[R] Create a time interval from a single time variable

[David Winsemius]

Not sure how to functionalize it either. What seemed promising  
(assuming it has first been sorted by ID and DaysEnrolled) would be  
three-step process:

sample1$Stop <-  c(sample1[2:nrow(sample1),"DaysEnrolled"],NA)  #  
shift the DaysEnrolled
sample1$next.id <-  c(sample1[2:nrow(sample1),"ID"],NA)         #  
shift ID
is.na(sample1$Start) <-   with(sample1, ID != next.id)          # NA  
the ends of ID groups

-- 
David winsemius


On Jun 3, 2009, at 2:15 PM, Katschke, Adrian R wrote:

> I am trying to set up a data set for a survival analysis with time- 
> varying covariates. The data is already in a long format, but does  
> not have a variable to signify the stopping point for the interval.  
> The variable DaysEnrolled is the variable I would like to use to  
> form this interval. This is what I have now:
>
>     ID   Age DaysEnrolled HAZ WAZ WHZ Food  
> onARV                         HIVStatus LTFUp Start Stop
> 1 71622 0.008            0  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 2 71622 0.085           28  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 3 71622 0.123           42  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 4 71622 0.277           98  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 5 71622 0.441          158  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 6 71622 0.517          186  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 7 71622 0.594          214  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 8 71622 0.715          258  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
> 9 71622 0.791          286  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    0
>
> This is what I would like to have:
>
>     ID   Age DaysEnrolled HAZ WAZ WHZ Food  
> onARV                         HIVStatus LTFUp Start Stop
> 1 71622 0.008            0  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0     0    28
> 2 71622 0.085           28  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    28    42
> 3 71622 0.123           42  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    42    98
> 4 71622 0.277           98  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    98    158
> 5 71622 0.441          158  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    158   186
> 6 71622 0.517          186  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    186   214
> 7 71622 0.594          214  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    214   258
> 8 71622 0.715          258  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    258   286
> 9 71622 0.791          286  NA  NA  NA   NA     0 HIV exposed,  
> status indeterminate     0    286    NA
>
> I am not sure how to put this in a function. I thought of using  
> embed() in tapply().
>
> astop <- tapply(sample1$DaysEnrolled, sample1$ID, function(x){
>                                                          
> ifelse(length(x) == 1,
>                                                         embed(x,1),  
> ifelse(length(x) > 1,
>                                                         embed(x,2),  
> NA))})
>
> This doesn't do what I thought it would. I know that I could write a  
> double loop to look at each subject and the differing number of  
> observations for each subject, but would like to avoid that it at  
> all possible.
>
>
> Sample of 2 subjects:
>           sample1 <-
> structure(list(ID = c(71622L, 71622L, 71622L, 71622L, 71622L,
> 71622L, 71622L, 71622L, 71622L, 1436L), Age = c(0.008, 0.085,
> 0.123, 0.277, 0.441, 0.517, 0.594, 0.715, 0.791, 6.968),  
> DaysEnrolled = c(0L,
> 28L, 42L, 98L, 158L, 186L, 214L, 258L, 286L, 0L), HAZ = c(NA_real_,
> NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
> NA_real_, NA_real_), WAZ = c(NA_real_, NA_real_, NA_real_, NA_real_,
> NA_real_, NA_real_, NA_real_, NA_real_, NA_real_, NA_real_),
>     WHZ = c(NA_real_, NA_real_, NA_real_, NA_real_, NA_real_,
>     NA_real_, NA_real_, NA_real_, NA_real_, NA_real_), Food =  
> c(NA_integer_,
>     NA_integer_, NA_integer_, NA_integer_, NA_integer_, NA_integer_,
>     NA_integer_, NA_integer_, NA_integer_, NA_integer_), onARV = c(0L,
>     0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), HIVStatus = structure(c(2L,
>     2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("", "HIV  
> exposed, status indeterminate",
>     "HIV infected", "HIV negative"), class = "factor"), LTFUp = c(0,
>     0, 0, 0, 0, 0, 0, 0, 0, NA), Start = c(0, 0, 0, 0, 0, 0,
>     0, 0, 0, 0), Stop = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0)), .Names =  
> c("ID",
> "Age", "DaysEnrolled", "HAZ", "WAZ", "WHZ", "Food", "onARV",
> "HIVStatus", "LTFUp", "Start", "Stop"), row.names = c(NA, 10L
> ), class = "data.frame")
>
>
> Adrian Katschke
> Biostatistician
> IU Department of Medicine
> Division of Biostatistics
> akatschk at iupui.edu
> 317-278-6665
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

David Winsemius, MD
Heritage Laboratories
West Hartford, CT