[R] multistate data w/out individual ID
Emilio A. Laca
ealaca at ucdavis.edu
Wed Nov 6 22:06:25 CET 2013
Dear R help community,
I have a version of the data below generated by counting the number of individuals in each state every 2-3 time units for a series of cohorts (sets).
States are A, B, C, E and T.
It is known that the following transitions are possible A->B, B->C, C->E, B->T, C->T and E->T.
The identity of the individuals in each set was not recorded.
I am interested in estimating the transition rates or probabilities and sojourn times as a function of covariates.
We assume that rates (probabilities) are independent of time.
Is there a package to estimate those quantities?
Is there a method to estimate those quantities?
Do you think a Bayesian approach integrating over all possible paths might work?
Everything I read about multi-state modeling packages seems to indicate that it is necessary to know the identity of individuals.
Any comment will be appreciated.
Emilio A. Laca, Professor
One Shields Avenue, 2306 PES Bldg.
Plant Sciences Mail Stop 1 ealaca at ucdavis.edu
University of California fax: (530) 752-4361
Davis, California 95616 voice: (530) 754-4083
mobile: (530) 220-5315
Include the Mail stop in the address or mail will be delayed.
More information about the R-help