personnel and computer time to exposure to hazardous waste sites. conduct them. Currently, NPCR and state I Sixteen cancer registries, including governments do not provide adequate Arkansas, Illinois, Michigan, Oklahoma resources for many state health agencies and West Virginia, have linked cancer and registries to conduct the data linkages data with occupational rosters to help they would like to undertake. Registries use REAL-TIME AND RAPID REPORTING Cancer is a chronic disease that may be initiat- With initial diagnostic information and minimal ed decades before the diagnosis is made, but information, such as age at diagnosis and residential that time lag does not exist in childhood can- address, state cancer prevention staff could cers. For example, prenatal X-ray exposure has “screen” the data periodically to look for emerging been shown to increase the risk of childhood patterns. This would require cooperation from the leukemia within a few years of birth. In some pediatric hospitals where a patient is first diagnosed communities, such as Woburn, MA, and Tom’s and some additional state cancer registry personnel River, NJ, there was an excess of childhood time, depending on the size of the state, but would leukemia in children whose mothers were include less than 10 percent of the statewide can- exposed to contaminated well water while preg- cer cases (the approximate percentage of cancer nant. For these reasons, it is desirable to exam- cases that occur in children). Such a system would ine childhood cancer patterns as rapidly as pos- allow for rapid review of these cases to see if unusu- sible to see if unusual numbers are occurring in al new patterns emerge. Later, the full data on each communities that may warrant further investi- childhood case, including first course of treatment gation and preventive action. and related clinical data could be added to enhance data completeness. Real-time reporting of Real-time reporting of childhood cancer has been childhood cancer would provide the potential proposed by a citizen group in Tom’s River, NJ, for early evaluation and intervention to pre- where a significant excess of childhood leukemia vent exposures and would respond to citizens’ and brain cancer was first noted by the group and special concerns about childhood cancer. confirmed by the state cancer registry. In states with large populations, the hospital where a child In addition, real-time reporting for children could is first diagnosed could transmit information elec- serve as a model for rapid reporting of adult can- tronically, with adequate privacy protection, to cer cases. With the use of new technologies, the state cancer registries. This would mean rapid reporting has become common in other pub- almost “real-time” reporting of the case. The lic health initiatives. The CDC and some local hospital subsequently would report additional communities have demonstrated this type of information to supplement the initial data it sent. reporting in their tracking of West Nile Virus and Real-time reporting requires modernizing to take severe acute respiratory syndrome (SARS). advantage of the latest available technology in Though cancer is not an acute infectious disease, electronic reporting systems. NPCR has called for a modern registry system could and should devel- electronic reporting for several years. NAACCR op additional procedures to compile useful data also has recognized the importance of moving rapidly for immediate reporting once a case of toward electronic, real-time case reporting. 30 cancer is accurately diagnosed. 11 their resources to focus first on collecting Many Community-Level Questions and compiling high-quality data – linkages Go Unanswered are a lower priority. Often there is too little TFAH’s study examined the performance of funding to hire enough staff to create data states in providing community-level data linkages, which can be labor-intensive. that can answer the public’s and Increased funding for data linkages should researchers’ questions about disparities in become a top priority for NPCR and the cancer rates among different neighbor- states, even though state and federal gov- hoods and ethnic groups. Of the 34 states ernments are facing significant deficits. and the District of Columbia that responded Given that the nation spends more than to TFAH’s survey, only 14 reported that they $180 billion on treatment and lost produc- provide data at the neighborhood level. tivity due to cancer, investment in data link- More than half the states participating in ages could help prevent new cases of cancer TFAH’s survey do not provide information in the future, decrease the number of at the community level, typically due to lim- deaths from cancer and potentially save ited funding and resources. health care dollars. Each year, more than a thousand communi- At a meeting hosted by CDC/NPCR, cancer ties or individuals raise concerns to their registry experts recognized the need to health departments about their neighbor- increase the value and utility of registry hoods’ cancer rates.24 Often these commu- databases for cancer prevention and control nities suspect they are experiencing a “can- by improving the linkages to other databas- cer cluster.” A higher-than-expected number es, such as geographic information systems of cancer cases in a certain geographic area (GIS) and census data.23 By incorporating over a defined period of time within a spe- data linkages into routine cancer registry cific group of people is considered to be a operations, public health