BMC Medical Research Methodology最新文献

Background: This report presents the study design and recruitment outcomes for the 'Health in Germany' panel, a long-term population-based health survey infrastructure developed by the Robert Koch Institute. The initial recruitment was conducted using a stratified random sample of the German population and a mixed-mode approach combining web-based and paper questionnaires. We examine participation rates across demographic subgroups, assess sample composition, and analyze potential selection effects.

Method: The panel recruitment survey for the 'Health in Germany' panel utilized the residents' registration offices as the sampling frame. A two-stage stratified (cluster) sample was drawn from 359 primary sampling units across Germany. A mixed-mode approach was employed, offering both online and paper survey-mode on the basis of age groups. The sequence of survey modes was differentiated by age groups based on information from the residents' registration offices. For respondents aged 16-69 years, a sequential mixed-mode design, offering the online mode first and only with the second reminder the paper survey-mode (also called push-to-web strategy), was applied. For respondents aged 70 years and older a simultaneous mixed-mode design was used, offering the online and paper survey-mode with the invitation. Participants completed a first panel recruitment survey in which socio-demographic and health data were collected. Selection effects and sample composition were analyzed via logistic regression models and compared with official population data.

Results: A total of 62,556 interviews were conducted, with 49,766 participants consenting to join the panel. After double opt-in registration process (for online participants), 47,863 remained. Response rates were higher among women and younger participants. Online participation predominated in the sequential design, whereas offline participation was more common in the simultaneous design. Logistic regression indicated higher participation among women and residents of smaller municipalities. Overall, the sample composition aligned broadly with population benchmarks, except for education and citizenship.

Conclusions: The first recruitment study for the 'Health in Germany' panel established one of Europe's largest population-based health panels through a mixed-mode design. Future expansions include regular health surveys, biometric measurements, complementary self-recruitment alongside the probability-based sample, and the development of a mobile survey app.

Background: Reporting treatment effects from clinical trials on both relative and absolute scales is crucial. While absolute measures like the Number Needed to Treat (NNT) are well-established for binary outcomes, their calculation for time-to-event outcomes remains challenging due to time-dependence, which hinders interpretation and communication. Traditional additive hazard models, while addressing time-dependence, have been limited by restrictive assumptions regarding outcome distributions.

Methods: This paper proposes to use a recently introduced class of parametric additive hazard models to compute time-constant absolute effect measures for time-to-event outcomes. These models allow for a wide range of parametric distributions, overcoming the limitations of previous approaches. The approach provides a single, absolute effect size (e.g., hazard difference or NNT) summarizing the effect over the entire study duration. We illustrate this method using digitized Kaplan-Meier data from the EMPA-REG OUTCOME trial, focusing on all-cause mortality, and fit six different parametric distributions (exponential, linear hazard rate, Weibull, log-logistic, Gompertz, and Gamma-Gompertz).

Results: Despite notable differences in model fit across the six distributions, the estimated rate differences, corresponding NNTs, and their confidence intervals were remarkably similar. The linear hazard rate and Gompertz models, which provided the best fit according to the BIC, yielded a rate difference of -8.8 per 1,000 person-years, with an NNT of 114. These models also demonstrated increasing hazards, aligning with expectations for all-cause mortality. The estimated modes of the distributions from the best-fitting models (10.4 and 13.0 years) were more plausible than those from simpler models.

Conclusions: The class of parametric additive hazard models offers a valuable tool for calculating time-constant absolute effect measures for time-to-event outcomes. This approach effectively addresses the issues of time-dependence and limited distribution flexibility, providing a single, interpretable absolute effect size. Future work could explore more general distributions and further derivation of absolute effect measures on the time scale.

Background: Electronic health record (EHR)-based observational studies can rapidly provide real-world data on vaccine effectiveness (VE), though EHR data may be prone to misclassification and unmeasured confounding.

Methods: In VEBIS-EHR, a retrospective multi-country COVID-19 VE cohort study, we examined unmeasured confounding using a negative control outcome (death not related to COVID-19) and misclassification due to timing of data extraction. The evaluation spanned two periods (November-December 2023, January-February 2024), encompassing up to 18.7 million individuals across six EU/EEA countries. Vaccine confounding-adjusted hazard ratios (aHRs) were pooled using random-effects meta-analysis.

Results: aHRs against non-COVID-19 mortality ranged from 0.35 (95% CI: 0.28-0.44) to 0.70 (0.66-0.73) when comparing vaccinated versus unvaccinated. Delaying EHR data extraction modestly increased the capture of outcome and exposure events, with some variation by vaccination status. Site-level fluctuations in aHRs did not meaningfully alter the overall pooled VE, suggesting stable estimates despite misclassification related to extraction timing.

