Issue with specifying vce type with 'stintcox' command

[Miranda Loutet]

Hello,
I am trying to run a Cox proportional Hazard model with interval censoring so am using the 'stintcox' command. I have multiple records per participant so I want to account for intragroup correlation. The help document says I cam use vce(cluster clustervar), but when I do that I get an error that says "cluster id not allowed in vce() option". Any idea why this is happening with the stintcox command? hoping the authors of the command might see this post

[Miranda Loutet]

Sorry I should have said that I have set-up the data using Single-record-per-subject interval-censored data, so that there is a lower and upper-bound time and the event happened at some point in between. but I have two ways that the outcome was measured (identified as the variable 'metric') so that is why there are two records per participant. I have an interaction term between the exposure and the metric to determine if there is a modifying effect of the metric. I need to account for the fact that there are two records per participant though.

Here is some sample data:
Participant metric education_bin rtime ltime
1 1 Some secondary or higher 248 241
1 2 Some secondary or higher 248 247
2 1 Primary or less 180
2 2 Primary or less 180
3 1 Some secondary or higher 302 295
3 2 Some secondary or higher 302 301
4 1 Primary or less 49 43
4 2 Primary or less 187 186

and this is what my code looks like right now: stintcox i.metric##i.education_bin, interval(ltime rtime), which is missing any accounting for intragroup correlation.