NONMEM®

ICON's nonlinear mixed effects modelling tool used in population pharmacokinetic/ pharmacodynamic analysis.

NONMEM®


NONMEM® is a nonlinear mixed effects modelling tool used in population pharmacokinetic/pharmacodynamic analysis

The Software:

NONMEM® versions up through VI are the property of the Regents of the University of California, but ICON Development Solutions has exclusive rights to license their use. NONMEM® 7 up to the current version 7.2.0 is the property of ICON Development Solutions.

The software consists of three parts.

  • The NONMEM® program itself is a very general (non-interactive) model analysis program that can be used to fit models to many different types of data. With version 7, Monte Carlo expectation-maximization and Markov Chain Monte Carlo Bayesian methods have been added to the classical likelihood methods available in previous versions. NONMEM® can be used to simulate data as well as fit data. 
  • PREDPP is a powerful package of subroutines that can be used by NONMEM® to compute predictions for population pharmacokinetic and pharmacodynamic data. Use of PREDPP obviates the need for the user to code kinetic-type equations and it also allows complicated patient-type data to be easily used. However, the user can also directly and very generally code prediction-type, likelihood or -2 log-likelihood equations, and thus a great variety of different types of models can be used. 
  • NM-TRAN is (non-interactive) preprocessor, allowing control and other needed inputs to NONMEM/PREDPP to be specified in a user friendly manner, with quick and comprehensive detection of a variety of errors that the user may have made in so doing.

NONMEM® 7 is programmed in Fortran 90/95 code and can be used with 32 or 64 bit Windows, Linux, Macintosh OSX, or Solaris operating systems incorporating a Fortran 90/95-compliant compiler. Since some NONMEM® runs can take considerable time, perhaps many hours or even days, depending on the speed of the computer and the size of the problem, it is advisable to use a fast machine with at least 1.0 Gb of available memory. For multiprocessor and multi-core environments 3-4 Gb of memory may be needed to accommodate several simultaneous NONMEM® runs.

NONMEM® 7.3 includes the following enhancements:

  • The following population analysis methods are available:
    • First Order Conditional Estimation (FOCE)
    • Laplace Conditional Estimation
    • Iterative Two Stage (ITS)
    • Importance Sampling Expectation-Maximization (IMP)
    • Stochastic Approximation Expectation-Maximization (SAEM)
    • Markov-Chain Monte Carlo Bayesian Analysis (BAYES)
  • Parallel computing of a single problem over multiple cores or computers, significantly reducing completion time.   
  • Dynamic memory allocation according to problem size, eliminating the need to recompile the NONMEM® program for unusually large problems. User may override program generated suggested values using a statement in the control stream.  
  • Improved incidence of success in problems using the first-order conditional estimation method.  
  • Improved incidence of completion when using the “Super Problem” feature.  
  • Additional result files, with number of significant digits selectable by the user, and which can be easily read by post-processing programs.  
  • Multiple mixed effects levels, with random effects across groups of individuals such as clinical site, may now be modelled. Sites themselves may be additionally grouped, such as by country, etc. 
  • Control stream files may be written in mixed case, any number of data items per data,  record label names may be as large as 20 characters.
  • Subscripted variables may be used in abbreviated code for use in DO loops.
  • Fewer restrictions on DOWHILE, and correlation of residual variables using CORRL2 may be written in abbreviated code. 
  • Easy to code inter-occasion variability.  ETA’s can be referenced by an index variable related to the inter-occasion data item. 
  • Symbolic reference to thetas, etas, and epsilons. 
  • Priors for SIGMA matrix may be added.
  • Variance matrix parameters output in covariance and correlation format.
  • Variance matrix parameters may be input in covariance, correlation, or Cholesky format.
  • Enhanced annealing feature for SAEM to facilitate estimation of fixed effect parameters (THETAs) that do not have associated inter-subject variance (ETA’s).
  • Boot-strap simulations may be performed in NONMEM. 
  • Initial ETAs may be introduced in the control stream file or from an external source.  
  • Certain options may be optimized for Monte Carlo (MC) and Expectation-Maximization (EM) methods.    For SAEM and IMP, NONMEM® assesses best ISAMPLE values for each subject. For IMP, NONMEM® assesses best IACCEPT and DF values for each subject. For BAYES and SAEM, NONMEM® assesses best CINTERVAL based on the degree of Markov chain correlation across iterations. 
  • AUTO option determines best set of options for MC/EM methods. 
  • Perform Monte Carlo search of population parameters, and select the set that provides the lowest starting objective function for an estimation. 
  • Perform Monte Carlo search for initial best estimates of etas for each subject. Together with a Monte Carlo search of best initial population parameters, this provides a global search technique for the traditional methods, with less reliance on starting position for incidence of success. 
  • FOCE/Laplace and ITS may be assessed using only numerical eta derivatives for search of best etas and/or eta Hessian matrix assessment.   
  • Conditional Individual Weighted Residual (CIWRES) added to residual variance diagnostics.   Includes evaluation for L2 correlated data as well.
  • An option is provided to obtain near identical results for repeated runs of Monte Carlo EM problems regardless of whether job is single CPU processed or parallel processed
  • NMTRAN will allow & as a continuation marker on abbreviated code lines.  Furthermore, the total length of a control stream record, whether on a single line or continued on several lines using &, may be up to 67000 characters long.  
  • Thetas may be inputted and reported in their natural domain, even when linear Mu referencing.
  • Informative record names for prior information of THETAs/OMEGAs/SIGMAs provide easier entry of prior information.
  • The XML version of the report file has been enhanced with additional elements.
  • Example control stream files modeling population densities of individual parameters that are t-distributed. 
  • Nelder-Mead optimization option for best fit individual etas, particularly useful to improve robustness for importance sampling. 
  • Option to use either eigenvalue square root or Cholesky square root algorithms for assessing weighted residual diagnostics. 
  • Select which ETAs are to be included in the average eta shrinkage calculation. A general option to exclude non-influential etas may be selected, and/or etas to specific subjects may be excluded by setting a reserved variable in $PK or $PRED.
  • Enhanced non-parametric analysis methods, such as extended grid of support points, use of an outsize inter-subject variance to obtain support points that fit outlier subjects better, and built-in bootstrap analysis methods for obtaining empirical confidence ranges to non-parametric probability parameters 
  • A utility program to fill in extra records with small time increments, to provide smooth plots.  
  • A utility program to fill in substitution variables in template control stream files.
  • Analytical and numerical derivatives of predicted and residual variance values with respect to eta may be outputted.  
  • Covariance assessment may be turned off for a particular estimation
  • A range of ETAs may be requested to be outputted to tables. For example, ETAS(3:LAST) in the $TABLE record may be requested to output etas 3 through last one.
  • Features to facilitate stochastic differential equations (SDE).

The NONMEM program is available on CD ROM, which together with the documentation and all updates and additions to the program and documentation, will be delivered for a license royalty fee to be paid annually. This fee is subject to change from year to year, and at each anniversary the licensee at its option may choose not to renew the license.