programs can bet- cluster. Cancer clusters tend to be contro- ter target limited resources to: versial because of the many challenges fac- I Modify policies and programs to improve ing cluster investigations - politics, percep- preventive measures, tions, statistical chance and limitations of sci- ence. Rarely do investigations result in the I Target populations for interventions, identification of true cancer clusters.25 I Provide information about risk factors, and These questions and concerns can only be I Evaluate current activities. properly addressed, and in most cases put CASE STUDY CALIFORNIA: California is one of Farmworkers of America (UFW), a largely America’s leaders in cancer tracking. Due Hispanic farm worker labor union. to the size and diversity of California’s Results showed that the risk of leukemia, population, more is known about the stomach, cervical and uterine cancers was occurrence of cancer among diverse popu- elevated in California farm workers. UFW lations than in most other states. A tradi- members also experienced later stage of tion of complete and accurate data has disease at diagnosis than other California allowed California data to be used in hun- Hispanics for most major cancer sites. dreds of research investigations. For exam- Additional research into the potential ple, from 1988 through 1999, the California causes of this increased risk for certain Cancer Registry studied the incidence of cancers is planned, including a study of cancer among members of the United farm workers’ exposure to pesticides.32 12 to rest, by providing information about can- at the neighborhood level. Accordingly, cer rates at a neighborhood level. The pub- some states have developed policies allowing lic should be able to obtain cancer registry data to be released at any geographic level data in two ways: by reviewing published under certain conditions. For example, sev- annual reports or by asking states for it eral states allow aggregate data to be directly. In reality, neither approach guar- released at a small geographic level, as long antees that the public and researchers as it contains more than five cancer cases. receive information specific enough to be The experience of cancer registries in other able to adequately address their questions countries demonstrates that confidentiality at a neighborhood level. can be protected while tracking cancer with- in specific communities and providing such States’ annual reports provide statewide can- data to citizens at a geographic level small cer rates and sometimes publish rates at small- enough to answer their questions.26 er geographic levels, such as by county, city or town. Typically, however, data specific to TFAH’s survey found that all states but neighborhoods is only available when special Rhode Island had written cancer cluster requests are made by the public. These investigation protocols. These protocols requests are often prompted by concerns that are important to the public because they there may be a potential cancer cluster. The establish formal rules that describe if, TFAH survey found that 21 of the 34 states when and how a state will act to investigate and the District of Columbia surveyed do not a possible cancer cluster. Cancer registries make neighborhood-specific data available to provide state health officials with a meas- citizens requesting it. In addition, of the 14 ure as to whether the observed incidence states that do provide neighborhood-level of cancer is elevated for a specific time data to citizens, 12 states publish breakdowns period. In addition, they also can provide of cancer cases by age, sex and race, but two a statistical estimate as to whether ‘chance’ do not. Many cancer-tracking experts point can be excluded as a competing explana- out that very high-quality control standards tion for the occurrence of the cancer clus- should be met before data is made available ter. If there are indications that there are to the public, to ensure that communities are a statistically significant greater number of receiving and relying on information that is as cases, then the health department may ini- accurate as possible. tiate an investigation, depending on the written protocols. All of the states have patient privacy laws that require them to limit the availability of data CASE STUDY OREGON: Before the 1992 Cancer summary data reports and using the data Registries Amendment Act, Oregon was one both internally and in collaboration with of the many states lacking a statewide cancer external researchers. For example, data is registry. With the help of national funds, being used to document the degree to Oregon has achieved the nation’s highest which inadequate cancer control services standards in data quality and completeness among people with disabilities results in in only a few years of operation. The higher risk for preventable cancers and for registry is now producing comprehensive diagnosis at later stages.33 13 CONCLUSION: A Vision for 21st-Century Cancer Tracking State cancer tracking, prevention and control 3) Implement real-time reporting for child- programs are critical to winning America’s hood cancer and rapid reporting for War on Cancer. They seek to reduce illness adult cancer; and and death due to cancer, and have achieved 4) Continue high standards for data quality. some real success. States have used cancer tracking data to save lives by improving cancer State cancer tracking programs have been screening guidelines and developing public successful at providing information that health programs