Efforts to provide for public health protection from environmental contaminants in fish have resulted in various advisories or recommendations with regard to fish consumption from local, state, and federal agencies. These advisories are based on measured levels of contaminants in fish that are combined with values for body weight and portion size to produce an estimate of an "acceptable" consumption frequency (e.g., eat no more than once per month). Because values for body weight and portion size are generally generic default values, they do not necessarily relate to a specific population or to any individual in that population. Thus, the use of default values may result in underprotection or overprotection in any given case. Given the benefits of fish consumption and the risks from overexposure to various toxicants, vigilance is required by custodians of public health to ensure that populations are protected while being cautious not to over- or underprotect them. In this analysis, we examine the "acceptable" consumption limits derived for fish species/groups consumed by three specific populations and determine the extent of public health protection afforded by these limits. To accomplish this, the "acceptable" consumption frequencies are derived based, in part, on default assumptions and are compared to intakes calculated from empirically derived species-specific individual consumption and demographic data. Sensitivity analyses and population-specific probabilistic assessments of exposure are conducted to identify those values and/or assumptions which might significantly influence the resulting fish consumption advisories. Three populations were chosen for study based on their ability to represent populations of greatest concern: those most sensitive and/or those most exposed. We conclude from this investigation that consumption pattern data, contaminant data and body weight data together can be used to make fish consumption advisories more focused and, therefore, less likely to be under- or overprotective.(c) 2004 Elsevier Inc. All rights reserved.
