[R] How to assign fixed beta coefficients in lrm for external validation

[David Winsemius]

On Jun 23, 2009, at 6:28 AM, jeweetwel wrote:

>
> Hi,
> I am planning to externally validate a logistic prediction model in  
> a new
> cohort. Outcome is mortality. The betacoefficients were derived from a
> previous published article. It seems not possible in R to assign  
> fixed beta
> coefficients to predictors like lrm (death ~
> intercept+beta1*var1+beta2*var2...). How do i solve this problem?

There is an offset model component for features that are assumed fixed.

-- 

David Winsemius, MD
Heritage Laboratories
West Hartford, CT