The cancer registry as a cancer-control tool (U.S.)

The cancer registry as a cancer-control tool

The Annual Report to the Nation on the Status of Cancer, 1976 to 2012, appears in this issue of Cancer.[1] This is the 18th year of the collaboration between scientists at the National Cancer Institute, the Centers for Disease Control and Prevention, the North American Association of Central Cancer Registries, and the American Cancer Society. It is an enduring public-private partnership that has paid tremendous dividends. Data from state and National Cancer Institute Surveillance, Epidemiology, and End Results program-funded registries are analyzed.

The study of registry data or vital registration is of tremendous value and often is underappreciated.[2] It can help define the cancer problem, determine where anticancer resources should be focused, and assess the effectiveness of cancer-control interventions. From a strong registry, trends in incidence, mortality, and survival can be documented, and risk factors for disease can be surmised. When registry data include cancer incidence and initial treatment, differences can be assessed in cancer treatment patterns.[3] Variation of care is a common cause of disparity. Documenting it is useful for health care providers, payers, and policy makers. Public disclosure creates pressure to improve the quality of care offered.

The assessment of registry data is an ecologic study. It can demonstrate whether a public health intervention is truly effective. Many of our public health interventions (preventative measures, screening tests, and treatments) are assessed in clinical studies that are divorced from reality. Clinical trials can demonstrate that an intervention is efficacious in the population enrolled, but they do not necessarily predict effectiveness when used in the general population.[4] The participants or patients in the general population may not be as healthy as those in the study, and the health care providers may not be as skilled.

This report to the nation documents significant declines in the age-adjusted death rate for the major causes of cancer death in the United States: cancers of the female breast, prostate, lung, and colorectal cancer. Although there has been some major cancer-control success, the data demonstrate that we can do better. There are some populations that have not shared as much in the declines. Most disconcerting, the report indicates that there are some who have not shared equally in the reduction in cancer death rates, and there are diseases in which mortality is increasing.

Registries have demonstrated a 40% decline in the age-adjusted breast cancer death rate over the past 25 years. Assessment demonstrates that the declines are because of improvements in treatment and early detection. Registries have also provided clear evidence of treatment differences and state-by-state variation in the mortality decline. This leads to the conclusion that we can do better for all by better disseminating current treatment technologies to the entire population.[5]

Lung cancer death rates have declined significantly for men over the past 25 years and have started to decline for women in the past decade. This is largely because the tobacco epidemic is modestly receding. The state-by-state variation in tobacco use today and the state-by-state differences in tobacco-related disease outcomes reveal an opportunity to do better.[6]

Lung screening has recently become available, but it is somewhat controversial. The American Cancer Society and the US Preventive Services Task Force have recommended that individuals at high risk for lung cancer because of age and smoking history with access to high-quality screening and treatment should consider screening after being counseled of the potential benefits and risks.[7, 8] The landmark study demonstrating a 20% relative risk reduction in screening also demonstrates significant harms when the screening is done even in the best of centers.[9] This includes very high false-positive rates and even some deaths from the workup of true-positive and false-positive screens.

Colorectal cancer rates have declined by 43% over the past 25 years. Studies of attributable fractions indicate that screening is responsible for 53% of the decline, changes in risk factors account for 35%, and 12% of the decline is because of improvements in treatment.[10]

There is substantial opportunity to save even more individuals from death because of colorectal cancer. States in the northeast of the United States (such as Massachusetts and New York) have had declines >40%, whereas states in the south (such as Mississippi and Alabama) have had declines <10%.[11] Nearly 40% of Americans aged >50 years are not up to date on colorectal cancer screening, and a substantial minority of those diagnosed receive less than optimal treatment.[12]

There has been debate regarding the value and accuracy of the 53% decline in age-adjusted prostate cancer mortality.[13] Indeed, the debate demonstrates some of the limitations of registry data. Registries can demonstrate that the decline in prostate cancer mortality began too early to be attributed solely to screening.[14] There is speculation that some of the decline is artifact.[15] The rise in mortality that started in the 1970s and the decline that began in the early 1990s coincided with World Health Organization-instituted changes in the coding for causes of death. Improvements in treatment, the use of prostate-specific antigen for assessment of prostate cancer progression, and the increasing use of hormone therapies may also have contributed to the decline in mortality.[16] Compounding the issue, some of these hormone therapies may reduce the risk of prostate cancer death by increasing the risk of death from cardiac disease.[17]