that have reduced people’s improves “secondary prevention” — cancer risk of developing and dying from cancer. screening and early detection. However, many states are falling short in efforts to use Yet states can do more. Many states could the information they collect to help reduce enhance their data tracking efforts to help cases of cancer that could be prevented alto- reduce cases of cancer that could have been pre- gether. Identifying factors that place indi- vented altogether. Many states need to work viduals or communities at risk for develop- towards expanding their data linkages to other ing cancer and working to reduce those fac- databases. This would accelerate the pace of tors in order to lower that risk is called “pri- valuable discoveries of the factors that contribute mary prevention.” to cancer and identify groups who are at risk today. Most states also need to improve efforts An example of primary prevention that to provide information that helps answer com- states could undertake is to link data on pat- munities’ questions about cancer rates in their terns of environmental behavior, such as neighborhoods. While budget pressures loom, tobacco exposure, diet or exercise, to data cancer tracking and control must be funded on cancer rates. Given current resources, adequately at the state and federal levels. many states do not include such data link- ages for primary prevention among their In order to modernize cancer tracking, pre- ongoing or regular cancer tracking activi- vention and control programs in the U.S., a ties. Yet such linkages have the potential to new, concerted commitment is required to become an even more effective tool in the ensure that states are able to: fight against cancer by expanding their 1) Conduct data linkages to combine reg- focus to include more efforts to stop new, istry data with other sources of informa- preventable cases of cancer from occurring. tion essential to cancer prevention, Reducing preventable cases of cancer and including lifestyle and behavior factors, increasing cancer screening rates could and environmental data, such as certain potentially reduce health care spending over viruses and pollutants; time for health care consumers, including 2) Develop community-friendly practices, the federal and state governments, as well as which help answer the public’s questions employers. More important, advancements about cancer rates in their neighbor- in understanding and improving cancer hoods, while maintaining the confiden- prevention can result in reduced illness and tiality of individuals; countless lives saved. 14 Recommendations For Action The results of TFAH’s survey suggest a series should examine childhood cancer patterns of actions for policymakers to consider with as rapidly as possible to see if unusual num- respect to modernizing America’s cancer bers are occurring in communities that may tracking system to help fight, and ultimately warrant further investigation and preventive win, the War on Cancer. actions. Moreover, real-time reporting could potentially contribute to time-sensitive 1. Expand data linkage activities. Cancer track- cancer treatments and research activities. ing data can be linked with other sources of Real-time reporting for childhood cancer health information to provide important could serve as a model for eventually estab- information and insights. The use and appli- lishing rapid reporting for adult cancers. cation of cancer data is critical to determin- ing cancer excesses in specific populations 5. Strengthen public accountability and and enhancing both primary and secondary enforce performance standards. Standards prevention efforts. As a condition of federal for data quality and timeliness are important support, states should work in conjunction first steps in improving cancer registry data with the CDC and researchers to identify and programs. However, nationally sanc- appropriate linkage studies. tioned standards do not exist to guide states’ policies and practices regarding data link- 2. Provide additional funding for cancer track- ages and response to public concerns. To ing. Although federal and state budgets are establish such standards and methods for under pressure, additional funding from tracking progress on an ongoing basis, both of these levels of government is crucial. NAACCR and NPCR should work directly Funding NPCR at $65 million, an increase of with researchers and community groups. approximately $19 million, would help mod- NPCR has worked hard to assist registries, ernize cancer tracking. Spending money on but it must exercise stronger leadership to tracking can save health care expenditures, motivate steady improvement. Should these and more important, save lives. Cancer organizations fail to prepare adequate stan- tracking should also be integrated into a dards in these areas, Congress should man- comprehensive Nationwide Health Tracking date NPCR set standards on data linkages Network administered by the CDC designed and response to public concerns for all reg- to monitor chronic diseases. istries receiving federal support.. 