Question

Received the following question from email:

I’m simulating recurrent survival data and I am using log-normal frailties. I want the frailties used in the simulation to appear in the data, so that I can see how well my model estimates the frailties. Right now the simulation looks like this:

nProcess <- 300
rho_weibull_ph <- function(x, lambda, p) lambda * p * x ^ (p - 1)
df=simEventData(z = cbind(rbinom(nProcess, 1, 0.5)),
                zCoef = c(0.5),frailty = stats::rlnorm, endTime = rnorm(nProcess, 10),
                recurrent = TRUE, rho = rho_weibull_ph,
                arguments = list(rho = list(lambda = 0.001, p = 1),frailty = list(sdlog = 0.25)))

My question is how can we get the frailties used in the simulation to be a column in the dataset?

Possible Workaround

The frailty effect is saved in the simEvent object. Therefore, a simple wrapper function is needed to use simEvent() instead
of using simEventData(). For instance,

library(reda)
set.seed(123)
nProcess <- 300
rho_weibull_ph <- function(x, lambda, p) lambda * p * x ^ (p - 1)
dat <- do.call(rbind, lapply(seq_len(nProcess), function(i) {
    obj <- simEvent(z = rbinom(1, 1, 0.5), zCoef = 0.5,
                    rho = rho_weibull_ph, frailty = stats::rlnorm,
                    arguments = list(rho = list(lambda = 0.2, p = 1),
                                     frailty = list(sdlog = 0.25)))
    out <- reda:::simEvent2data(i, obj)
    out$frailty <- obj@frailty$frailty
    out
}))
head(dat, 10)
##    ID      time event origin X  frailty
## 1   1 0.1292489     1      0 0 1.221572
## 2   1 0.3593253     1      0 0 1.221572
## 3   1 1.6547921     1      0 0 1.221572
## 4   1 2.9409516     1      0 0 1.221572
## 5   1 3.0000000     0      0 0 1.221572
## 6   2 1.2635873     1      0 0 1.525842
## 7   2 2.1844353     1      0 0 1.525842
## 8   2 3.0000000     0      0 0 1.525842
## 9   3 1.2199754     1      0 0 1.563189
## 10  3 3.0000000     0      0 0 1.563189

Dear Wenjie:

First, I would like to let you know this is a great package. I am working on a project to calculate a large number of mcf. It is very slow due the number of mcf calculations. I believe the confidence interval is more time consuming than mcf. Is there a way I can turn off the confidence interval calculation.

Thanks,

Sincerely yours,

Hi,

I think there is a bug in simEventData. I used the following function to generate recurrent event with time-variant covariate. zt is a vector of length 553, which is attached.
zt.csv

AT<-function(time,zt) zt[pmax(0,floor(time-100))+1]
zfun<-function(time,zt,Z2) cbind(AT(time,zt),Z2) ## time-variant covariate vector
x=simEventData(z = zfun, zCoef = c(0.1,2),origin=0,endTime=length(zt),arguments = list(z = list(zt=zt,Z2=1)),rho=0.001)

With the above codes, I get the following results:

x
ID time event origin X.1 X.2
1 1 26.28915 1 0 100.00000 1
2 1 72.58110 1 0 100.00000 1
3 1 97.80675 1 0 100.00000 1
4 1 139.48710 1 0 14.96476 1
5 1 148.87831 1 0 12.75497 1
6 1 171.86007 1 0 11.22835 1
7 1 204.00038 1 0 9.49048 1
8 1 212.41692 1 0 9.10000 1
9 1 216.82910 1 0 9.70141 1
10 1 239.12504 1 0 9.49023 1
11 1 298.58581 1 0 9.37119 1
12 1 448.02677 1 0 28.17732 1
13 1 453.27768 1 0 30.40548 1
14 1 454.55875 1 0 30.87229 1
15 1 465.77483 1 0 36.38604 1
16 1 553.00000 0 0 87.94848 1

However, when I concatenate a vector of 10 to zt as follows, simEventData generate a lot of events. Since the time-variant covariate values did not change for the time period 0 to 553, so many events in this time period does not make sense.

