The public health importance of clusters: a note from the Centers for Disease Control

Cancer clusters in the USA: what do the last twenty years of state and federal investigations tell us?

Background

Cancer clusters garner considerable public and legislative attention, and there is often an expectation that cluster investigations in a community will reveal a causal link to an environmental exposure. At a 1989 national conference on disease clusters, it was reported that cluster studies conducted in the 1970s and 1980s rarely, if ever, produced important findings. We seek to answer the question: Have cancer cluster investigations conducted by US health agencies in the past 20 years improved our understanding of cancer etiology, or informed cancer prevention and control?

Methods

We reviewed publicly available cancer cluster investigation reports since 1990, obtained from literature searches and by canvassing all 50 states and the District of Columbia. Investigations were categorized with respect to cancer type(s), hypothesized exposure, whether perceived clusters were confirmed (e.g. by elevated incidence), and conclusions about a link between cancer(s) of concern and hypothesized environmental exposure(s).

Results

We reviewed 428 investigations evaluating 567 cancers of concern. An increase in incidence was confirmed for 72 (13%) cancer categories (including the category “all sites”). Three of those were linked (with variable degree of certainty) to hypothesized exposures, but only one investigation revealed a clear cause.

Conclusions

It is fair to state that extensive efforts to find causes of community cancer clusters have not been successful. There are fundamental shortcomings to our current methods of investigating community cancer clusters. We recommend a multidisciplinary national dialogue on creative, innovative approaches to understanding when and why cancer and other chronic diseases cluster in space and time.

An update on cancer cluster activities at the Centers for Disease Control and Prevention

The Centers for Disease Control and Prevention (CDC) continues to be aware of the need for response to public concern as well as to state and local agency concern about cancer clusters. In 1990 the CDC published the "Guidelines for Investigating Clusters of Health Events," in which a four-stage process was presented. This document has provided a framework that most state health departments have adopted, with modifications pertaining to their specific situations, available resources, and philosophy concerning disease clusters. The purpose of this present article is not to revise the CDC guidelines; they retain their original usefulness and validity. However, in the past 15 years, multiple cluster studies as well as scientific and technologic developments have affected duster science and response (improvements in cancer registries, a federal initiative in environmental public health tracking, refinement of biomarker technology, cluster identification using geographic information systems software, and the emergence of the Internet). Thus, we offer an addendum for use with the original document. Currently, to address both the needs of state health departments as well as public concern, the CDC now a) provides a centralized, coordinated response system for cancer cluster inquiries, b) supports an electronic cancer cluster listserver, c) maintains an informative web page, and d) provides support to states, ranging from laboratory analysis to epidemiologic assistance and expertise. Response to cancer clusters is appropriate public health action, and the CDC will continue to provide assistance, facilitate communication among states, and foster the development of new approaches in duster science.

Things to know and do about cancer clusters

Perceived cancer clusters present difficulties and opportunities for clinicians and public health officials alike. Public health officials receive reports of perceived cancer clusters, evaluate the validity of these reports, and/or launch investigations to identify potential causes. Clinicians interact directly with the affected patients, families, or community representatives who question the occurrence of cancer and the underlying causes. Clinicians may identify cancer clusters when they question the unusual occurrence of a rare form of cancer within their practice or community. In addition, clinicians may be asked to discuss cancer clusters and inform local debates. In this paper, we describe the public health practice experience with cancer clusters and identify cancer prevention and control opportunities for clinicians and public health officials. Scientific investigations of cancer clusters rarely uncover new knowledge about the causes of cancer. However, a set of common characteristics, unique to etiologic cluster investigations have uncovered new information about the causes of cancer or demonstrated a preventable link to a known carcinogen. These characteristics may provide useful clues for sorting out the small number of clusters worthy of further scientific investigation. Public awareness of cancer clusters may promote an opportunity to inform and motivate people about the preventable causes of cancer and effective cancer screening methods.

Use of computer simulation to evaluate a putative cluster of genetic or teratologic outcomes: adjustment for "multiple hypotheses" and application to a reported excess of Down's syndrome

The identification of an apparent excess of a genetic outcome in a particular area and/or a particular time often provokes considerable public alarm about the presence of an environmental mutagen. It is often difficult to determine in any particular case whether the observation, whatever its nominal statistical significance, is due to chance concatenation of events or to an environmental factor. Statistical evaluation is made more difficult by the profuse number of possible hypotheses that could have triggered concern about an excess. This renders it difficult to calculate the actual probability of the observation (or one more extreme). By attempting to identify similar types of outcomes that could have provoked an apparent excess and then undertaking computer simulations assuming random deviations from a constant rate, one may attempt to adjust for the problem of multiple hypotheses. We apply this approach to a reported excess of Down's syndrome in Norway in 1985-1986 in younger mothers, and conclude that there is a high probability that it arose by chance.

