Glossary

Absolute difference

The absolute difference in rates of good or harmful outcomes between experimental groups (experimental group risk [EGR]) and control groups (control group risk [CGR]), calculated as the risk in the control group minus the risk in the experimental group (CGR – EGR). For instance, if the rate of adverse events is 20% in the control group and 10% in the treatment group, the absolute difference is 20% − 10% = 10%.

ACS NSQIP-P

See National Surgical Quality Improvement Program Pediatric (NSQIP-P).

American College of Surgeon Trauma Quality Improvement Program (ACS TQIP)

Implemented in 2010 to improve the quality of care for trauma patients. See also National Trauma Data Bank (NTDB).

Analysis of covariance (ANCOVA)

The linear model used for covariate adjusting. It assumes, for all possible values of covariates, that covariate effect size (ES) is equal to typical ES; that is, that there is no interaction between the covariates and the treatment effect.

Analysis of variance (ANOVA)

Statistical method used to compare a continuous dependent variable and more than 1 nominal independent variable. Often used for analyzing longitudinal data. ANOVA does not have the flexibility of mixed models of analysis and can yield misleading results if its more rigid assumptions (eg, all effects are considered fixed) are not met.

Area under the ROC curve (AUROC)

Technique used to measure the performance of a test plotted on a receiver operating characteristic (ROC) curve or to measure drug clearance in pharmacokinetic studies. When measuring test performance, the larger the AUC, the better the test performance. A model with perfect sensitivity and specificity would have an AUROC of 1. See also Receiver operating characteristic (ROC) curve.

Bayesian analysis

A statistical method that uses prior knowledge (eg, prior probability, conditional probability or likelihood) combined with data to obtain a new probability.

Bayesian hierarchical model (BHM)

A statistical procedure that integrates information across many levels, so multiple quantities are estimated simultaneously, and explicitly separates the observed variability into parts attributable to random differences and true differences.

Bias

Systematic deviation from the underlying truth because of a feature of the design or conduct of a research study (eg, overestimation of a treatment effect because of failure to randomize).

Bonferroni correction

A statistical adjustment to the threshold P value to adjust for multiple comparisons. The usual threshold for statistical significance (α) is 0.05. To perform a Bonferroni correction, one divides the critical P value by the number of comparisons being made. For example, if 10 hypotheses are being tested, the new critical P value would be α/10, usually 0.05/10 or 0.005. The Bonferroni correction represents a simple adjustment but is very conservative (ie, less likely than other methods to give a significant result).

C statistic

The C statistic is the probability that, given 2 individuals (one who experiences the outcome of interest and the other who does not or who experiences it ...