zt=c(zt,rep(10,533))
x=simEventData(z = zfun, zCoef = c(0.1,2),origin=0,endTime=length(zt),arguments = list(z = list(zt=zt,Z2=1)),rho=0.001)

x
ID time event origin X.1 X.2
1 1 0.006242786 1 0 100 1
2 1 0.006839542 1 0 100 1
3 1 0.017938353 1 0 100 1
4 1 0.028651440 1 0 100 1
5 1 0.033637245 1 0 100 1
6 1 0.046612646 1 0 100 1
7 1 0.051124358 1 0 100 1
8 1 0.052338844 1 0 100 1
9 1 0.063594936 1 0 100 1
10 1 0.067455526 1 0 100 1
11 1 0.083278221 1 0 100 1
12 1 0.083860555 1 0 100 1
13 1 0.092751879 1 0 100 1
14 1 0.092862057 1 0 100 1
15 1 0.096591501 1 0 100 1
16 1 0.106756317 1 0 100 1
17 1 0.109465642 1 0 100 1
18 1 0.118664398 1 0 100 1
19 1 0.118995341 1 0 100 1
20 1 0.129560419 1 0 100 1
21 1 0.140414691 1 0 100 1
22 1 0.144648840 1 0 100 1
23 1 0.152500856 1 0 100 1
24 1 0.153273102 1 0 100 1
25 1 0.156321097 1 0 100 1
26 1 0.174974314 1 0 100 1
27 1 0.180033237 1 0 100 1
28 1 0.186355416 1 0 100 1
29 1 0.191305847 1 0 100 1
30 1 0.200493667 1 0 100 1
31 1 0.203004847 1 0 100 1
32 1 0.203831227 1 0 100 1
33 1 0.213427058 1 0 100 1
34 1 0.219884156 1 0 100 1
35 1 0.226711902 1 0 100 1
36 1 0.238861370 1 0 100 1
37 1 0.248352938 1 0 100 1
38 1 0.249743697 1 0 100 1
39 1 0.251025185 1 0 100 1
40 1 0.259059073 1 0 100 1
41 1 0.260553979 1 0 100 1
42 1 0.265022980 1 0 100 1
43 1 0.265408599 1 0 100 1
44 1 0.266536777 1 0 100 1
45 1 0.268670326 1 0 100 1
46 1 0.292304101 1 0 100 1
47 1 0.302883653 1 0 100 1
48 1 0.303276499 1 0 100 1
49 1 0.306170839 1 0 100 1
50 1 0.312236723 1 0 100 1
51 1 0.313719985 1 0 100 1
52 1 0.317544096 1 0 100 1
53 1 0.320767841 1 0 100 1
54 1 0.323275411 1 0 100 1
55 1 0.335794068 1 0 100 1
56 1 0.336983959 1 0 100 1
57 1 0.344310057 1 0 100 1
58 1 0.349678224 1 0 100 1
59 1 0.360598594 1 0 100 1
60 1 0.372449873 1 0 100 1
61 1 0.378575765 1 0 100 1
62 1 0.392175993 1 0 100 1
63 1 0.394257982 1 0 100 1
64 1 0.395818677 1 0 100 1
65 1 0.397277389 1 0 100 1
66 1 0.408018247 1 0 100 1
67 1 0.413464217 1 0 100 1
68 1 0.417847208 1 0 100 1
69 1 0.427381956 1 0 100 1
70 1 0.458754522 1 0 100 1
71 1 0.460189161 1 0 100 1
72 1 0.467391060 1 0 100 1
73 1 0.474431766 1 0 100 1
74 1 0.482510073 1 0 100 1
75 1 0.495227621 1 0 100 1
76 1 0.497469003 1 0 100 1
77 1 0.497635925 1 0 100 1
78 1 0.508283200 1 0 100 1
79 1 0.509404770 1 0 100 1
80 1 0.511629727 1 0 100 1
81 1 0.513421083 1 0 100 1
82 1 0.519522748 1 0 100 1
83 1 0.527497089 1 0 100 1
84 1 0.529607839 1 0 100 1
85 1 0.543771408 1 0 100 1
86 1 0.544726510 1 0 100 1
87 1 0.546149712 1 0 100 1
88 1 0.549375325 1 0 100 1
89 1 0.551445292 1 0 100 1
90 1 0.555493815 1 0 100 1
91 1 0.563718817 1 0 100 1
92 1 0.567767155 1 0 100 1
93 1 0.577351468 1 0 100 1
94 1 0.581826004 1 0 100 1
95 1 0.584861912 1 0 100 1
96 1 0.587809261 1 0 100 1
97 1 0.589983361 1 0 100 1
98 1 0.597754619 1 0 100 1
99 1 0.600915567 1 0 100 1
100 1 0.600990019 1 0 100 1
101 1 0.603297320 1 0 100 1
