To perform a meta-analysis of single nucleotide polymorphism needs to calculate gene frequency. This paper employs allele model as an example to introduce how to calculate gene frequency and display the process of a meta-analysis of single nucleotide polymorphism data using Review Manager 5.3 software.
Objective When making causal inferences in observational studies, in order to improve the robustness of the results of observational studies, statistical analysis techniques are often used to estimate the impact of unmeasured potential confounding factors. By systematically reviewing the application progress of the E-value, one of the sensitivity analysis methods, the advantages and limitations of using the E-value were discussed, to provide references for the application, reporting and interpretation of the E-value. Methods In the PubMed database, E-value was used as a keyword for title, abstract and key paper citation retrieval, and the literature that used the E-value analysis method for sensitivity analysis during 2016-2021 was screened. Results The E-value was widely used not only in cohort studies (n=215) and case-control studies (n=15), but also in cross-sectional studies (n=28), randomized controlled trials (n=6) and meta-analysis (n=16). The E-value was often combined with other sensitivity analysis methods, such as hierarchical analysis, instrumental variables, and multiple statistical regression models that correct different covariates, to further explore the reliability and robustness of the results. Conclusion When the E-value is used to evaluate the confounding factors in observational studies, the confidence interval and P value can be combined to evaluate the sensitivity of the results more comprehensively.
Survival data were widely used in oncology clinical trials. The methods used, such as the log-rank test and Cox regression model, should meet the assumption of proportional hazards. However, the survival data with non-proportional hazard (NPH) are also quite usual, which will decrease the power of these methods and conceal the true treatment effect. Therefore, during the trial design, we need to test the proportional hazard assumption and plan different analysis methods for different testing results. This paper introduces some methods that are widely used for proportional hazard testing, and summarizes the application condition, advantages and disadvantages of analysis methods for non-proportional hazard survival data. When the non-proportional hazard occurs, we need to choose the suitable method case by case and to be cautious in the interpretation of the results.
Objective Clinical trials for Alzheimer's disease feature long follow-up periods and high dropout rates, with missing outcome data being commonplace, which can impair the accuracy of treatment effect estimation. This study aimed to compare the applicability and performance of several commonly used and regulatory-recommended missing data handling strategies, including the mixed-effects model for repeated measures (MMRM), standard multiple imputation (MI), reference-based imputation (RBI), and δ-adjusted multiple imputation, in Alzheimer’s disease clinical trials. Methods The data were derived from a multicenter, randomized, double-blind, parallel-group, placebo-controlled clinical trial for Alzheimer’s disease. The endpoint was the Alzheimer's disease assessment scale-cognitive (ADAS-Cog) score, and the change from baseline in ADAS-Cog score at Week 26 was the primary outcome, and the difference between the treatment and placebo groups was estimated. The primary analysis used MMRM under the missing at random (MAR) assumption. Sensitivity analyses were performed using standard MI, reference-based imputation (J2R, CR, CIR), and δ-adjusted multiple imputation. Effect estimates, standard errors, confidence intervals, and P-values were compared across methods. Results Treatment effect estimates were consistent in direction across all methods. Compared with MMRM and MI under the MAR assumption, RBI yielded more conservative estimates under the missing not at random (MNAR) assumption. Under conservative δ settings the conclusions remained robust (all P-values <0.001), indicating that the findings were stable against deviations from MAR to MNAR. Conclusion In this clinical trial dataset, treatment effect inferences show good consistency and robustness across multiple missing data handling methods and missing data mechanisms. The analytical workflow proposed in this paper can serve as a reference for missing data handling and sensitivity analysis in clinical trials for neurodegenerative diseases.
Conducting real-world studies inevitably faces the challenge of bias. Researchers need to employ appropriate bias analysis methods to determine whether the strength of causal associations derived from the study is distorted by potential biases. Traditional qualitative assessments and statistical methods often struggle to effectively quantify the impact of unknown or unmeasured residual biases on study conclusions. Quantitative bias analysis can address this by constructing quantitative models to quantify sources of bias (such as misclassification or unmeasured confounding) and systematically evaluate the direction and magnitude of the effect of residual biases on effect estimates after conventional analysis, thereby confirming the robustness of the estimates. This method has gained widespread international recognition for addressing key questions such as, "How much uncertainty might potential biases introduce?" and "How credible are the study conclusions in the presence of bias?" However, there is still a lack of systematic introduction to this method domestically. Therefore, this article will begin with the practical challenges of bias analysis in real-world studies and sequentially elaborate on the development, theoretical framework, implementation cases, and application of quantitative bias analysis in real-world research. It aims to systematically introduce the methodological characteristics of quantitative bias analysis, providing researchers with a methodological reference for addressing bias analysis issues.