3. Make data available to the public while pro- 6. Institute of Medicine (IOM) should con- tecting patient confidentiality. States should duct a study on the federal role in cancer actively work to make community-level cancer registries. Congress should direct the IOM data easily accessible and available to the pub- to conduct a review of the existing, bifurcat- lic and researchers. Health agencies in the ed registry system in the U.S. The National states should also adopt policies that allow Cancer Institute and the CDC separately them to provide information at a neighbor- fund cancer registries. Both programs hood level that guard patient confidentiality. serve important purposes, but maintain dis- 4. Improve reporting time. Within five years, tinct goals and missions. Progress has been all registries should set the goal of achieving made toward coordinating these major pro- real-time reporting for childhood cancers. grams, yet obstacles remain. The IOM Cancer is a chronic disease that may be initi- should examine the respective SEER and ated decades before the diagnosis is made, NPCR programs and make recommenda- but that time lag does not exist in childhood tions that will guide development of a seam- cancers. State public health departments less cancer tracking system in the U.S. 15 Appendix A: DESCRIPTION OF THE GRADING CRITERIA Data Quality I Link cancer data with other individual and/or population-level databases to pro- Nineteen criteria were used to assess data qual- vide enhanced health tracking;37 and ity. All but one of the criteria have been used by CDC’s National Program of Cancer Registries I Implement an NPCR standard mandating (NPCR) to evaluate the first five years of their that state cancer registries have procedures program.35 Nine of the 19 criteria are used by allowing public health researchers access to the North American Association of Central confidential case information. Note that Cancer Registries (NAACCR) in their registry some state laws prohibit access to confiden- certification program.36 Where NPCR and tial case information, and NPCR standards NAACCR criteria were the same, TFAH used do not supercede state law. Some states NAACCR’s criteria when assigning points. choose to make their data available in pub- Criteria derived from NAACCR standards were lic use databases, which can meet the needs scored using a 3-point system. A state scored 0 for cancer control planning, but may not points if it performed below NAACCR’s “silver” meet the needs for all types of research. level, 2 points if it performed at NAACCR’s “sil- This is not ideal from the cancer research ver” level and 4 points if it performed at and prevention standpoint. NAACCR’s “gold” level. Criteria based on Performance on each criterion in this catego- NPCR’s standards were scored using a 2-point ry was scored using a 2-point scoring system. system. A state scored 0 points if it did not meet A state received 1 point if it had such proce- the NPCR criteria and 1 point if it did meet the dures in place and 0 points if it did not. Scores criteria. Scores on each criterion were added on each criterion were added together to cre- together to create a category-specific grade. ate a category-specific grade. To get an “A” in The Data Quality scoring method assigned this category, states had to receive the maxi- more weight (more points) to NAACCR crite- mum possible points on every standard. ria to account for the relative value and impor- tance of the NAACCR registry program as an Community-Level Answers indicator of state cancer registries’ perform- TFAH established four criteria to evaluate how ance with regard to data quality, completeness well states provide information about cancer and timeliness. In addition, the scoring system rates at a neighborhood level to the public to translated NPCR’s and NAACCR’s non-numer- help better understand possible disparities. ic evaluation procedures into comparable Performance on each criterion was scored using numeric values. This study used NAACCR’s list a point system. States that provided citizens with of silver- and gold-certified states for 2000 inci- the least informative data scored 1 point. States dence data to validate the Data Quality grading that provided citizens with moderately informa- system and to ensure accurate reporting by the tive data scored 2 points. States that provided cit- states. The Data Quality grades will differ izens with the most informative data scored 4 slightly from NAACCR’s certification results points.38 The criterion regarding written cancer due to the differences in scoring methods and protocols was scored using a 2-point system. the inclusion of additional criteria from NPCR. States with a written cancer cluster protocol in Data Linkage and Availability place scored 2 points, states without one scored 0 points. Scores on each criterion were added TFAH established four criteria to assess together to create a category-specific grade. To whether state cancer registries or other state get an “A” in this category, states had to receive health department programs: the maximum possible points on every standard. 16 State Legislation point for performing the criteria and 0 points it did not. Scores on each criterion TFAH’s State Legislation criteria were based were added together to create a category- on 9 NPCR program evaluation criteria. specific grade. To get an “A” in this catego- Performance on each criterion was scored ry, states had to receive the maximum possi- using a 2-point system. States received 1 ble points on every standard. GRADING SYSTEM GRID Criteria Point System Final Weight 0 1 2 4 Collect and report all NPCR-required data items** Missing 1 or more No missing Use NAACCR standardized record layout** No Yes Employ at least 1 certified tumor registrar** No Yes Written data quality policies & procedures** No Yes Passing EDITS*/** < 97% ± 97% 100% Text information is included in computerized form as well as data codes** No Yes Re-abstracting audits in reporting facilities