error after predictms - using the exact code by Prof Crowther (stata meeting slides)

[Denise Vella]

As you may be aware I am calculating transition probabilities using Prof Crowther slides using this website :

Ref: Slide 49 https://www.stata.com/meeting/italy1...8_Crowther.pdf

clear all
use http://fmwww.bc.edu/repec/bocode/m/multistate_example,clear
msset, id(pid) states(rfi osi) times(rf os) covariates(age)
mat tmat = r(transmatrix)

//Declare survival data
stset _stop, enter(_start) failure(_status=1) scale(12)

qui stpm2 age sz2 sz3 nodes hormon pr_1 if _trans1==1, scale(h) df(3) tvc(sz2 sz3 pr_1) dftvc(1)
estimates store m1
qui streg age sz2 sz3 nodes hormon pr_1 if _trans2==1, distribution(weibull)
estimates store m2
qui stpm2 age sz2 sz3 nodes hormon pr_1 if _trans3==1, scale(h) df(3) tvc(pr_1) dftvc(1) nolog
estimates store m3

predictms , transmat(tmat) at1(age 45) timevar(temptime) graph models(m1 m2 m3)

(Question 1) Why am I getting this error after -predictms-

predictms, transmat(tmat) at1(age 45)

ERROR: At least one prediction type must be specified

I don't understand, I've copied and pasted the code from Prof Crowther's slides' (slide 49-50)
I'm sorry I can't paste the screenshot, looks like there is an error on Statalist.

Appreciate any insight

[Denise Vella]

Michael Crowther appreciate any input.. there have been similar posts from other users. Does -predictms- still work?

[Bjarte Aagnes]

Does -predictms- still work?

https://github.com/RedDoorAnalytics/multistate

ssc install multistate
help predictms

[Bjarte Aagnes]

from help predictms *! version 4.4.1 17jan2023 MJC

statistic                      Description
    ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------
      probability                  calculate transition probabilities
      los                          calculate length of stay in each state
      rmst                         calculate restricted mean survival time
      visit                        probability of ever visiting each state
      hazard                       transition-specific hazard function
      survival                     transition-specific survival function
      userfunction(func_name)      user-defined Mata function for bespoke predictions; see details
    ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------
    At least one of the above options must be specified

(my emphasis above)

[Denise Vella]

thanks for the above. But I don't understand why MJC didn't specify any of this in his slides.... I've linked his 2018 and attached a screenshot of his 2017 slides

https://www.stata.com/meeting/italy1...8_Crowther.pdf

Using the help above:

clear all
use http://fmwww.bc.edu/repec/bocode/m/multistate_example,clear
msset, id(pid) states(rfi osi) times(rf os) covariates(age)

//Stores matrix callted tmat name, stores the values from transmatrix
mat tmat = r(transmatrix)

//Displays tmat ie the values of transmatrix
mat list tmat

//Declare survival data
stset _stop, enter(_start) failure(_status=1) scale(12)
streg age_trans1 age_trans2 age_trans3 _trans2 _trans3, dist(weibull)
range temptime 0 15 100
predictms, transmat(tmat) at1(age 45) timevar(temptime)

ERROR: At least one prediction type must be specified
r(198);


//Following statalist help:
predictms, transmat(tmat) probability at1(age 45) timevar(temptime)

ERROR: Only merlin models are supported

What is this merlin model? I've haven't found anyone with a similar problems. iT WOULD be great if these 2018/2017 slides are updated.

[Denise Vella]

Of note, I also check the latest predictms help section


and modified my code as recommended by the help

clear all
use http://fmwww.bc.edu/repec/bocode/m/multistate_example,clear
msset, id(pid) states(rfi osi) times(rf os) covariates(age)

//Stores matrix callted tmat name, stores the values from transmatrix
mat tmat = r(transmatrix)

//Displays tmat ie the values of transmatrix
mat list tmat

//Declare survival data
stset _stop, enter(_start) failure(_status=1) scale(12)

stmerlin age if _trans1==1, distribution(weibull)
estimate store m1

stmerlin age if _trans2==1, distribution(weibull)
estimate store m2

stmerlin age if _trans3==1, distribution(weibull)
estimate store m3

///transition probabilities for a patient with age 50  
predictms, transmatrix(tmat) models(m1 m2 m3) probability at1(age 50)

However I obtain the following error:

Any thoughts? THe 2017/2018 slides looked so simple, yet there's been an. update somewhere in the process