Special Issues in Using Epidemiologic Data Guidelines

Policy decisions and risk management strategies based upon risk assessments may have substantial societal implications. Consequently, approaches have been developed for standardizing the approach to risk assessment for specific classes of agents. These guidelines, such as those published by the Environmental Protection Agency, offer a framework for evaluating the available information, whether from epidemiologic or experimental research, and then using the data in the risk assessment process. The rising use of epidemiologic evidence in risk assessments with regulatory implications has led to concern that specific guidelines are needed for epidemiologic studies with regulatory implications. This section also addresses proposals for such guidelines.

Guidelines have been most extensively developed for determining carcinogenicity. Perhaps the longest-standing guidelines are those of the Interna tional Agency for Research on Cancer (IARC), which began a program in 1969 to evaluate the carcinogenic risk of chemicals to humans (WHO IARC 1972). The procedures of the agency only take the evaluation through the step of hazard identification, determining whether the evidence of carcingenicity is sufficient. The monographs on the chemicals do not attempt to quantify the risk, although dose-response relationships are considered as the causality of associations is assessed. The IARC working groups are charged with considering only epidemiologic studies of adequate quality and with considering the methodology of the study: definition of the study population, disease, and exposure; confounding and the selection of the comparison population; adequate description of the basic data; and use of proper analytic methods. The totality of the epidemiologic evidence is to be evaluated with criteria that emphasize the strength of the association, replication, dose response, and specificity. The guidelines also provide criteria for interpreting studies as showing a lack of carcinogenicity. In the final classification of a chemical, the strength of the evidence from both human and animal studies is considered and integrated for the overall evaluation.

One of the earliest attempts to develop regulatory guidelines was the report of the Interagency Regulatory Liaison Group (IRLG) (IRLG 1979). This wide-ranging report covered the types of input information needed to conduct a cancer risk assessment and offered recommendations for the evaluation of this information. It described the epidemiologic approach to studying carcinogens and indicated the need for attention to biologic plausibility, bias, confounding, and chance.

The Environmental Protection Agency has published a series of guidelines on cancer and other health endpoints. The Cancer Guidelines date to 1986, but proposed revisions have been published (US EPA 1986). The 1986 guidelines follow the structure of the IARC approach then in use and classified agents into five groups: group A (Human Carcinogen), group B (Probable Human Carcinogen), group C (Possible Human Carcinogen), group D (Not Classifiable as to Human Carcinogenicity), and group E (Evidence for Non-Carcinogenicity for Humans). The guidelines acknowledge the potentially unique contribution that human data can make but caution that epidemiologic data "are inherently capable of detecting only comparatively large increases in the relative risk of cancer." The criteria offered for assessing epidemiologic studies parallel those in the IARC guidelines. The agency's proposed new guidelines offer a markedly different classification scheme that replaces the five categories with verbal descriptors. Information on mode of action receives greater emphasis, reflecting the substantial evolution in the understanding of mechanisms of carcinogenesis. The guidelines offer 10 criteria for evaluating epidemiologic studies (Table 5-3). These criteria are fully

Table 5-3. Guidelines For Assessing Epidemiologic Studies (US Environmental Protection Agency's Cancer Guidelines)

1. Clear articulation of study objectives or hypothesis

2. Proper selection and characterization of the exposed and control groups

3. Adequate characterization of exposure

4. Sufficient length of follow-up for disease occurrence

5. Valid ascertainment the causes of cancer morbidity

6. Proper consideration of bias and confounding factors

7. Adequate sample size to detect and effect

8. Clear, well-documented, and appropriate methodology for data collection and analysis

9. Adequate response rate and methodology for handling missing data

10. Complete and clear documentation of results consistent with usual practices of epidemiologists in conducting research studies. The guidelines' scheme for weighing human evidence follows closely the widely applied criteria for causality.

More general guidelines for the conduct of epidemiologic research with regulatory implications have been proposed. The Chemical Manufacturers Association has published guidelines for "good epidemiologic practices" that provide specifications for both design and documentation. The calls for rigorous and standardized documentation of protocols in these recommendations and those of the new Cancer Guidelines of the Environmental Protection Agency are warranted, and current practices may often not meet these guidelines. The general need for guidelines was the topic of a 1994 conference (Graham 1995). In follow-up of this conference a set of guidelines entitled "Principles for Evaluating Epidemiologic Data in Regulatory Risk Assessment" was published in 1996. These guidelines were based on a panel of 18 individuals who were convened by Federal Focus Inc., a nonprofit foundation established to "engage in research and educational activities pertaining to Federal government policy issues, particularly ones of inter-agency concern."

The Federal Focus guidelines offer principles for assessing the utility of an epidemiologic study for risk assessment, for evaluating epidemiologic reports on cause-effect relationships, and for using human and animal data in dose-response evaluation. Checklists are provided for evaluating the adequacy of studies for the hazard identification of a risk assessment. The listed items are appropriate but criteria are not offered for their use nor are any tests of their application shown. Recommendations emphasizing interaction be tween epidemiologists and risk assessors are also made to improve the use of epidemiologic evidence in risk assessment.

Additional guidelines, focused on the dose-response step, were recently offered by Hertz-Picciotto (1995). She proposed a three-tiered evaluation scheme based on five criteria (Table 5-4). Category 1 studies were considered as offering a direct basis for deriving a dose-response relationship. Studies in this category are to meet four of five criteria. The first criterion, having a moderate-to-strong association, may inappropriately exclude studies of agents for which more modest effects are anticipated on a biological basis. Criteria 3 and 4 are reflective of the extent of the data that may be needed to address risk assessment needs. Studies in categories 2 and 3 are less informative than those in category 1 and cannot contribute directly to the dose-response step. Like other approaches to evaluating epidemiologic studies, application of the criteria could prove difficult and a test of the proposed classification has not yet been made.

These newer proposed guidelines, from the Federal Focus panel and Hertz-Picciotto, have not yet been applied, nor have they been endorsed by professional societies. In fact, professional organizations in epidemiology have remained silent on the use of epidemiologic evidence in risk assessment. While the guidelines remain untested, they are indicative of the scrutiny applied to epidemiologic findings in the risk assessment context and represent a starting point for broader discussion on evaluation of epidemiologic studies used in risk assessment.

0 0

Post a comment