An official website of the United States government.

Due to a lapse in appropriations, EPA websites will not be regularly updated. In the event of an environmental emergency imminently threatening the safety of human life or where necessary to protect certain property, the EPA website will be updated with appropriate information. Please note that all information on the EPA website may not be up to date, and transactions and inquiries submitted to the EPA website may not be processed or responded to until appropriations are enacted.

We've made some changes to If the information you are looking for is not here, you may be able to find it on the EPA Web Archive or the January 19, 2017 Web Snapshot.

Benchmark Dose Tools

Benchmark Dose Software (BMDS) v3.0 Support Articles

On this page

Reinstalling Excel May Solve Random Performance Issues

Some users have reported unusual error messages (such as “File not found: cmodels.dll”) or problems running analyses despite having a stable Microsoft Excel installation. In most cases, re-installing Excel solves these random problems.

For BMDS 3.0, we recommend a desktop version of Microsoft Excel 2016 for Windows or later with macros enabled (visit the Microsoft support site for information on enabling Excel macros). We recommend using the Office 365 or standalone Microsoft Office installation methods rather than installing via the Microsoft Store.

Note: Running BMDS 3.0 with the 64-bit version of Office will also generate error messages such as “File not found: cmodels.dll.” Reinstalling 64-bit Office will not solve that issue. An upcoming release will resolve most or all of these compatibility issues.

Top of Page

Planned Features

The following features will be rolled out in future BMDS releases:
  • Analysis of two-dose datasets
  • Report parameter standard error (SE) values
  • Import BMDS 2.7 datasets
  • Implement NCTR nested dichotomous model
  • Recompile models for improved speed (particularly for “Added Risk”)
  • Add Asymptotic Correlation Matrix of Parameter Estimates table to dichotomous output
  • Implement improvements to Results Workbook and Word Report feature
  • Implement confidence intervals for plotted group responses
  • Implement Cochran-Armitage trend test for dichotomous datasets
  • Implement data transformations on a column of data (for example, multiplying by a constant, returning the Log() base 10 of all values in a column, or converting standard errors to standard deviations, among other types of transformations)
  • Auto-update feature
  • Bayesian models for continuous response data
  • Bayesian model averaging for continuous response data
  • Sensitivity analysis for Bayesian priors (posterior distributions for alternative priors)
  • P-value (or equivalent) for Bayesian models
  • Covariate capability for all models
  • Risk-at-a-dose capability for all models
  • Model Averaging plots

Top of Page

Specifying a Background Dose of Zero

This option is currently supported only for the frequentist dichotomous models.

If you specify the background dose to be zero , the resulting maximum-likelihood estimation (MLE) background parameters can sometimes still vary slightly above zero.

Top of Page