What is EPICURE?
Risk Sciences International (RSI) is excited to announce the release of Epicure 2.0 for Windows, the first official release of the window-based version of the premiere software for risk regression and person-year tabulation. The following sections provide an overview of the history and functionality of Epicure and highlight some of the new features.
See below for information on pricing.
EPICURE: what to expect
RSI, acquired Epicure from Hirosoft International Corporation, and is marketing Epicure while actively working with Dale Preston and his colleagues to improve and enhance Epicure’s capabilities.
Brief History and Description of Epicure
Epicure was originally developed for regression modeling of radiation effects on cancer rates in atomic bomb survivors. The development was motivated by the recognition that models focused on the excess relative risk (i.e. the RR-1) were more suitable for describing dose response and effect modification than the loglinear Cox-regression proportional hazards model and by the need for explicitly modeling excess rates (rate differences) as a function of dose and other, often time-dependent, factors.
For almost 30 years, Epicure has provided a powerful set of tools for risk regression using a rich class of models that extends the commonly used log-linear relative risk (Cox regression) and relative odds (logistic regression) models to include excess relative risk / odds models and excess rate models. Epicure also includes a module for the straightforward specification and creation of highly stratified person-year (or more generally event-time) tables including, as needed, stratification on multiple time scales and time-dependent factors (such as lagged cumulative doses).
Epicure is an interactive, command-driven program with a simple and intuitive, b ut powerful scripting language. This new release features a graphical user interface that makes the program even easier to use.
While Epicure is the de-facto standard for modeling radiation health effects, the models and methods in Epicure have been used for a wide variety of medical, public health, epidemiological, economic, environmental, and reliability data. The methods are particularly useful for dose-response modeling and investigating joint effects of and interactions between multiple risk factors.
Epicure’s DATAB module is the most powerful and flexible tool available for creating high dimensional person year (event-time) tables for use in Poisson-regression-based analyses of survival data including analysis of standardized mortality/incidence ratios.
RSI's EPICURE offer in greater detail
Epicure Risk Model Examples
The following examples describe some of the models that can be used for modeling risks and rates in Epicure. Additional details are available in the online manuals (http://epicurehelp.risksciences.com).
Classical relative risk / relative odds models:
baseline – non-parametric, stratified, fully parametric
RR – typically log-linear ()
- Includes the semi-parametric standard proportional hazards model, Cox regression ()
- Fully stratified or parametric baseline rate/risk models can also be used
- General time-dependent covariates
Excess relative risk / excess odds ratio:
ERR – excess relative risk
- Models include simple dose response models: e.g., or
- Easily fit categorical, threshold, and spline dose-response models
- Multiplicative effect modification: e.g. where x could include fixed (e.g., exposure age, sex) or time-dependent (e.g. age, time since exposure) factors
Joint effects for excess relative risk / excess odds ratio:
additive excess relative risk
multiplicative excess relative risk model
- Examples include joint effects of chemical exposure (ERR1) and smoking (ERR2)
- Up to four excess relative risk terms
- Can model joint effects as geometric mixture of additive and multiplicative models
Risk difference / Excess rate / excess odds models†:
- Up to four excess absolute rate terms
- Excess absolute rate terms have same form as ERR terms
- † risk difference modeling requires parametric baseline model
Epicure Risk Regression Modules
GMBO / PECAN – Binomial data including conditional and unconditional logistic regression
- Matched (conditional) and unmatched logistic regression for odds ratios in case-control studies
- Stratified logistic regression
- Regression analyses of probabilities and log(probabilities)
- User-defined variance functions for over-dispersed data
PEANUTS –Partial-likelihood methods for censored survival data (including Cox regression)
- Easily define and use general user-defined time-dependent covariates
- Case-cohort data
- Allows late entry data
- Kaplan-Meier and Nelson-Aalen non-parametric estimates
AMFIT – Grouped survival / (Poisson (piecewise constant hazard) data
- Relative risk and rate difference models
- Efficient handling of time-dependent covariates
- Either fully or semi-parametric hazard functions in relative risk models
- Standardized mortality/incidence ratio analyses
Person-year (rate) table creation (DATAB Module)
- Easily create multi-dimensional event-time (person-year / rate) tables
- Multiple category variables
- Simple specification of categories
- Multiple time scales (e.g. attained age, time since exposure/diagnosis, calendar time)
- User-defined time category boundaries
- Calendar time or user defined time scales
- Time dependent categories (e.g. lagged cumulative dose, years smoked, time-since-smoking cessation)
- Multiple summary variables (case counts, cell-specific means etc.)
