// Updates to standsurv
0.1 - 2019-1-20
test version released on webpage 20/1/2019
0.2 - 2019-03-04
0.34 - 2019-07-03
0.36 - 2019-11-07
0.43 - 2019/02/29
-- fixed bug with no at option - now behaves as at(.)
-- added derivative of hazard functions for streg models
-- now work with ode calculations for competing risks.
-- error given if a cure models
-- fixed some typos in help file.
0.44 - now compiled in 15.1
0.45 - 2020-06-09
- fixed problem with more than 2 competing risks models
- added error of missing indweights
0.46 - 2020-06-11
- genind works for competing risks models
- fixed cif bug (from 0.45)
0.47 - 2020-11-28
- toffset() option now works again
- contrast(ratio) returned log of ratio without ci option. Now fixed.
- trans(loglog) gave lci and uci the wrong way round. Now fixed.
0.48 - 2020-12-08
- cif with rmft now works for competing risks models to estimate loss in life expectancy.
0.50 - 2021-02-03
- Fixed bug introduced in last release that ode methods for crude probabilities and CIFS gave
incorrect estimates if not sorted by timevar.
0.51 - crude probabilities use ode method by default
- can now use file from web location in expsurv(using()) option.
0.52 - help file edit + genind() now works when calculating cifs in competing risks models
0.53 - fixed bug when calculation ci or se for rmst for exponential models
0.55 - for crude probabilities of death - contrast was not returned without ci option - fixed.
0.56 - 2021-03-21
-- popmort filename can now have spaces for windows compatability.
-- genind option now works for most predictions.
-- at() suboption of expsurv now works with multiple variables.
0.57 - 2021-03-22
-- genind() option for competing risks and crude probabilities now allows if statements.
0.58 - 2021-04-09
-- expsurv option when combined with ode did not use expected rates
0.59 - 2021-05-10
-- introduced a bug in previous version for CIs when doing contrasts - now fixed.
0.60 - 2021-05-12
-- too aggressive error checks for attimevar suboption removed.
0.61 - 2021-08-23
-- using -cif rmft- gave an unhelpful Mata error without the ci option - now fixed.
-- ode evaluator now slightly more accurate. Works better with genind() option.
0.62 - 2021-09-03
-- using -cif rmft- combined with a contrast gave an unhelpful Mata error without the ci option - now fixed.
0.63 - 2021-12-19
-- stpm2 models using reverse option gave incorrect estimates - now fixed
-- new suboption genind as part of expsurv option.
-- more meaningful errors for lincom() option if specified incorrectly
0.64 - 2022-01-20
-- reverse option now working with strcs models
0.65 - 2022-02-06
-- previous update (0.63) broke genind option for standard models
0.66 - 2022-04-06
-- fixed error when combining with rmst, with expsurv and ode
0.67 - 2022-05-07
-- add storeg option to save derivatives of standardized estimates.
0.69 - 2022-05-25
-- add storeg option for competing risks / crude probabilities