performed within past year** No Yes Independent audit of the registry performed within past 5 years** No Yes DATA QUALITY Missing age at diagnosis*/** > 3% † 3% † 2% 50% Missing race*/** > 5% † 5% † 3% Missing sex*/** > 3% † 3% † 2% Missing county/state*/** > 3% † 3% † 2% Duplicates*/** > 2/1,000 † 2/1,000 † 1/1,000 Data ≥ 90% complete available within 23 months of the diagnosis year* No Yes Current published annual reports available within 12 months of the diagnosis year (as of December 2002)** > 24 months † 24 months † 12 months Interstate data sharing agreements with all border & retirement states** No Yes Completeness*/** < 90% ≥ 90% ±95% Death clearance & follow-back*/** No DCO or >5% † 5% † 3% Case-finding audits in reporting facilities performed within past year** No Yes Data linked to health data (access/quality of care) and/or lifestyle behavioral risk factor data by registry or in partner with other state programs or departments*** No Yes Data Linkage & Availability Data linked to environmental risk factor databases by registry or in partner with other state programs or departments*** No Yes 20% Data linked to occupational risk factor databases by registry or in partner with other state programs or departments*** No Yes Raw data available to external public health researchers*** No Yes Informative geographic level*** County City/Zip Code Census Tracts/ Community-Level Block Groups Informative demographic categories at lowest geographic level*** < 2 variables 2 variables Gender, Race & Age Answers Informative tumor site data at lowest geographic level*** All sites combined Site Specific or All Sites Combined 20% Appropriate and Site Specific aggregations Written cancer cluster investigation protocols*** No Yes Complete case reporting from all facilities diagnosing or treating cancer** No Yes Complete case reporting from all practitioners diagnosing or treating cancer** No Yes State Legislation Access to medical records** No Yes Uniform reporting format** No Yes Protection of confidentiality** No Yes 10% Use of data for research** No Yes Release of data for research** No Yes Protection of individuals from liability** No Yes Cases reported to central registry within 6 months** No Yes *NAACCR registry certification program criteria; **NPCR program evaluation criteria; ***TFAH criteria 17 DESCRIPTION OF THE GRADING SYSTEM GRID Data Quality (50% of overall grade) 9) Missing age at diagnosis – No more than 3% of reported cases should be missing 1) Report all NPCR data items – data on an individual’s age. Standardized data collection is key to a surveillance system that ensures useful 10) Missing race – No more than 5% of data on every case of cancer is collected reported cases should be missing data and allows cancer rates to be compared on an individual’s race. and aggregated across state registries. 11) Missing sex – No more than 3% of 2) Use NAACCR standardized record layout reported cases should be missing data – Standardized record layouts ensure on an individual’s sex. that data can be compiled from different 12) Missing county/state – No more than registries. 3% of reported cases should be missing 3) Employ at least one certified tumor data on the county or state in which an registrar – Certified tumor registrars individual resides. receive specialized training and are essen- 13) Duplicates – No more than 1 case per tial for effective registry functioning. 2,000 should have duplicate records for 4) Written policies and procedures – Written each primary tumor. protocols available to registry staff with 14) Timely data – Complete, quality data varied experience and tenure are neces- need to be available within 23 months of sary to help ensure that registry opera- the diagnosis. tions are consistent, and of high quality. 15) Annual cancer reports – Annual cancer 5) Passing EDITS – EDITS is a software reports need to be made publicly avail- logic program developed by CDC and able within 12 months of the close of maintained by NAACCR that registries the diagnosis year. An annual report use to test the validity and logic of indi- contains at minimum, age-adjusted inci- vidual cancer records. dence rates and age-adjusted mortality 6) Text information included in computer- rates for the diagnosis year by sex for ized form as well as data codes – Text selected cancer sites and, where appro- information is vital for including more priate, by sex, race and ethnicity for information than is conveyed with stan- selected cancer sites.39 dardized coded data. 16) Interstate data sharing agreements with 7) Re-abstracting audits in reporting facili- all border and retirement states – Data ties performed within the past year – Re- sharing agreements ensure that if a state abstracting audits provide not only valu- resident is diagnosed outside of the able information regarding errors state, the diagnosis is reported back to reported in the original data file, but can the state registry where the patient per- prevent such errors in the future. manently resides. 8) Independent audits of the registry per- 17) Completeness – Registries should con- formed within the past five years – tain at least 90% of state residents diag- Reviews of registry performance by inde- nosed with cancer. Registries estimate pendent experts or academic users of the the percent of expected, unduplicated data provide a fresh perspective on reg- cases to provide a measure of the com- istry operations. pleteness of ascertaining all cases of can- cer diagnosed among state residents. 