Conclusions: We observed some evidence of unmeasured confounding when using non-COVID-19 deaths as a negative outcome, though the specificity of our negative control must be considered. This result may suggest overestimation of VE, but also the need for further analysis with more specific negative control outcomes and confounding-adjustment techniques. Addressing such confounding using richer data sources and more refined approaches remains critical to ensure accurate, timely VE estimates based on retrospective cohorts constructed using registry data. Extending the delay between the end of observation and data extraction modestly improves the completeness of exposure and outcome data, with limited effect on pooled VE estimates.

Background: Early cancer diagnosis is essential for improving prognosis and guiding treatment. However, the high dimensionality and complexity of omics data present major challenges. Computational approaches that extract stable biomarkers and enable reliable classification across cancer types and stages are needed.

Methods: A novel feature selection method, sDCFE (synergistic Discriminative Cluster-based Feature Extraction), was developed by extending Fisher-like variance analysis with a median absolute deviation (MAD) regularization term and a cluster separation component to enhance robustness and interpretability. Features selected by sDCFE were compared with those obtained from XGBoost, and the intersected set of 82 genes was evaluated through functional enrichment (KEGG, Reactome, GO BP), survival analysis (Kaplan-Meier, Cox regression), and biomarker novelty assessment against six external resources. Hybrid classification models integrating XGBoost, sDCFE, and deep learning were applied to pancancer classification, and the framework was further extended to lung squamous cell carcinoma (LUSC) staging using RNA-seq and methylation data.

Results: The overlap between sDCFE and XGBoost yielded 82 candidate biomarkers enriched in cancer-related pathways, including cell cycle regulation, immune signalling, and DNA repair. Novelty assessment stratified these genes into established, emerging, and novel categories. Six genes-HFE2, LOC339674, SERINC2, SFTA3, SOX2OT, and ACPP-emerged as the most promising candidates, supported by enrichment and survival associations across multiple cancers. The hybrid model achieved near-perfect pancancer classification on TCGA (accuracy = 99.3%, MCC = 0.992, AUC = 1.0) and demonstrated strong generalizability on PCAWG (accuracy = 94%, MCC = 0.929, AUC = 0.997). In the LUSC staging task, multiomics integration improved classification performance: the CNN-based model reached 84% accuracy, while logistic regression applied to sDCFE-ranked features achieved 88.5% accuracy with superior calibration, highlighting the robustness of the selected features.

Conclusion: sDCFE provides a principled extension of Fisher-like methods, enabling stable and interpretable biomarker selection. When combined with XGBoost and deep learning, the framework achieves highly accurate and biologically grounded cancer classification across both cancer types and stages. The identification of novel and prognostic biomarkers, including HFE2, LOC339674, SERINC2, SFTA3, SOX2OT, and ACPP, underscores its translational potential. These results position the framework as a promising precision oncology tool to support early diagnosis, risk stratification, and treatment decision-making.

Background: There is growing interest in utilizing Bayesian approaches to borrow information across tumor types in basket trials. Several innovative designs, primarily extensions of the Bayesian hierarchical model (BHM), have been proposed to dynamically borrow information based on observed data. However, there is no recognized solution to quantify the degree of information borrowing in such a context, posing a great challenge for non-statisticians to understand these complex designs.

Methods: The tool of effective sample size (ESS) is leveraged to Bayesian basket trials and several ESS-based borrowing strategies are proposed. The mean squared error (MSE), which explicitly accounts for the trade-off between estimation bias and variance reduction, is selected as the target measure for deriving ESS. Through a reanalysis of the RAGNAR study as well as simulation studies, the interpretability of ESS is demonstrated at both the analysis and design stages of basket trials.

Results: ESS reflects the impact of information borrowing on MSE and intuitively characterizes the degree of borrowing. It aligns with the type I error rate and power, showing potential as a valuable complement in statistical analyses and simulation studies.

Conclusions: Quantifying the degree of information borrowing by ESS can greatly help trialists design Bayesian basket trials, reasonably evaluate and interpret the results of Bayesian analyses, conduct sensitivity analyses, and ultimately borrow proper amount of information in basket trials.

Background: The panel `Health in Germany` has been established to gather nationwide health-related information, replacing cross-sectional surveys as primary data sources. However, panel designs involve multiple selection stages, potentially introducing additional nonresponse bias. This study aims to describe this drop-out bias and the weighting strategy used to improve representativeness.

