Fit multiple parametric survival regressions to the multiple required endpoints
Source:R/fitting.R
fit_ends_mods_par.RdFits multiple parametric survival regressions, according to the distributions stipulated, to the multiple endpoints required in fitting partitioned survival analysis, clock forward and clock reset semi-markov models.
Usage
fit_ends_mods_par(
simdat,
cuttime = 0,
ppd.dist = c("exp", "weibullPH", "llogis", "lnorm", "gamma", "gompertz"),
ttp.dist = c("exp", "weibullPH", "llogis", "lnorm", "gamma", "gompertz"),
pfs.dist = c("exp", "weibullPH", "llogis", "lnorm", "gamma", "gompertz"),
os.dist = c("exp", "weibullPH", "llogis", "lnorm", "gamma", "gompertz"),
pps_cf.dist = c("exp", "weibullPH", "llogis", "lnorm", "gamma", "gompertz"),
pps_cr.dist = c("exp", "weibullPH", "llogis", "lnorm", "gamma", "gompertz"),
expvar = NA
)Arguments
- simdat
Dataset of patient level data. Must be a tibble with columns named:
ptid: patient identifier
pfs.durn: duration of PFS from baseline
pfs.flag: event flag for PFS (=1 if progression or death occurred, 0 for censoring)
os.durn: duration of OS from baseline
os.flag: event flag for OS (=1 if death occurred, 0 for censoring)
ttp.durn: duration of TTP from baseline (usually should be equal to pfs.durn)
ttp.flag: event flag for TTP (=1 if progression occurred, 0 for censoring).
Survival data for all other endpoints (time to progression, pre-progression death, post-progression survival) are derived from PFS and OS.
- cuttime
Cut-off time for a two-piece model, equals zero for one-piece models.
- ppd.dist
Vector of distributions (named per
flexsurv::flexsurvreg()) to be fitted to Pre-Progression Death (PPD).- ttp.dist
Vector of distributions (named per
flexsurv::flexsurvreg()) to be fitted to Time To Progression (TTP).- pfs.dist
Vector of distributions (named per
flexsurv::flexsurvreg()) to be fitted to Progression-Free Survival (PFS).- os.dist
Vector of distributions (named per
flexsurv::flexsurvreg()) to be fitted to Overall Survival (OS).- pps_cf.dist
Vector of distributions (named per
flexsurv::flexsurvreg()) to be fitted to Post Progression Survival, where time is from baseline (clock forward).- pps_cr.dist
Vector of distributions (named per
flexsurv::flexsurvreg()) to be fitted to Post Progression Survival, where time is from progression (clock reset).- expvar
Explanatory variable for modeling of PPS
Value
A list by endpoint, then distribution, each containing two components:
result: A list of class flexsurvreg containing information about the fitted model.
error: Any error message returned on fitting the regression (NULL indicates no error).
See also
Spline modeling is handled by fit_ends_mods_spl()
Examples
bosonc <- create_dummydata("flexbosms")
fit_ends_mods_par(bosonc, expvar=bosonc$ttp.durn)
#> $ttp
#> $ttp[[1]]
#> $ttp[[1]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> rate 0.0437 0.0360 0.0530 0.0043
#>
#> N = 204, Events: 103, Censored: 101
#> Total time at risk: 2358.845
#> Log-likelihood = -425.5134, df = 1
#> AIC = 853.0269
#>
#>
#> $ttp[[1]]$error
#> NULL
#>
#>
#> $ttp[[2]]
#> $ttp[[2]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 0.8379 0.7110 0.9875 0.0702
#> scale 0.0680 0.0448 0.1033 0.0145
#>
#> N = 204, Events: 103, Censored: 101
#> Total time at risk: 2358.845
#> Log-likelihood = -423.0927, df = 2
#> AIC = 850.1853
#>
#>
#> $ttp[[2]]$error
#> NULL
#>
#>
#> $ttp[[3]]
#> $ttp[[3]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.0300 0.8726 1.2158 0.0872
#> scale 15.4175 11.8479 20.0625 2.0716
#>
#> N = 204, Events: 103, Censored: 101
#> Total time at risk: 2358.845
#> Log-likelihood = -422.068, df = 2
#> AIC = 848.136
#>
#>
#> $ttp[[3]]$error
#> NULL
#>
#>
#> $ttp[[4]]
#> $ttp[[4]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> meanlog 2.762 2.462 3.062 0.153
#> sdlog 1.777 1.535 2.057 0.133
#>
#> N = 204, Events: 103, Censored: 101
#> Total time at risk: 2358.845