18 18) Death clearance and follow-back – No 4) Raw data available to researchers – States more than 5% of cases should be identi- need to have provisions in place to fied through death certificates. ensure that confidential case data can be Registries calculate the percent of cases provided to external researchers for the that are identified by death certificate purpose of public health research. only, meaning the case escaped the nor- Community-Level Answers mal reporting process through hospital (20% of overall grade) or clinic procedures and was identified only because the patient’s death certifi- 1) Informative geographic level – Cancer cate noted the type of cancer. data provided to citizens should be at an appropriate geographic resolution to 19) Case-finding audits in reporting facili- answer community-level questions, while ties performed within the past year – maintaining patient confidentiality. Data Case-findingaudits help to verify com- at too large a geographic level often plete reporting of cancer by diagnosing dilutes the ability to observe trends in a facilities and practitioners. localized area. Data Linkage & Availability 2) Informative demographic categories at (20% of overall grade) lowest geographic level – Data provided to the public should be stratified by 1) Data linked to health care intervention informative demographic variables, such data and/or lifestyle behavioral risk factor as age, sex and race, in order to observe databases by registry or in partner with possible trends, while maintaining other state programs or departments – patient confidentiality. Linking cancer registry data with health care intervention data and/or lifestyle 3) Informative tumor site data at lowest geo- behavioral risk factor data in order to graphic level – Data provided to the pub- help identify possible trends that warrant lic should contain overall cancer rates, as investigation or intervention. well as site-specific and or appropriate aggregated data (e.g., leukemias), while 2) Data linked to environmental risk factor maintaining patient confidentiality. databases by registry or in partner with other state programs or departments – 4) Written cluster investigation protocols – Linking cancer registry data with haz- The existence of a written cluster investi- ardous exposure data (air emissions, toxic gation protocol helps to ensure that the exposures from hazardous waste sites, state provides a thorough response to water quality, etc.) ensures that states are inquiries and concerns regarding a possi- using cancer registry data in order to ble cancer cluster. identify possible health trends that war- rant investigation or intervention. State Legislation (10% of overall grade) 3) Data linked to occupational risk factor databases by registry or in partner with Cancer registries require support and other state programs or departments – authority through state legislation and/or Linking cancer registry data with occupa- regulations for effective operation and utili- tional risk factor databases, such as occu- ty. All standards used are NPCR program pational rosters, ensures that cancer data evaluation standards. NPCR requires states is being used to identify trends in specif- that receive funds through their program to ic occupations that warrant investigation comply with standards 1-8 as delineated in or intervention. the Cancer Registries Amendment Act. 19 1) Complete case reporting from all facilities 6) Use of data for research – State legisla- diagnosing or treating cancer. State legis- tion/regulations should authorize the lation/regulations should mandate the state cancer registry or other persons or reporting of cancer cases to the registry organizations to conduct studies utilizing by all hospitals, or other facilities provid- registry data, including studies of sources ing screening, diagnostic or therapeutic and causes of cancer, evaluations of the services to patients with respect to cancer. cost, quality, efficacy and appropriateness of diagnostic, therapeutic, rehabilitative 2) Complete case reporting from all practi- and preventive services and programs tioners diagnosing or treating cancer. relating to cancer, and any other clinical, State legislation/regulations should man- epidemiological or cancer research. date that the reporting of cancer cases to the cancer registry by physicians, sur- 7) Release of data for research – State legis- geons and all other health care practi- lation/regulations should contain provi- tioners diagnosing or providing treat- sions by which confidential case data ment of cancer patients. may be disclosed to cancer researchers (in accordance with state law) for the 3) Access to medical records – State legisla- purposes of cancer prevention, control tion/regulations should mandate that and research. the registry is to have access to all med- ical records that identify cases of cancer 8) Protect individuals from liability – State or establish characteristics of the cancer, legislation/regulations should protect treatment of the cancer or medical status individuals who report information of the individual. regarding cancer cases to the registry or who have access to cancer case informa- 4) Uniform reporting format – State legisla- tion provided to the registry from liabili- tion/regulations should mandate that ty in any civil action. reporting facilities use standardized data formats when reporting data to the can- 9) Cases reported to central registry within cer registry. 