Bias Analysis Gone Bad

Socioeconomic Status, Smoking, and Lung Cancer: Mediation and Bias Analysis in the SYNERGY Study

BACKGROUND

Increased lung cancer risks for low socioeconomic status (SES) groups are only partially attributable to smoking habits. Little effort has been made to investigate the persistent risks related to low SES by quantification of potential biases.

METHODS

Based on 12 case-control studies, including 18 centers of the international SYNERGY project (16,550 cases, 20,147 controls), we estimated controlled direct effects (CDE) of SES on lung cancer via multiple logistic regression, adjusted for age, study center, and smoking habits and stratified by sex. We conducted mediation analysis by inverse odds ratio weighting to estimate natural direct effects and natural indirect effects via smoking habits. We considered misclassification of smoking status, selection bias, and unmeasured mediator-outcome confounding by genetic risk, both separately and by multiple quantitative bias analyses, using bootstrap to create 95% simulation intervals (SI).

RESULTS

Mediation analysis of lung cancer risks for SES estimated mean proportions of 43% in men and 33% in women attributable to smoking. Bias analyses decreased the direct effects of SES on lung cancer, with selection bias showing the strongest reduction in lung cancer risk in the multiple bias analysis. Lung cancer risks remained increased for lower SES groups, with higher risks in men (fourth vs. first [highest] SES quartile: CDE, 1.50 [SI, 1.32, 1.69]) than women (CDE: 1.20 [SI: 1.01, 1.45]). Natural direct effects were similar to CDE, particularly in men.

CONCLUSIONS

Bias adjustment lowered direct lung cancer risk estimates of lower SES groups. However, risks for low SES remained elevated, likely attributable to occupational hazards or other environmental exposures.

Core Concepts in Pharmacoepidemiology: Quantitative Bias Analysis

Pharmacoepidemiological studies provide important information on the safety and effectiveness of medications, but the validity of study findings can be threatened by residual bias. Ideally, biases would be minimized through appropriate study design and statistical analysis methods. However, residual biases can remain, for example, due to unmeasured confounders, measurement error, or selection into the study. A group of sensitivity analysis methods, termed quantitative bias analyses, are available to assess, quantitatively and transparently, the robustness of study results to these residual biases. These approaches include methods to quantify how the estimated effect would be altered under specified assumptions about the potential bias, and methods to calculate bounds on effect estimates. This article introduces quantitative bias analyses for unmeasured confounding, misclassification, and selection bias, with a focus on their relevance and application to pharmacoepidemiological studies.

Reliability and validity of the online Pittsburgh sleep quality index in college students from low-income regions

Objectives

We aimed to test the reliability and structural validity (also called dimensionality) of the online Pittsburgh Sleep Quality Index among college students from low-income regions.

Methods

We assessed 195 Brazilian college students from a low-income region (Gini index of 0.56), of whom 117 were reassessed to evaluate the reliability. We collected all data in a self-reported online twice, 2-week apart. We evaluated reliability and structural validity.

Results

All questionnaire components showed reliability, correlation coefficient ≥0.49. In the structural validity, the confirmatory analysis showed better global model adjustment for the one-factor (RMSEA = 0.019; SRMR = 0.041; CFI = 0.992; TLI = 0.986) solution compared with two-factor (RMSEA = 0.099; SRMR = 0.070; CFI = 0.764; TLI = 0.619) and three-factor (RMSEA = 0.108; SRMR = 0.066; CFI = 0.763; TLI = 0.548) solutions, respectively.

Discussion

The online questionnaire presents acceptable reliability and structural validity in Brazilian low-income regions.

Higher-order evidence

Higher-order evidence is evidence about evidence. Epidemiologic examples of higher-order evidence include the settings where the study data constitute first-order evidence and estimates of misclassification comprise the second-order evidence (e.g., sensitivity, specificity) of a binary exposure or outcome collected in the main study. While sampling variability in higher-order evidence is typically acknowledged, higher-order evidence is often assumed to be free of measurement error (e.g., gold standard measures). Here we provide two examples, each with multiple scenarios where second-order evidence is imperfectly measured, and this measurement error can either amplify or attenuate standard corrections to first-order evidence. We propose a way to account for such imperfections that requires third-order evidence. Further illustrations and exploration of how higher-order evidence impacts results of epidemiologic studies is warranted.

Exploring links between Chinese military recruits' psychological stress and coping style from the person-environment fit perspective: The chain mediating effect of self-efficacy and social support

The choice of coping style of recruits under psychological stress in the process of military task execution has been an important topic in the promotion of military operations and cohesion of military forces. Taking a positive coping style under psychological stress can help recruits overcome the negative effects of stress and improve military morale and group combat effectiveness. Although soldiers' psychological stress in the process of military mission execution having an impact on coping style has been studied by a large body of literature, very little literature has focused on the mechanism of self-efficacy and social support between recruits' psychological stress and coping style from the person-environment fit perspective. Therefore, this study was conducted to analyze the impact of recruits' psychological stress on coping style through a chain mediation model and to discuss the role of self-efficacy and social support in this relationship. Two waves of survey data were utilized to test the research hypotheses on a sample of 1028 Chinese recruits performing military tasks. The results indicated that recruits' psychological stress negatively impacted positive coping styles and positively correlated with negative ones. In addition, self-efficacy and social support mediated the relationship between psychological stress and positive coping style, and self-efficacy mediated the relationship between psychological stress and negative coping style. More importantly, self-efficacy and social support play the chain mediating effect between psychological stress and positive coping style.