Studies of hazardous waste sites and human health: evaluation of internal validity

Abstract

We conducted a systematic review of English-language analytical epidemiological studies assessing whether contaminants from hazardous waste sites cause adverse human health effects. From 900 candidate studies identified from computerized searches, bibliographic reviews, reference lists, and personal contacts, 43 potentially eligible studies were chosen. These were independently screened by two reviewers and 14 studies were selected for the final review. Each study was assessed independently by three reviewers. The methodology and main results were summarized. Studies were rated for different aspects of internal validity and criteria for causation. Inter-rater reliability was evaluated and differences were resolved by consensus. Most (11 of 14) studies selected for the final review were case-control studies. Hazardous exposures ranged from a single chemical in groundwater to multiple compounds in contaminated landfill sites. Health outcomes assessed included various cancers, various reproductive outcomes, and a variety of nonspecific health conditions and symptoms. Reported statistically significant relative risks (RRs) or odds ratios (ORs) were generally modest. Overall, for these 14 studies exposure measures were poorly rated; outcome measures were well rated and measurement bias, selection bias, and confounding biases were intermediately rated. Adherence to criteria for temporal association and biological plausibility was better than for other criteria of causality. The studies reviewed did not provide convincing evidence of causal relationships between hazardous waste site exposure and adverse human health effects, in particular because of poor exposure measurement. This does not mean that such causal relationships do not exist. Substantially better exposure assessment methodology is needed to establish the harmful effects caused by hazardous waste exposures.