[R] help with nomogram function

[Frank E Harrell Jr]

On Wed, 05 Nov 2003 09:22:34 +1000
Ross Darnell <r.darnell at uq.edu.au> wrote:

> I have fitted a logistic regression model
> 
> > failed.lr2$call
> lrm(formula = failed ~ Age + task2 + Age:task2, data = time.long, 
>     na.action = na.omit)

Use Age*task2 and omit na.action as na.omit is the default

> 
> using the Design package functions and would like to generate a
> nomogram from this model.
> 
> the datadist information is generated and stored in
> 
> > ddist
>                 time.long$Age time.long$task2
> Low:effect                 45            <NA>
> Adjust to                  56       both.foam
> High:effect                68            <NA>
> Low:prediction             21       both.foam
> High:prediction            80           right
> Low                        20       both.foam
> High                       80           right

This looks most strange.  You did not include the original code but I
suspect you had $ in a term.  $ should not appear in column headings
above.  Design wants you to use data= or attach, and avoid $ in terms.

Frank Harrell

> 
> Values:
> 
> time.long$task2 : both.foam left right 
> > 
> 
> The model fitted and then when I try the nomgram function
> 
> > nomogram(failed.lr2)
> Error in value.chk(at, i, NA, -nint, Limval, type.range = "full") : 
> 	variable Age does not have limits defined by datadist
> > 
> 
> I get an error. The NA values in ddist seem to be the problem but I
> don't understand the datadist information.
> 
> Can anyone help explain why this is failing.
> 
> Thanks
> 
> Ross Darnell
> 
> -- 
> 
> University of Queensland, Brisbane QLD 4067 AUSTRALIA
> Email: <r.darnell at uq.edu.au>
> Phone +61 7 3365 6087
> http://www.shrs.uq.edu/shrs/school_staff/ross_darnell.html
---
Frank E Harrell Jr    Professor and Chair            School of Medicine
                      Department of Biostatistics    Vanderbilt University