Do you want to refer to procedure B on the basis of success or failure or whatever outcome?

Hi Syed.

I mean whatever outcome. For example, suppose a chemotherapy unit. What percentage of patients are referred to radiology, hormone therapy, oncologist, etc.

Thanks in advance for your interest.

Edgar

Dear Edgar,

Yes,

Firstly, I think, there are many possibilities to understand, what kind of treatments, procedures, and medicines, among others, are derived from that disease. At this point, I think you should design a mixed study. Because you need to select only one kind of Cancer, and what is the traceability of the illness?; in this part is better to have Cancer specialists, not literature. There are two ways, interviews or reading articles. As a result a Bayesian network (probably).

Secondly, In Colombia studying Cancer is essential, due to that illness is high cost.

Thirdly, although medical traceability on databases in Colombia is not too good -in a national scope-, you can base on those registers for the quantitative study, the Colombian Health Ministry can provide that data, but In my opinion, you need to analyze that information with tweezers. Getting that information is relatively easy and static (per period, not by a person).

Finally, you need to delimit the scope. I think, probably, at a Hospital you can get some medical records and also traceability, but they are jealous of that; the procedure to analyze that information is the problem.

Good luck with your research!

Sincerely,

Johanna :)

Dear Johanna,

It's great to hear from you. Thanks a lot for your answer. In fact, I've been working with secondary data from SISPRO (Colombian national healthcare registry) (https://www.sispro.gov.co/Pages/Home.aspx). This system has detailed records about the services received by every patient in Colombia, including insurance companies, location, age, CUPS code, etc. My question goes beyond these records and perhaps I didn't state it properly.

I'd like to know about studies that may have previously obtained generic transition probabilities matrices for cancer treatments.

I'll try to be more clear with the following example:

Let i and j be services belonging to a cancer clinical pathway (diagnosis, surgery, therapy, etc). The index i represents origin services while j represents destiny services.

Then, I wanto to get a probability matrix such as this:

j1 j2 j3

i1 0 0,45 0,1

i2 0 0 0,15

i3 0 0 0

For example 0.1 is the probability for patients to be referred from service i1 to service j3.

A few years ago, SISPRO offerred this information with anonimized data, but it is now impossible. Therefore I thought that maybe some other researchers have worked with similar questions.

I expect not to be so boring anyway. Jajajaja.

Hope to see you soon!

Edgar