6 months – State legislation/regulations should require that facilities and individ- 5) Protection of confidentiality – State legis- uals report cases to the cancer registry lation/regulations should mandate the within 6 months of the diagnosis. protection of confidentiality of all cancer case data reported to the cancer registry. 20 Appendix B: NAACCR CERTIFICATION LEVELS FOR 2000 INCIDENCE DATA40 NAACCR Certification Level State Gold Alaska Michigan California Minnesota Colorado Missouri Connecticut Nebraska District of Columbia Nevada Florida New Hampshire Georgia New Jersey Hawaii New York Idaho North Carolina Illinois Oregon Iowa Pennsylvania Kansas Rhode Island Kentucky Washington Louisiana West Virginia Maryland Wisconsin Massachusetts Wyoming Silver Alabama North Dakota Arizona Ohio Indiana South Carolina Montana Utah New Mexico IMPROVING CANCER PEER REVIEWERS TRACKING TODAY SAVES The opinions expressed in this report do not necessarily reflect the views of these individuals LIVES TOMORROW: or organizations. DO STATES MAKE THE GRADE? Division of Cancer Prevention and Control AUTHORS National Center for Chronic Disease Shelley A. Hearne, DrPH Prevention and Health Promotion Executive Director Centers for Disease Control and Prevention Trust for America’s Health Division of Cancer Control and Susan L. Polan, PhD Population Sciences Director of Government Relations National Cancer Institute Trust for America’s Health National Institutes of Health Laura M. Segal, MA Elizabeth Ward, PhD Director of Communications Director, Surveillance Research Trust for America’s Health Department of Epidemiology and Surveillance Research Patti J. Unruh American Cancer Society Senior Communications Associate Trust for America’s Health Daniel Wartenberg, PhD Professor CONTRIBUTING AUTHOR University of Medicine and Dentistry Richard W. Clapp, MPH, DSC New Jersey – Robert Wood Johnson Medical School TECHNICAL SUPPORT Department of Environmental and Tellus Institute, Environmental Health Community Medicine, Environmental and Occupational Health Sciences Institute Program and Cancer Institute of New Jersey 21 Endnotes 1 The Mellman Group, Inc. and Public Opinion 12 “2003 Program Fact Sheet, Cancer Registries: Strategies. “National Survey on Public The Foundation for Cancer Prevention and Perceptions of Risk from Bioterrorism and Control.” Centers for Disease Control and Other Health Threats,” Trust for America’s P r e v e n t i o n . < h t t p : / / w w w. c d c . g o v / Health. <http://healthyamericans.org/news- cancer/npcr/register.htm.> 19 August 2003. room/surveymemo0515.pdf.> 15 May 2002. 13 United States Cancer Statistics: 1999 2 “Cancer Facts & Figures 2003.” American Incidence. Atlanta (GA): Department of Cancer Society. Atlanta, Georgia. 2003. Health and Human Services U.S. Cancer Statistics Working Group, Centers for Disease 3 “Cancer Registries: The Foundation for Cancer Control and Prevention and National Cancer Prevention and Control.” Centers for Disease Institute; 2002. Control and Prevention. <http://www.cdc.gov/ nccdphp/bb_cancer_reg/.> 2001. 14 Weigs JW. “Epidemic Lung Cancer in Women,” The Journal of the American 4 “Cancer Facts & Figures 2003.” American Medical Association, 1977: 238:10. Cancer Society. Atlanta, Georgia. 2003. This estimate excludes noninvasive cancer of any 15 Ries LAG, Eisner MP, Kosary CL, Hankey BF, site except urinary bladder, and does not Miller BA, Clegg L, Mariotto A, Fay MP, Feuer include basal and squamous cell skin cancers. EJ, Edwards BK (eds). SEER Cancer Statistics Review, 1975 -2000. National Cancer Institute, 5 “Cancer Facts & Figures 2003.” American Bethesda, MD, <http://www.seer.cancer.gov/ Cancer Society. Atlanta, Georgia. 2003. csr/1975_2000/.> 2003. 6 Ibid. 16 “Cancer Registries: The Foundation for Cancer Prevention and Control AT-A- 7 TFAH calculation based on 1996-2000 prostate GLANCE 2001.” Centers for Disease Control cancer incidence rates of 164.3 per 100,000 for and Prevention. <http://www.cdc.gov/can- white males and 272.1 per 100,000 for African- cer/npcr/2001AAG-npcr.htm.> 2001. Americans. Source for prostate cancer incidence 17 Longnecker, et al. “Risk of Breast Cancer In rates: “Cancer Health Disparities,” National Relation To Lifetime Alcohol Consumption.” Cancer Institute. <http://www.cancer.gov/ news- J Natl Cancer Institute, 1995; 87:923-929. center/healthdisparities.> 23 April 2003. 18 One limitation of TFAH’s survey is that it 8 “Cancer Health Disparities.” National Cancer does not take into account the fact that state Institute. <http://www.cancer.gov/newscen- cancer tracking programs are at varying ter/healthdisparities.> 23 April 2003. stages of development and have varying levels of resources. Older programs may have had 9 “Cancer Facts & Figures 2003.” American many opportunities for data linkages and Cancer Society. Atlanta, Georgia. 2003. community interactions than newer registries that have only recently received federal fund- 10 “Importance of Cancer Registry.” National ing. Newer cancer tracking programs may Cancer Institute, <http://training.seer.can- therefore be focusing their attention and cer.gov/module_cancer_registration/unit1_i resources on establishing themselves and mportance.html.> Accessed 4 August 2003. meeting minimum quality standards. 11 Several states that participated in the Trust 19 Tennessee and South Dakota are the only two for America’s Health evaluation have reg- states receiving ‘capacity-building funds’ from istries (either state-wide or regional) that cur- NPCR for the purpose of developing cancer rently participate in the SEER program. registries where a limited or no statewide reg- These states include Alaska (Alaska Native istry previously existed (“NPCR Funded Tumor Registry), California, Georgia (the States.” Center for Disease Control and Atlanta and Rural Georgia Registries), Prevention. <www.cdc.gov/cancer/npcr/ Washington (Seattle-Puget Sound Registry) awards.htm>). 