Each year, the scientists involved in this report do a deep dive on 1 specific issue in cancer control.[1] This year, it is liver cancer. Liver cancer is the second most common cause of cancer death worldwide. For years, liver cancer was less a problem in the United States, but its incidence and mortality rates are now rising.

Major risk factors for liver cancer include alcohol intake, obesity, diabetes, and infection with viral hepatitis B (HBV) and hepatitis C (HCV). Worldwide, 60% of all liver cancers are caused by infection, primarily with HBV and HCV. It is noteworthy that HBV and HCV cause approximately 30% of liver cancers in the United States. The most common cause in the United States is fatty liver disease as a consequence of obesity and diabetes.

All of the major causes of liver cancer are modifiable. Success in stemming the US epidemic of obesity can lower rates of liver cancer. Simple precautions can prevent infection. There is evidence that the HBV vaccine is at least partially responsible for declines in HBV infection and liver cancer. Treatment for viral hepatitis infection is often successful and can prevent liver cancer, but it is very expensive.

The US Centers for Disease Control and Prevention and the US Preventive Services Task Force have recommended screening for the cohorts at highest risk for HCV and HBV. This involves 1-time testing for HCV infection in all Americans born between 1945 and 1965 and HBV screening for individuals living in the United States but born in sub-Saharan Africa, Asia, and countries in which HBV infection prevalence is ≥2%.

The report also discusses cancer incidence and mortality rates by race. One limitation of registry data is how populations are categorized. In the United States, Federal law requires the US government to collect health statistics using racial categories that are defined by the US Office of Management and Budget 2 to 3 years before every decennial census.[18] These racial and ethnic definitions have changed literally every 10 years.

The Office of Management and Budget says these categories are sociopolitical groupings that are not based in biology. They are categories of convenience.
Some, such as Asian/Pacific Islander, make little sense geographically, culturally, or biologically.[19] Even the more recent move to split Asian/Pacific Islander into Asian and the separate category “Native Hawaiian and Pacific Islander” creates challenges. Still lost in the averaging of data are the generally high prostate and lung cancer incidence rates among Samoan Americans, the high breast and colorectal cancer rates among Native Hawaiians, and the very high rates of liver cancer among Laotian Americans and stomach cancer among Korean Americans. The subdivision of categories also creates the problem of small numbers and especially limits the assessment of mortality data.

Misattribution of race is a significant issue.[2] Beginning with the 2000 Census, individuals were allowed to choose multiple categories. Although patients often declare their race while they are alive and receiving health care, the health care provider usually does it for the dead. The misattribution of Hispanic ethnicity among Filipino patients with Spanish surnames in mortality data is well documented. All of this makes the study of cancer in populations, and especially the study of long-term trends, difficult.

The Institute of Medicine notes that there is tremendous promise in the focused epidemiological study of better defined populations. Findings from such research will likely transcend to other populations. We commonly describe populations by sex and race but can also use ethnicity, area of residence, area of geographic origin, and socioeconomic status. Currently, most registries do not collect data on area of geographic origin or socioeconomic status. However, area of geographic origin, compared with race, aligns far better with genetics; and a correlation (if not a direct influence) has been demonstrated between socioeconomic status or social deprivation over a period of years with cancer etiology and genomics.[20]

Today's cancer registry is a powerful public health and cancer-control tool. The future of the cancer registry involves acquisition of big data from the electronic medical record. Harmonization of record systems eventually will permit the direct transfer of data. The future registry will allow for the correlation of molecular genomics and outcomes. This will provide new insights into risk and effectiveness. It will make race and some other population categorizations less relevant or irrelevant. The challenge will be to use the technology wisely, as privacy concerns are already threatening some registries.