1 item has been added to your cart.
Participants of the webinar Analyzing interval-censored survival-time data in Stata, which took place on January 24, 2018, asked StataCorp developers the following:
| Question: | Why did you choose a Weibull model? |
|---|---|
| Response: | The Weibull model is chosen simply as an example. Xiao will show how to compare the goodness-of-fit of the models later in the slides. |
| Question: | How do you apply this model to time-updated survival analysis? |
| Response: | By "time-updated" survival analysis, do you mean time-varying or time-dependent covariates? If so, these are not yet supported with stintreg. |
| Question: | What is ln_p? |
| Response: | ln_p is the log of the shape parameter p of the Weibull distribution. |
| Question: | Hello, which dataset is used? |
| Response: | For this example, you can type webuse lungtumor in Stata 15 to access this dataset. |
| Question: | Does stintreg support the svy prefix? |
| Response: | Yes, svy is supported with stintreg. |
| Question: | What command would we use instead if we have time-varying covariates? |
| Response: | If you have uncensored or right-censored survival data, you can use streg or stcox. Time-varying covariates are not yet supported for general interval-censored survival data. |
| Question: | Any suggestions to model interval-censored outcomes when there is a competing risk? |
| Response: | If you have uncensored or right-censored survival data, you can use stcrreg. Competing risks are not supported with general interval-censored survival data yet. |
| Question: | Can the plot of survivor functions be used for a combination of several factor values? |
| Response: | Yes, you can plot the survivor function for different levels of several factor variables. |
| Question: | Is it possible in some way to analyze case II interval-censored data where multiple similar events can occur in the interval? |
| Response: | Multiple failures are not yet supported for general interval-censored survival data. We will put this on our development list. With uncensored or right-censored data, stcox and streg support multiple failures. |
| Question: | Is there an ado-file in Stata for nonparametric estimation of the survivor function in case of interval-censored data. There are several appoaches published in the literature, but I could not find one that has been implemented in Stata. |
| Response: | estat gofplot is using one nonparametric method. And we are considering adding more in the future. |
| Question: | You mentioned to include stratification in a multivariate survival analysis. Is it a test to check if it is necessary to consider group variation rather than including it? |
| Response: | Stratification allows you to model baseline hazard and ancillary parameters differently for each group. Whether the stratification is needed can be inferred from the tests reported in the output table for stratum-specific parameters. |
| Question: | When there are multiple intervals of time (3, 6, 9 months as an example), can you not use this (b/c of left/right)? Do you instead have to run multiple models? |
| Response: | Different observations can have different width intervals, and the likelihood for stintreg handles this type of censoring. |
If you have additional questions about the webinar or about analyzing interval-censored survival-time data using Stata, please contact our technical services department.
