Importance of responsible interpretation of observational studies with an impact on public health policy

Abstract

Observational studies are often referenced when informing policy decisions in public health. However, there limitations associated with the nature of observational studies may lead to erroneous interpretations and recommendations. A peer-reviewed publication published in 2018 reported that communities living near hog Concentrated Animal Feeding Operations (CAFO) have increased negative health outcomes and mortalities, in the U.S. state of North Carolina (NC). While authors stated that associations does not imply causation, speculative interpretation of their results by media caused detrimental effects on the swine industry. The objective here was to repeat their study using updated data and to evaluate the strength of conclusions and appropriateness of methods used with the ultimate goal of alerting on the impact that study limitations may have when used as evidence for decision-making. The referenced manuscript reported results from an ecological study design that used available data from 2007–2013. They analyzed exposure to CAFOs using a logistic regression model conducted at the individual-level, while presumably correcting for six confounding factors. We used same data sources and analytical methods, with additional years of data (2007–2018). Mortality, hospital admissions, and emergency department visits related eight diseases were analyzed (anemia, kidney disease, infectious diseases, tuberculosis, low birth weight, HIV, and diabetes). The study population and their exposure was defined by categorizing NC zip codes into three by pig density; where, > 1 hogs/km2 (G1), > 232hogs/km2 (G2), and no hogs (Control). Our repeated analysis recognized that the early study approach was inadequate to assess the intended hypothesis and had shortcomings including presence of ecological fallacy, residual confounding, inconsistency of associations, and over estimation of exposure. Diseases that are not causally relatable to hog CAFOs, such as HIV and diabetes, were also prominent in the communities neighboring CAFOs (G1 and G2). These associations, in part, reflects an alternative explanation of systemic health disparities of the communities near hog CAFOs. An ecological study would not confirm or deny exposure over time nor indicate causation. Hence, we emphasize the importance of responsible interpretation of ecological studies that concerns public health, agriculture, and economy, for science to provide appropriate and accurate support to the ultimate objective of targeting policy to enhance the quality of life and health of the communities.