Methods: Panelists were recruited through a recruitment study. At completion of the recruitment questionnaire, participants were invited to register for the panel. Registered panelists were subsequently invited to the first annual survey in 2024, which was divided into three sub-waves. Each included one of four questionnaires with different topics. Logistic regression models are used to estimate the probability for panel registration and participation in panel 2024 questionnaires, using sociodemographic and health-related variables from the recruitment study to illustrate drop-out bias. The weighting scheme and techniques are described. To assess the potential for drop-out bias and how it is reduced through weighting, unweighted and weighted estimates are compared to internal and external reference distributions using standardized differences.

Results: Drop-out analysis showed that sociodemographic characteristics (age, education, German citizenship) had a stronger association with panel registration than health-related parameters (e.g., self-rated health, smoking status, sport activity, chronic diseases) among the participants of the recruitment study. Similar patterns as for registration were found for participation in the 2024 questionnaires, except for age, which showed a reverse effect. No differences were observed across questionnaire types. Standardized differences confirmed the findings: sociodemographic characteristics-particularly education and German citizenship-showed larger deviations than health-related parameters. The largest deviations occurred in the recruitment study. The weighting procedure reduced most standardized differences to below 0.5% points. An exception is German citizenship, which showed only slight improvement.

Conclusions: Drop-out within the first year of a newly established panel is mainly affected by sociodemographic variables, with minor effects due to health-related parameters. The additional recruitment steps did not lead to concerning deviations in sample composition compared to the recruitment study. Remaining differences were addressed through drop-out and calibration weighting, so the weighted panel 2024 sample does not substantially differ from what would be expected in a cross-sectional design such as the recruitment study. However, continued analyses are needed, as sample composition may change due to future panel attrition.

Two-Sample Mendelian Randomization (TSMR) analysis is a widely used method for inferring causal effect in the presence of unmeasured confounding. However, causal inferences may be biased if the distributions of key variables (e.g., exposures, outcomes, and confounders) differ across populations. Such discrepancies in the distributions of key variables between the two populations are referred to as different local mechanisms. This paper aims to clarify the impact of different local mechanisms on the estimation of the Local Average Treatment Effect (LATE) in TSMR analyses using selection diagrams. We first uncover and formally define the Complete and Partial Inconsistent TSMR Estimations (InTSMRE). Subsequently, we propose a criterion of No InTSMRE in the context of continuous and binary outcomes. Following this, we introduce the LATE Ratio to evaluate the deviation of the LATE estimate from the true causal effect. Finally, we demonstrate that the violation of the Monotonicity condition exacerbates the occurrences of the Complete InTSMRE; otherwise only the Partial InTSMRE occurs. Additionally, through simulation studies, we illustrate the specific conditions under which these InTSMRE arise. We explore the LATEs of Waist-to-hip ratio on Type 2 diabetes in European and mixed populations, demonstrating the phenomenon of the InTSMRE.

Background: Chronic illness impacts not only individuals affected by it, but also those who care for them. Care partnerships recognize that health conditions are often shared, dyadic experiences. Qualitative dyadic analysis, which foregrounds the dyad as the unit of analysis, is a method that can enhance understanding of illness as a joint experience. However, when perspectives of dyad members are collected separately, their subsequent analysis as a unit can be challenging.

Objective: To review and summarize qualitative literature where data have been collected through separate individual interviews with patient and care partner dyads and analyzed at the dyadic level.

Methods: A scoping review guided by Joanna Briggs Institute methodology was undertaken. Databases (Ovid's Medline, Embase, and PsycINFO; EBSCO CINAHL; and ProQuest Sociological Abstracts) were searched in February 2024. Eligible articles included peer-reviewed literature published in English from 2010 onwards documenting qualitative dyadic analysis of individual interviews collected from patient and care partner dyads. Title and abstracts were screened and the full text of all potentially eligible articles was reviewed by two independent reviewers. Data were extracted using a table and results were summarized using frequency counts and qualitative content analysis.

Results: 7,494 records were identified and screened. 113 reports of 112 unique studies fulfilled eligibility criteria and were included. Numerous methodologies and analytic methods were reported, many of which incorporated methods from different qualitative traditions, often with variable sequencing of analytic steps that were infrequently well described. Studies were not routinely conceptualized at the dyadic level and underlying epistemological assumptions were rarely discussed despite their essential role in grounding dyadic analysis.

Conclusions: When conducting qualitative dyadic analysis, researchers should consider dyadic study conceptualization from study outset. The purpose of the analysis, the analytic steps taken, and their alignment with underlying epistemology and other incorporated methodologies should be clearly documented and reported.