Risk Regression Package
Risk Regression Package - Related Videos
04:57
Establishing a Competing Risk Regression Nomogram Model for Survival Data
0 Views •
Cited by 4 •
2020
Presented here is a protocol to build nomograms based on the Cox proportional hazards regression model and competing risk regression model. The competing method is a more rational method to apply when competing events are present in the survival analysis.
01:52
Regression Toward the Mean
0 Views •
2020
Regression toward the mean (“RTM”) is a phenomenon in which extremely high or low values—for example, and individual’s blood pressure at a particular moment—appear closer to a group’s average upon remeasuring. Although this statistical peculiarity is the result of random error and chance, it has been problematic across various medical, scientific, financial and psychological applications. In particular, RTM, if not taken into account, can interfere when researchers try to extrapolate results...
05:37
An R-Based Landscape Validation of a Competing Risk Model
0 Views •
Cited by 2 •
2022
The present protocol describes codes in R for evaluating the discrimination and calibration abilities of a competing risk model, as well as codes for the internal and external validation of it.
07:32
Freezing, Thawing, and Packaging Cells for Transport
0 Views •
Cited by 2 •
2008
Cultured mammalian cells are used extensively in cell biology studies. This video describes the basic skills required to freeze and store cells and how to recover frozen stocks.
01:32
Chromatin Packaging
0 Views •
2023
Each human somatic cell contains 6 billion base pairs of DNA. Each base pair is 0.34 nm long, meaning each diploid cell contains a staggering 2 meters of DNA. This long DNA strand is packed inside a nucleus measuring only 10-20 microns in diameter with the help of specialized DNA-binding proteins called histones. Together they form a compact DNA-protein complex called chromatin. The chromatin is further compacted into higher-order structures. The highest level of compaction is achieved during...
FAQs
In a competing-risks analysis, an event of one type can occur before the event being evaluated, so the possible outcomes cannot be treated independently by ignoring the other event. The package accounts for competing events while estimating absolute risk at specified times. This produces risk quantities that better represent the probability of each outcome in the observed time-to-event setting.
A time-dependent receiver operating characteristic, or ROC, analysis examines how well predicted risks separate individuals who experience an outcome from those who do not as follow-up progresses. Unlike an absolute risk estimate, it focuses on discrimination, meaning the model’s ability to rank people by outcome likelihood. This helps compare prognostic models when predictive separation may change across the time horizon.
Prediction error and the Brier score evaluate how closely predicted risks correspond to observed outcomes over time, complementing discrimination measures such as time-dependent ROC curves. A model can distinguish higher- from lower-risk individuals while showing weaker overall predictive performance. Considering these measures together gives model comparisons a broader basis than relying on discrimination alone.
Accounting for censoring is important because time-to-event datasets may not contain complete event information for every participant. The package incorporates censoring when assessing prediction error, Brier scores, and related performance measures, rather than treating incomplete follow-up as a straightforward observed outcome. This supports a more appropriate evaluation of models using survival data with incomplete observation periods.
An analysis can produce estimated absolute event risks at selected follow-up times together with measures of discrimination and prediction error. Examining these outputs side by side lets researchers assess both the magnitude of predicted risk and the model’s performance in separating outcomes. The results can support comparison of candidate prognostic models and selection of one for biomedical or public health use.
Researchers can use the package after developing prognostic models to evaluate how well they perform on time-to-event data. The workflow centers on applying the models, estimating risks over time, and examining time-dependent ROC, prediction-error, and Brier-score results while accounting for censoring and competing events. This creates a structured basis for validation and evidence-based model selection.
In clinical and public health research, predicted risk may be needed for an individual patient or summarized for a population. The package supports both perspectives by quantifying event risk over time and evaluating predictive performance in survival and competing-risks settings. These results can inform prognostic assessment, population risk characterization, and decisions about which model provides the most useful evidence.