- Easy to add external rates or job-exposure-matrix values
Powerful tools for dealing with time-dependent exposures and risks
Easy to use standard models or to extend to more general risk models
- Excess relative risk models
- Excess rate models
- Additive and multiplicative models for joint effects
- Simple model specification
- Asymptotic-normality-based (Wald) test and confidence bounds
- Score tests
- Likelihood ratio tests
- Profile likelihoods and likelihood-based bounds
- Efficient procedures for fitting stratified regression models
- Create multi-dimensional person-year tables with user-defined time-dependent categories
- Simple, but useful tools for summarizing and plotting data
- Create new variables as needed
- Designed for interactive modeling
- Easily select subsets
- Simple creation of categorical variables from continuous variables
- Command driven scripting with a graphical user interface
- Windows-based graphical user interface
- Multiplicative ERR models
- Linear parameter constraints in regression models
- Improved case-cohort modeling (PEANUTS)
- Simplified input from delimited (comma, tab, space) text files with variable name headers
- Array functions to simplify computation of time-dependent events such as cumulative doses
- External dataset lookup available as a function (with up to six levels of indexing) in both DATAB and the regression modules.
- Enhanced data export routines
- Improved output of model form and parameter estimate data (easily readable in commonly used statistical packages)
- Save ungrouped person year data (one record for each cell in which a “person” is at risk) to text file to facilitate minimally grouped Poisson regression analyses of survival data. Useful for data with many risk factors or for post-hoc analyses of covariate (e.g. dose) uncertainties.
- Improved automatic session logging
- Context sensitive help
A single user license is 1200 USD but licensed users of the previous edition of Epicure can upgrade for 750 USD.
Optional annual maintenance is 240 USD. Our discount program is outlined below:
- Full price for the first license
- 10% off the second
- 15% for the third
- 20% for the fourth
- 25% off any additional licenses
Contact email@example.com if you need more information about Epicure models and features.
If you wish to purchase Epicure 2.0, please provide us with your billing details and indicate how many licenses you would like to purchase.
Selected examples of analyses using EPICURE:
- Grant et al. Risk of death among children of atomic bomb survivors after 62 years of follow-up: a cohort study. Lancet Oncol. 2015 Oct;16(13):1316-23. http://www.ncbi.nlm.nih.gov/pubmed/26384241
- Gudzenko et al. Non-radiation risk factors for leukemia: A case-control study among chornobyl cleanup workers in Ukraine. Environ Res. 2015 Oct;142:72-6. http://www.ncbi.nlm.nih.gov/pubmed/26117815
- Lee et al. Occupational ionising radiation and risk of basal cell carcinoma in US radiologic technologists (1983-2005). Occup Environ Med. 2015 Dec;72(12):862-9. http://www.ncbi.nlm.nih.gov/pubmed/26350677
- Sokolnikov M, Preston D, Gilbert E, Schonfeld S, Koshurnikova N. Radiation effects on mortality from solid cancers other than lung, liver, and bone cancer in the Mayak worker cohort: 1948-2008. PLoS One. 2015;10(2):e0117784. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4342229/
- Zablotska et al. Leukemia, lymphoma and multiple myeloma mortality (1950-1999) and incidence (1969-1999) in the Eldorado uranium workers cohort. Environ Res. 2014 Apr;130:43-50. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4002578/
- Furukawa et al. Radiation and smoking effects on lung cancer incidence among atomic bomb survivors. Radiat Res. 2010 Jul;174(1):72-82. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3857029/
- Lubin and Caporaso. Cigarette smoking and lung cancer: modeling total exposure and intensity. Cancer Epidemiol Biomarkers Prev. 2006 Mar;15(3):517-23. http://cebp.aacrjournals.org/content/15/3/517.long