Computational protocol: The PiGeOn project: protocol of a longitudinal study examining psychosocial and ethical issues and outcomes in germline genomic sequencing for cancer

Similar protocols

Protocol publication

[…] Mean differences in outcomes will be compared using t-test (continuous) or chi-squared tests (dichotomous). Non-parametric tests will also be used where appropriate. Temporal changes in scales will be investigated by calculating the difference between time-points. Multiple (continuous outcomes) or logistic (dichotomous outcomes) regression will be employed to adjust for the effect of confounders and identify predictors of outcome. The correlation structure between blood relatives or when outcomes are analysed over more than 2 time points will be suitably adjusted for when needed using mixed effects models. Linear mixed models and logistic mixed models will be performed in R: A language and environment for statistical computing employing package nlme for the linear mixed model and lme4 for the logistic mixed model. Assumptions of normality of residuals and homogeneity of variance will be confirmed visually though diagnostic residual plots. Multivariable models will be created with the inclusion of all potential confounders, and those predictors that show little evidence for an association with the outcome in univariate analysis. Backwards elimination followed by forwards addition will be employed to select predicting variables in the final model. Known and identified confounders will be incorporated regardless of their statistical significance. Collinear independent variables will be identified and eliminated. […]

Pipeline specifications

Software tools nlme, lme4
Application Mathematical modeling
Organisms Homo sapiens
Diseases Neoplasms