26 August 2003. and Michigan (Detroit registry). 22 20 South Dakota participated in TFAH’s study. 29 Haynes, Alfred M and Smedley, Brian D However, as the registry is presently in transi- (eds). The Unequal Burden of Cancer: An tion from a limited cancer data collection Assessment of NIH Research and Programs system to a statewide, population-based reg- for Ethnic Minorities and the Medically istry. Accordingly, the data provided were Underserved. Committee on Cancer not sufficient for grading in comparison to Research Among Minorities and the other states. Medically Underserved, Institute of Medicine. 1999. 21 “Cancer Registries: The Foundation for Cancer Prevention and Control AT-A- 30 “North American Association of Central GLANCE 2001.” Centers for Disease Control Cancer Registries Overarching Goals and and Prevention. <http://www.cdc.gov/can- Tasks,” NAACCR Leadership Retreat. cer/npcr/2001AAG-npcr.htm.> 2001. <http://www.naaccr.org/filesystem/pdf/Ove rarchinggoals2003-2005%2001-25-03.pdf.> 22 The Behavioral Risk Factor Surveillance October 2002. System (BRFSS) is a telephone survey con- ducted by all state health departments, the 31 “The New York State Cancer Surveillance District of Columbia, Puerto Rico, the Virgin Improvement Initiative.” New York State Health Islands, and Guam with assistance from CDC. Department. <http://www.health.state.ny.us/ The BRFSS is the largest continuously con- nysdoh/cancer/sublevel/csii.htm.> ducted telephone health survey in the world. December 2000. States use BRFSS data to track critical health problems and to develop and evaluate public 32 “2002 Program Fact Sheet Cancer Registries: health programs. The Foundation for Cancer Prevention and Control.” Centers for Disease Control and 23 “Cancer Surveillance System (NPCR-CSS).” Prevention. <http://www.cdc.gov/cancer/ Centers for Disease Control and Prevention. npcr/register2002.htm.> 2002. <http://www.cdc.gov/cancer/npcr/npcr- css.htm.> July 1999 33 “NPCR - The First 10 Years NPCR.” Centers for Disease Control and Prevention. 24 Trumbo, Craig W. “Public Requests for <http://www.cdc.gov/cancer/npcr/npcr- Cancer Cluster Investigations: A Survey of pdfs/handout.pdf.> 2002. State Health Departments,” American Journal of Public Health. August 2000: 90(8). 34 “Case-control Study of Childhood Cancers in Dover Township (Ocean County), New Jersey,” 25 Neutra, R.. “Reviews and Commentary: Volume II: Final Technical Report. Division of Counterpoint from a Cluster Buster,” Epidemiology, Environmental and American Journal of Epidemiology. July Occupational Health, New Jersey Department 1990: 132(1). of Health and Human Services in cooperation with the Agency for Toxic Substances and 26 Sankila R, et al. “Informed Consent in Disease Registry, U.S. Department of Health Cancer Registries,” Lancet (letter), 12 May and Human Services. <http://www.state.nj.us/ 2001: 357: 1536. health/eoh/hhazweb/case-control_pdf/ Volume_II/vol_ii.pdf.> January 2003. 27 Hodge, James G Jr. Letter to Holly L. Howe, Executive Director, North American 35 Hutton, M, Simpson, L, Miller, D, et al. Association Central Cancer Registries. “Progress Toward Nationwide Cancer <http://www.naaccr.org/filesystem/pdf/Aca Surveillance: An Evaluation of the National demic_Letter_Interpreting_HIPAA2.pdf.> Program of Cancer Registries: 1994-1999,” July 2001: 13. Journal of Registry Management. 2001: 28:3: 113-120. 28 “Cancer Health Disparities.” National Cancer Institute. <http://www.cancer.gov/newscen- ter/healthdisparities.> 23 April 2003. 23 36 “Criteria and Standards for NAACCR 39 “Program Announcement 02060: National Certification.” North American Association of Cancer Prevention and Control Program Central Cancer Registries. <http://www.naac- Notice of Availability of Funds.” Centers for cr.org/filesystem/pdf/finalcriteriaforRegistry Disease Control and Prevention. certificationpage.pdf.> Accessed 2 September <http://www.cdc.gov/cancer/dcpcpdf/Prog 2003. ram-Announcement-02060.pdf.> 2002: 45. 37 Data linkages conducted by a registry for the 40 “2002 Certified Registries of 2000 Incidence purpose of case ascertainment or data quality Data.” North American Association of were excluded. Central Cancer Registries. <http://www.naac- cr.org/index.asp?Col_SectionKey=12&Col_C 38 ‘Most informative’ is defined as the data: ontentID=98.> 2002. were at the census tract/census block level (or relevant aggregations); did include spe- cific demographic stratifications (race sex, age); and included cancer data for site spe- cific cancer sites and all sites combined. Data availability was also categorized into two addi- tional groups: ‘Least informative,’ defined as the data were: at the county level or higher; did not include all specific demographic stratifications (race, sex and age); and only included cancer data for all cancer sites com- bined, or ‘Moderately informative,’ defined as the data: were at the city/town zip-code level; did not include all specific demograph- ic stratifications (race, sex, age); and only included cancer data for site specific cancer sites or appropriate aggregations. States received 4 points for the most informative data, 2 points for moderately informative data and 1 point for least informative data. 24 1707 H Street, NW, 7th Floor Washington, DC 20006 (t) 202-223-9870 (f) 202-223-9871