Capture-recapture method to determine the best design of a surveillance system. Application to a thyroid cancer registry

The capture-recapture method is often confronted, when assessing completeness of surveys, to problems of dependence of data sources. The objective of this paper is to discuss the application of capture-recapture methods to choose the optimal combination of sources for a surveillance system. Our approach is based on: (1) using multiple sources, (2) assessing dependence between sources and between pools of dependent sources and other sources, (3) ruling out combinations that yield biased estimates, and (4) choosing the combinations of sources that have the best ratio between precision and applicability. We studied the independence for each pair of sources by computing the capture-recapture odds ratio. We characterized all combinations of sources by their sensitivity, coefficient of variation of the estimated number of cases, and level of resources needed to ascertain cases. Application of the approach is illustrated by data from a survey of thyroid cancer in New Caledonia, where five sources were used to estimate the incidence. The five sources provided 119 cases; the exhaustivity of sources and combinations of sources varied from 27.1 to 99.2%. Determination of dependence revealed ten dependencies out of 22 combinations. Coefficients of variation of the estimated number of cases varied from 0.83 to 27.79. The preferred combination included four sources and had a sensitivity of 97.5 and a coefficient of variation of 0.94. An assessment of dependence, based on simple criteria, can be used to choose the best combination of sources for a registry or a surveillance system.

Geographic analysis of pertussis infection in an urban area: a tool for health services planning

OBJECTIVE

This study examined whether incident cases of pertussis cluster in urban census tracts and identified community characteristics that predict high-incidence areas.

METHODS

An ecological study design was used. The study population included all persons diagnosed with pertussis from January 1, 1986, through December 31, 1994. Maps of rates were constructed via a geographic information system and clustering was statistically confirmed. Associations between pertussis rates and community characteristics were tested.

RESULTS

Mapping and statistical analysis revealed spatial clustering of pertussis. Higher age-adjusted rates of pertussis infection were associated with higher proportions of residents below poverty level.

CONCLUSIONS

In urban areas pertussis infection clusters in areas of poverty.

Molecular epidemiology study of a suspected community cluster of childhood cancers

We investigated the report of a community cluster of cancers in 33 children, which included two siblings known to have dominantly inherited Li-Fraumeni syndrome and a germline p53 mutation. After defining criteria for inclusion in the cluster, the 12 eligible childhood cancer probands diagnosed between 1980 and 1989 were not excessive (expected, ten cases). The corresponding childhood cancer mortality rates for the community fluctuated between 1950 and 1989 and were not increased overall. However, three additional probands had family histories of childhood cancer that suggested a forme fruste of Li-Fraumeni syndrome. The epidemiological data suggested a geographic cluster of this rare hereditary disorder, but absence of germline p53 mutation in the three other multicase families indicates genetic heterogeneity. Laboratory studies can assist analyses of suspected clusters, although investigations of geographic clusters of hereditary cancers raise complex issues of confidentiality and protection of affected individuals, their families, and the community.

Effect of relative risk and cluster configuration on the power of the one-dimensional scan statistic

The scan test for clustering in time is based on the maximum number of events in an interval (window) of width w as the window moves across the entire time frame. Power estimates of the scan statistic are simulated for a variety of epidemiologically motivated situations. Two cluster configurations are used: a rectangular pulse, and a triangular pulse designed to emulate environmental contamination. For a rectangular pulse, the relative risk R of disease in the cluster region is R-fold as high as it is for the background region. The power is strongly influenced by the sample size, the relative risk, and the width or duration of the cluster region, whereas the effect of the cluster configuration is small. Using a 5 per cent significance level, a relative risk of 4, a standardized cluster duration of 0.10, a relative window width of 1.5, and a (non-random) sample size of 50, the simulated power is approximately 80 per cent, indicating that the minimum sample size in the cluster region for adequate power is in the 12-32 range for values of the parameters used in this study.

Role of exposure databases in epidemiology

At present, exposure databases record data primarily for regulatory purposes; they have not focused on serving the needs of epidemiologists or public health. However, the modification of exposure databases could facilitate their use in epidemiology. Characteristics necessary to enhance the use of all databases include easy access by users; documentation of methods, sampling bias, error, and inconsistences; widespread coverage in time and space; and methods and measures for estimating exposure of individuals as well as populations. Also needed are exposure scenarios and models to estimate exposures for geographic areas and time intervals not currently sampled. Multidisciplinary teams are needed to examine current databases, to review strategies for improving data collection, and to suggest and help implement appropriate changes. A long-term goal is to develop and validate data from exposure scenarios and models using data on the relationship of exposure to doses measured in humans.