#> Log-likelihood = -423.7432, df = 2
#> AIC = 851.4864
#>
#>
#> $ttp[[4]]$error
#> NULL
#>
#>
#> $ttp[[5]]
#> $ttp[[5]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 0.80871 0.65294 1.00164 0.08828
#> rate 0.03166 0.02120 0.04730 0.00649
#>
#> N = 204, Events: 103, Censored: 101
#> Total time at risk: 2358.845
#> Log-likelihood = -423.5205, df = 2
#> AIC = 851.0411
#>
#>
#> $ttp[[5]]$error
#> NULL
#>
#>
#> $ttp[[6]]
#> $ttp[[6]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape -0.0376 -0.0652 -0.0101 0.0141
#> rate 0.0603 0.0455 0.0800 0.0087
#>
#> N = 204, Events: 103, Censored: 101
#> Total time at risk: 2358.845
#> Log-likelihood = -421.3123, df = 2
#> AIC = 846.6246
#>
#>
#> $ttp[[6]]$error
#> NULL
#>
#>
#>
#> $ppd
#> $ppd[[1]]
#> $ppd[[1]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> rate 0.01229 0.00854 0.01769 0.00228
#>
#> N = 204, Events: 29, Censored: 175
#> Total time at risk: 2358.845
#> Log-likelihood = -156.5603, df = 1
#> AIC = 315.1206
#>
#>
#> $ppd[[1]]$error
#> NULL
#>
#>
#> $ppd[[2]]
#> $ppd[[2]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.43500 1.09372 1.88278 0.19884
#> scale 0.00352 0.00105 0.01183 0.00218
#>
#> N = 204, Events: 29, Censored: 175
#> Total time at risk: 2358.845
#> Log-likelihood = -153.7336, df = 2
#> AIC = 311.4673
#>
#>
#> $ppd[[2]]$error
#> NULL
#>
#>
#> $ppd[[3]]
#> $ppd[[3]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.553 1.174 2.055 0.222
#> scale 43.036 29.365 63.072 8.393
#>
#> N = 204, Events: 29, Censored: 175
#> Total time at risk: 2358.845
#> Log-likelihood = -154.0749, df = 2
#> AIC = 312.1498
#>
#>
#> $ppd[[3]]$error
#> NULL
#>
#>
#> $ppd[[4]]
#> $ppd[[4]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> meanlog 3.923 3.450 4.397 0.242
#> sdlog 1.289 0.993 1.674 0.172
#>
#> N = 204, Events: 29, Censored: 175
#> Total time at risk: 2358.845
#> Log-likelihood = -154.541, df = 2
#> AIC = 313.0821
#>
#>
#> $ppd[[4]]$error
#> NULL
#>
#>
#> $ppd[[5]]
#> $ppd[[5]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.5585 1.0961 2.2159 0.2799
#> rate 0.0303 0.0152 0.0605 0.0107
#>
#> N = 204, Events: 29, Censored: 175
#> Total time at risk: 2358.845
#> Log-likelihood = -153.7877, df = 2
#> AIC = 311.5754
#>
#>
#> $ppd[[5]]$error
#> NULL
#>
#>
#> $ppd[[6]]
#> $ppd[[6]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 0.03933 0.00680 0.07185 0.01659
#> rate 0.00777 0.00436 0.01385 0.00229
#>
#> N = 204, Events: 29, Censored: 175
#> Total time at risk: 2358.845
#> Log-likelihood = -154.1277, df = 2
#> AIC = 312.2554
#>
#>
#> $ppd[[6]]$error
#> NULL
#>
#>
#>
#> $pfs
#> $pfs[[1]]
#> $pfs[[1]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> rate 0.05596 0.04718 0.06637 0.00487
#>
#> N = 204, Events: 132, Censored: 72
#> Total time at risk: 2358.845
#> Log-likelihood = -512.5726, df = 1
#> AIC = 1027.145
#>
#>
#> $pfs[[1]]$error
#> NULL
#>
#>
#> $pfs[[2]]
#> $pfs[[2]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 0.9313 0.8080 1.0733 0.0675
#> scale 0.0676 0.0453 0.1009 0.0138
#>
#> N = 204, Events: 132, Censored: 72
#> Total time at risk: 2358.845
#> Log-likelihood = -512.0729, df = 2
#> AIC = 1028.146
#>
#>
#> $pfs[[2]]$error
#> NULL
#>
#>
#> $pfs[[3]]
#> $pfs[[3]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.1921 1.0310 1.3784 0.0883
#> scale 11.4917 9.3244 14.1628 1.2254
#>
#> N = 204, Events: 132, Censored: 72
#> Total time at risk: 2358.845
#> Log-likelihood = -513.4304, df = 2
#> AIC = 1030.861
#>
#>
#> $pfs[[3]]$error
#> NULL
#>
#>
#> $pfs[[4]]
#> $pfs[[4]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> meanlog 2.412 2.179 2.645 0.119
#> sdlog 1.535 1.353 1.742 0.099
#>
#> N = 204, Events: 132, Censored: 72
#> Total time at risk: 2358.845
#> Log-likelihood = -516.8304, df = 2
#> AIC = 1037.661
#>
#>
#> $pfs[[4]]$error
#> NULL
#>
#>
#> $pfs[[5]]
#> $pfs[[5]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 0.91518 0.75087 1.11544 0.09240
