Estimating uncertainty in observational studies of associations between continuous variables: example of methylmercury and neuropsychological testing in children

A systematic review of quantitative bias analysis applied to epidemiological research

BACKGROUND

Quantitative bias analysis (QBA) measures study errors in terms of direction, magnitude and uncertainty. This systematic review aimed to describe how QBA has been applied in epidemiological research in 2006-19.

METHODS

We searched PubMed for English peer-reviewed studies applying QBA to real-data applications. We also included studies citing selected sources or which were identified in a previous QBA review in pharmacoepidemiology. For each study, we extracted the rationale, methodology, bias-adjusted results and interpretation and assessed factors associated with reproducibility.

RESULTS

Of the 238 studies, the majority were embedded within papers whose main inferences were drawn from conventional approaches as secondary (sensitivity) analyses to quantity-specific biases (52%) or to assess the extent of bias required to shift the point estimate to the null (25%); 10% were standalone papers. The most common approach was probabilistic (57%). Misclassification was modelled in 57%, uncontrolled confounder(s) in 40% and selection bias in 17%. Most did not consider multiple biases or correlations between errors. When specified, bias parameters came from the literature (48%) more often than internal validation studies (29%). The majority (60%) of analyses resulted in >10% change from the conventional point estimate; however, most investigators (63%) did not alter their original interpretation. Degree of reproducibility related to inclusion of code, formulas, sensitivity analyses and supplementary materials, as well as the QBA rationale.

CONCLUSIONS

QBA applications were rare though increased over time. Future investigators should reference good practices and include details to promote transparency and to serve as a reference for other researchers.

A New Method for Partial Correction of Residual Confounding in Time-Series and Other Observational Studies

Methods exist to detect residual confounding in epidemiologic studies. One requires a negative control exposure with 2 key properties: 1) conditional independence of the negative control and the outcome (given modeled variables) absent confounding and other model misspecification, and 2) associations of the negative control with uncontrolled confounders and the outcome. We present a new method to partially correct for residual confounding: When confounding is present and our assumptions hold, we argue that estimators from models that include a negative control exposure with these 2 properties tend to be less biased than those from models without it. Using regression theory, we provide theoretical arguments that support our claims. In simulations, we empirically evaluated the approach using a time-series study of ozone effects on asthma emergency department visits. In simulations, effect estimators from models that included the negative control exposure (ozone concentrations 1 day after the emergency department visit) had slightly or modestly less residual confounding than those from models without it. Theory and simulations show that including the negative control can reduce residual confounding, if our assumptions hold. Our method differs from available methods because it uses a regression approach involving an exposure-based indicator rather than a negative control outcome to partially correct for confounding.

A proposal for assessing study quality: Biomonitoring, Environmental Epidemiology, and Short-lived Chemicals (BEES-C) instrument

The quality of exposure assessment is a major determinant of the overall quality of any environmental epidemiology study. The use of biomonitoring as a tool for assessing exposure to ubiquitous chemicals with short physiologic half-lives began relatively recently. These chemicals present several challenges, including their presence in analytical laboratories and sampling equipment, difficulty in establishing temporal order in cross-sectional studies, short- and long-term variability in exposures and biomarker concentrations, and a paucity of information on the number of measurements required for proper exposure classification. To date, the scientific community has not developed a set of systematic guidelines for designing, implementing and interpreting studies of short-lived chemicals that use biomonitoring as the exposure metric or for evaluating the quality of this type of research for WOE assessments or for peer review of grants or publications. We describe key issues that affect epidemiology studies using biomonitoring data on short-lived chemicals and propose a systematic instrument--the Biomonitoring, Environmental Epidemiology, and Short-lived Chemicals (BEES-C) instrument--for evaluating the quality of research proposals and studies that incorporate biomonitoring data on short-lived chemicals. Quality criteria for three areas considered fundamental to the evaluation of epidemiology studies that include biological measurements of short-lived chemicals are described: 1) biomarker selection and measurement, 2) study design and execution, and 3) general epidemiological study design considerations. We recognize that the development of an evaluative tool such as BEES-C is neither simple nor non-controversial. We hope and anticipate that the instrument will initiate further discussion/debate on this topic.

