Modelling (or fitting) data sets

The fitting or modelling of single or repeat dose pharmacokinetic concentration-time data can be performed easily using PCModfit. The program utilises three very different and powerful algorithms (all very fast) for generating the lines of 'best fit' through data with extensive predictive output to text files and and in Excel, with high quality graphics. Weighted and Iteratively Reweighted Least Squares approaches are used (with or without constraints) with inbuilt calculation of parameter errors and other such essential information required by any professional PK package. The graphical output is of high quality and these stored pictures can be imported into Microsoft Word etc. Data can be entered into Excel spreadsheets in columns and transferred to PCModfit in seconds. Summary results and pictures are stored in the PCModfit 'Results' directory for reports etc.

There are numerous models built into the program including for example, i.v. bolus and infusion (up to 3-compartments), oral (first and zero order), polynomials and many others to wet your appetite! A few examples of pictures generated from fitting data are shown for interest.

From V7.0 onwards, the modelling option has been further updated. The setting up and the graphics (fitted line and data) are now all displayed in the 'Modelling' Excel spreadsheet. For setting up the 'Control' parameters, which was often a bit tricky, there is now a 'Keywords' button which helps with the required layout based on the Fitting Options selected in the spreadsheet (models, algorithm, weighting etc.). The graphics are of high quality (both linear and logarithmic plots are now on separate Charts within Excel) and the number of points for the computed line can now be selected to ensure a representative line over extended time periods - useful for repeat dose fitting where the overall time can be quite long (up to 10000 points maximum). There is additional help describing the models and parameters which are also shown in the spreadsheet (drop down boxes containing model numbers and what the models actually are and if user parameter starting estimates are required). The graphics files produced within Excel®, are now stored as *.png and not *.wmf files to improve the whole appearance (Fitplotlin*.png for Linear and Fitplotlog*.png for Logarithmic Charts). Users can now analyse up to 1000 data points per profile and up to 100 profiles in a single run (should the user be so lucky!). When modelling is completed, all of the graphs can be viewed within the spreadsheet to aid the user in deciding if it was acceptable or otherwise.


Fitting single dose infusion data with a 3-compartment model



Fitting repeat dose data with a 2-compartment oral model (differing doses and intervals)