#> rate 0.04985 0.03633 0.06840 0.00805
#>
#> N = 204, Events: 132, Censored: 72
#> Total time at risk: 2358.845
#> Log-likelihood = -512.1786, df = 2
#> AIC = 1028.357
#>
#>
#> $pfs[[5]]$error
#> NULL
#>
#>
#> $pfs[[6]]
#> $pfs[[6]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape -0.01338 -0.03435 0.00759 0.01070
#> rate 0.06342 0.04926 0.08165 0.00817
#>
#> N = 204, Events: 132, Censored: 72
#> Total time at risk: 2358.845
#> Log-likelihood = -511.7442, df = 2
#> AIC = 1027.488
#>
#>
#> $pfs[[6]]$error
#> NULL
#>
#>
#>
#> $os
#> $os[[1]]
#> $os[[1]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> rate 0.02788 0.02285 0.03402 0.00283
#>
#> N = 204, Events: 97, Censored: 107
#> Total time at risk: 3479.143
#> Log-likelihood = -444.2435, df = 1
#> AIC = 890.4871
#>
#>
#> $os[[1]]$error
#> NULL
#>
#>
#> $os[[2]]
#> $os[[2]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.51432 1.29287 1.77369 0.12215
#> scale 0.00559 0.00254 0.01233 0.00226
#>
#> N = 204, Events: 97, Censored: 107
#> Total time at risk: 3479.143
#> Log-likelihood = -433.3483, df = 2
#> AIC = 870.6965
#>
#>
#> $os[[2]]$error
#> NULL
#>
#>
#> $os[[3]]
#> $os[[3]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.814 1.541 2.136 0.151
#> scale 23.832 20.361 27.895 1.914
#>
#> N = 204, Events: 97, Censored: 107
#> Total time at risk: 3479.143
#> Log-likelihood = -436.9493, df = 2
#> AIC = 877.8987
#>
#>
#> $os[[3]]$error
#> NULL
#>
#>
#> $os[[4]]
#> $os[[4]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> meanlog 3.1627 2.9845 3.3409 0.0909
#> sdlog 1.0166 0.8806 1.1734 0.0744
#>
#> N = 204, Events: 97, Censored: 107
#> Total time at risk: 3479.143
#> Log-likelihood = -439.3941, df = 2
#> AIC = 882.7881
#>
#>
#> $os[[4]]$error
#> NULL
#>
#>
#> $os[[5]]
#> $os[[5]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 1.7620 1.3981 2.2207 0.2080
#> rate 0.0605 0.0435 0.0842 0.0102
#>
#> N = 204, Events: 97, Censored: 107
#> Total time at risk: 3479.143
#> Log-likelihood = -434.1308, df = 2
#> AIC = 872.2617
#>
#>
#> $os[[5]]$error
#> NULL
#>
#>
#> $os[[6]]
#> $os[[6]]$result
#> Call:
#> .f(formula = ..1, dist = ..2)
#>
#> Estimates:
#> est L95% U95% se
#> shape 0.04179 0.02473 0.05885 0.00871
#> rate 0.01532 0.01079 0.02174 0.00274
#>
#> N = 204, Events: 97, Censored: 107
#> Total time at risk: 3479.143
#> Log-likelihood = -433.9344, df = 2
#> AIC = 871.8687
#>
#>
#> $os[[6]]$error
#> NULL
#>
#>
#>
#> $pps_cf
#> $pps_cf[[1]]
#> $pps_cf[[1]]$result
#> NULL
#>
#> $pps_cf[[1]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cf[[2]]
#> $pps_cf[[2]]$result
#> NULL
#>
#> $pps_cf[[2]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cf[[3]]
#> $pps_cf[[3]]$result
#> NULL
#>
#> $pps_cf[[3]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cf[[4]]
#> $pps_cf[[4]]$result
#> NULL
#>
#> $pps_cf[[4]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cf[[5]]
#> $pps_cf[[5]]$result
#> NULL
#>
#> $pps_cf[[5]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cf[[6]]
#> $pps_cf[[6]]$result
#> NULL
#>
#> $pps_cf[[6]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#>
#> $pps_cr
#> $pps_cr[[1]]
#> $pps_cr[[1]]$result
#> NULL
#>
#> $pps_cr[[1]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cr[[2]]
#> $pps_cr[[2]]$result
#> NULL
#>
#> $pps_cr[[2]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cr[[3]]
#> $pps_cr[[3]]$result
#> NULL
#>
#> $pps_cr[[3]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cr[[4]]
#> $pps_cr[[4]]$result
#> NULL
#>
#> $pps_cr[[4]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cr[[5]]
#> $pps_cr[[5]]$result
#> NULL
#>
#> $pps_cr[[5]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#> $pps_cr[[6]]
#> $pps_cr[[6]]$result
#> NULL
#>
#> $pps_cr[[6]]$error
#> <simpleError in model.frame.default(formula = survival::Surv(time = durn1, time2 = durn2, event = evflag) ~ { { expvar }}): variable lengths differ (found for '{ { expvar } }')>
#>
#>
#>