Atrazine and pregnancy outcomes: a systematic review of epidemiologic evidence

Atrazine (ATR) is a commonly used agricultural herbicide that has been the subject of epidemiologic studies assessing its relation to reproductive health problems. This review evaluates both the consistency and the quality of epidemiologic evidence testing the hypothesis that ATR exposure, at usually encountered levels, is a risk factor for birth defects, small for gestational age birth weight, prematurity, miscarriages, and problems of fetal growth and development. We followed the current methodological guidelines for systematic reviews by using two independent researchers to identify, retrieve, and evaluate the relevant epidemiologic literature on the relation of ATR to various adverse outcomes of birth and pregnancy. Each eligible paper was summarized with respect to its methods and results with particular attention to study design and exposure assessment, which have been cited as the main areas of weakness in ATR research. As a quantitative meta-analysis was not feasible, the study results were categorized qualitatively as positive, null, or mixed. The literature on ATR and pregnancy-related health outcomes is growing rapidly, but the quality of the data is poor with most papers using aggregate rather than individual-level information. Without good quality data, the results are difficult to assess; however, it is worth noting that none of the outcome categories demonstrated consistent positive associations across studies. Considering the poor quality of the data and the lack of robust findings across studies, conclusions about a causal link between ATR and adverse pregnancy outcomes are not warranted.

Bisphenol A and indicators of obesity, glucose metabolism/type 2 diabetes and cardiovascular disease: a systematic review of epidemiologic research

INTRODUCTION

Bisphenol A (BPA), a high-volume chemical with weak estrogenic properties, has been linked to obesity, cardiovascular diseases (CVD) and diabetes mellitus (DM). This review evaluates both the consistency and the quality of epidemiological evidence from studies testing the hypothesis that BPA exposure is a risk factor for these health outcomes.

METHODS

We followed the current methodological guidelines for systematic reviews by using two independent researchers to identify, review and summarize the relevant epidemiological literature on the relation of BPA to obesity, CVD, DM, or related biomarkers. Each paper was summarized with respect to its methods and results with particular attention to study design and exposure assessment, which have been cited as the main areas of weakness in BPA epidemiologic research. As quantitative meta-analysis was not feasible, the study results were categorized qualitatively as positive, inverse, null, or mixed.

RESULTS

Nearly all studies on BPA and obesity-, DM- or CVD-related health outcomes used a cross-sectional design and relied on a single measure of BPA exposure, which may result in serious exposure misclassification. For all outcomes, results across studies were inconsistent. Although several studies used the same data and the same or similar statistical methods, when the methods varied slightly, even studies that used the same data produced different results.

CONCLUSION

Epidemiological study design issues severely limit our understanding of health effects associated with BPA exposure. Considering the methodological limitations of the existing body of epidemiology literature, assertions about a causal link between BPA and obesity, DM, or CVD are unsubstantiated.

Climate change and environmental impacts on maternal and newborn health with focus on Arctic populations

BACKGROUND

In 2007, the Intergovernmental Panel on Climate Change (IPCC) presented a report on global warming and the impact of human activities on global warming. Later the Lancet commission identified six ways human health could be affected. Among these were not environmental factors which are also believed to be important for human health. In this paper we therefore focus on environmental factors, climate change and the predicted effects on maternal and newborn health. Arctic issues are discussed specifically considering their exposure and sensitivity to long range transported contaminants.

METHODS

Considering that the different parts of pregnancy are particularly sensitive time periods for the effects of environmental exposure, this review focuses on the impacts on maternal and newborn health. Environmental stressors known to affects human health and how these will change with the predicted climate change are addressed. Air pollution and food security are crucial issues for the pregnant population in a changing climate, especially indoor climate and food security in Arctic areas.

RESULTS

The total number of environmental factors is today responsible for a large number of the global deaths, especially in young children. Climate change will most likely lead to an increase in this number. Exposure to the different environmental stressors especially air pollution will in most parts of the world increase with climate change, even though some areas might face lower exposure. Populations at risk today are believed to be most heavily affected. As for the persistent organic pollutants a warming climate leads to a remobilisation and a possible increase in food chain exposure in the Arctic and thus increased risk for Arctic populations. This is especially the case for mercury. The perspective for the next generations will be closely connected to the expected temperature changes; changes in housing conditions; changes in exposure patterns; predicted increased exposure to Mercury because of increased emissions and increased biological availability.

CONCLUSIONS

A number of environmental stressors are predicted to increase with climate change and increasingly affecting human health. Efforts should be put on reducing risk for the next generation, thus global politics and research effort should focus on maternal and newborn health.