I am building a simulation model in R. In this model, I would like to simulate a patient with a certain baseline age, and then simulate the time-to-event for breast cancer. I want to sample the time-to-event from a distribution that I fit on data from literature, but I am not sure how to go about. I have included the table with literature data that I want to use below.
Ideally, I'd like to fit some sort of cumulative incidence curve and sample from that. But I am not sure how to get this to work without individual patient data. The data is also not traditional survival data in the sense that you start out with an X amount of patients and during follow up some get events at certain times and others do not (or get censored). That would probably make it a bit easier to approach.
Currently, I have tried solving my problem by making a file with 100 'fake' patients that get events at either 25, 35, 45, 55, etc. years of age according to the cumulative incidence that was reported at each age category. So, for instance, I have 4 patients with a follow-up time of 25 years before getting an event, and then I have 20 patients with a follow-up time of 35 years before getting an event, etc. The patients without an event are censored at 85 years of age. However, I feel there should be a better solution that uses the given data in a better way.
I did find an R package that can reconstruct Kaplan-Meier curves based on literature, but it is more aimed at published curves and the aim to find estimates from that, which is not quite the same as my problem here.