102 1 0.605375022 1 0 100 1
103 1 0.614614531 1 0 100 1
104 1 0.619029684 1 0 100 1
105 1 0.620205188 1 0 100 1
106 1 0.625461296 1 0 100 1
107 1 0.633465294 1 0 100 1
108 1 0.633507280 1 0 100 1
109 1 0.640759321 1 0 100 1
110 1 0.642790922 1 0 100 1
111 1 0.644641142 1 0 100 1
112 1 0.656446538 1 0 100 1
113 1 0.660237826 1 0 100 1
114 1 0.663517186 1 0 100 1
115 1 0.663615622 1 0 100 1
116 1 0.664658161 1 0 100 1
117 1 0.665626199 1 0 100 1
118 1 0.678755968 1 0 100 1
119 1 0.681003655 1 0 100 1
120 1 0.681719499 1 0 100 1
121 1 0.690110526 1 0 100 1
122 1 0.696531586 1 0 100 1
123 1 0.710909445 1 0 100 1
124 1 0.711612790 1 0 100 1
125 1 0.728442951 1 0 100 1
126 1 0.741216006 1 0 100 1
127 1 0.745733117 1 0 100 1
128 1 0.768354068 1 0 100 1
129 1 0.769063435 1 0 100 1
130 1 0.771498263 1 0 100 1
131 1 0.772921513 1 0 100 1
132 1 0.791330999 1 0 100 1
133 1 0.815834509 1 0 100 1
134 1 0.815878029 1 0 100 1
135 1 0.816734014 1 0 100 1
136 1 0.817243638 1 0 100 1
137 1 0.819825244 1 0 100 1
138 1 0.823388786 1 0 100 1
139 1 0.844831005 1 0 100 1
140 1 0.846703023 1 0 100 1
141 1 0.856267187 1 0 100 1
142 1 0.858514492 1 0 100 1
143 1 0.861123661 1 0 100 1
144 1 0.875305698 1 0 100 1
145 1 0.875617931 1 0 100 1
146 1 0.876870024 1 0 100 1
147 1 0.886627804 1 0 100 1
148 1 0.887348842 1 0 100 1
149 1 0.889334413 1 0 100 1
150 1 0.902878098 1 0 100 1
151 1 0.906231215 1 0 100 1
152 1 0.907070927 1 0 100 1
153 1 0.907322761 1 0 100 1
154 1 0.907600294 1 0 100 1
155 1 0.907841225 1 0 100 1
156 1 0.910605047 1 0 100 1
157 1 0.927590676 1 0 100 1
158 1 0.941816746 1 0 100 1
159 1 0.947595752 1 0 100 1
160 1 0.953814089 1 0 100 1
161 1 0.973006674 1 0 100 1
162 1 0.978912536 1 0 100 1
163 1 0.999507157 1 0 100 1
164 1 1.007360715 1 0 100 1
165 1 1.010971548 1 0 100 1
166 1 1.015471441 1 0 100 1
[ reached 'max' / getOption("max.print") -- omitted 31878 rows ]

Hi Steven, what is the intended usage of the isDate slot?

First of all, this is an amazing versatile package. It is very useful, especially for simulating event data.

I am currently looking at the following example from this vignette.

zFun3 <- function(time, a, b) cbind(time / 10 + a, as.numeric(time > b))
(simDat <- simEventData(nProcess = 3, z = zFun3, zCoef = c(- 0.1, 0.5),
                        arguments = list(z = list(a = quote(rpois(1, 10) / 10),
                                                  b = quote(runif(1, 1, 3))))))

I was trying to modify this code to fix the values of a:

zFun3 <- function(time, a, b) cbind(time / 10 + a, as.numeric(time > b))
(simDat <- simEventData(nProcess = 3, z = zFun3, zCoef = c(- 0.1, 0.5),
                        arguments = list(z = list(a = c(0.1, 0.2, 0.3),
                                                  b = quote(runif(1, 1, 3))))))

However, this does not work. Any suggestions how to have different but deterministic values for inputs of z?