Egret by Cytel | Statistical software | Timberlake Consultants

+44 (0) 20 8697 3377

Home

About Us

Charities

info@timberlake.co.uk

Select Region:

Egret: Statistical software for Epidemiologists

Intuitive tools for design and analysis of biomedical and epidemiological studies.

Developer:
Operating systems:
Type:

Cytel Inc.
Windows 7, Vista & XP
Statistical & Clinical Trial

Overview

Egret is the friendliest statistical package designed specifically for epidemiologists and provides a comprehensive set of regression and other tools for analyzing case-control studies. Egret also offers graphical and tabular diagnostic tools that let you inspect the fitted model and assess it, as well as “delta-beta” plots that tell you how much a regression parameter would change if a given case were omitted from the analysis.

Analysis Models

Egret’s regression models include:

Logistic regression, fixed effects (both conditional and unconditional)
Logistic regression, random effects
Poisson regression
Cox proportional hazards regression
Exponential regression
Weibull regression (with separate sets of terms for the scale and shape parameters)

For random effects logistic regression, Egret provides betabinomial, logistic binomial, logistic binomial for distinguishable data, and logistic normal models, and lets you specify risk as additive, relative, or additive-relative. For all logistic and Poisson regression models, Egret offers two automated model selection procedures: stepwise (forward) selection, and backward elimination. Fitting algorithm option include Newton-Raphson, modified Newton-Raphson, quasi-Newton or Nelder-Mead simplex.

For survival data, you can define time dependent covariates using interpolation and step functions and staggered entry times.The output includes deviances, p-value, standard errors, likelihood ratio tests, odds ratios, and conditional odds ratios.

Statistical tests in Egret include KaplanMeier procedures for censored survival data, a stratified and unstratified log-rank test, as well as both exact and asymptotic analysis for contingency table data.

Egret’s Kaplan-Meier model allows for arbitrary grouping for multiple estimates, and gives you both graphical and listtype output. Its plots include KM, -log (KM), log 1 -log (KM), 1-KM, and kernelsmoothed hazards.

For 2xK contingency table analysis Egret provides both asymptotic and exact procedures for various odds ratio and trend tests (including Mantel-Haenszel), as well as asymptotic tests for their stratified counterparts. Up to 100 levels of the exposure variable are permitted.

CaseEditor

The Egret CaseEditor is a spreadsheetlike facility for importing, creating or editing case data files. You can import LogXact, StatXact, ASCII, Excel, Excel, CSV, SPSS, SYSTAT and SAS Transport files. You can have up to 120,000 records and 250 variables. Spreadsheet facilities let you create derived variables, create new categorical variables from existing categorical and continuous variables, add weights to regression variables, and create subsets of the full data set.

Post-fit Diagnostics and Graphics

For analysis of the fitted model (logistic, conditional logistic, random- effects logistic regression, Cox Proportional Hazards, and Poisson regression), Egret offers both fitted values and delta-betas plotted and in spreadsheet format. You can edit the plots on screen, and plot subsets of data based on delta-betas or fitted values. Observations with extreme delta-betas can be specially flagged. The plots available include: