Using the library for disease progression

[ogreyesp]

Hi

I'm very interested in using this library for generating time series that represent some disease progression. In particular, I'm working on diseases where disability levels in the persons commonly increase along time, like in Parkinson or Multiple Sclerosis.

How can I use this library to obtain time series where a disability-level (continuous variable) increases over time?

[manitadayon]

Hi
Do you have a DAG or the Bayesian network for your problem?
Do you know the distribution of the variables in your network?
There are many ways to generate the continuous time series using this package. I will be more than happy to help you if you can tell me more about your problem. What is the range of your continuous variable? and how many variables you have?
If you can share more details about your problem I will be more than happy to help you.

[ogreyesp]

Hi

Thank you for your response.

Well, I don’t the DAG of my problem, but I can briefly summarise it as follows:

• The problem studied is regarding multiple sclerosis, which is a chronic lifelong disease with currently no known cure.

• The most common disease course – is characterized by clearly defined attacks of new or increasing neurologic symptoms. These attacks – also called relapses or exacerbations – are followed by periods of partial or complete recovery (remissions).

• Along the time, the disability level of patients with MS increases, as shown in the figure 1 at this link.
  https://www.nationalmssociety.org/What-is-MS/Types-of-MS/Relapsing-remitting-MS

• So, we can say that we have the following states:

  • relapse state, wherein the disability level of the patient commonly increases. A relapse state is always followed by a remission one.
  • remission state, wherein the patient recovers completely or partially the disability level that he/she previously had before the relapse.
  • active without worsening, wherein the disability level does not increase (stable)
  • stable without activity, wherein the disability level does not increase (stable)

There is one continuous variable that I want to model, the disability level. We can say that this variable goes from 0 (no disability) to 1 (death). The probabilities to move from one state to another may be assigned randomly.

My main doubt here is that the disability level must tend to increase along the time disregarding the state and I don’t know how to model this kind of dependence. Also, bear in mind that it is quite difficult to define a CDF for the continuos variable for each state, because the disability level increases along time. Consequently, the disability level could have a CDF for early relapses, and another completely CDF for later relapses.

I'm new in bayesian network analysis, I don't know if it is possible to create a DAG that models such a progression of MS, and thus allow me to create synthetic time series that represent the disability progression. I would be more than happy to have a toy sample that could approximate this kind of problem.

I would appreciate you help,

Thanks in advance