How to take the difference of marginals.fitted.values from two inla models


May someone familiar with the R-INLA Project assist with the question below.

I am trying to produce the posterior difference of marginals.fitted.values from two SPDE models e.g. Chapter 9 Spatial modeling of geostatistical data. Malaria in The Gambia | Geospatial Health Data: Modeling and Visualization with R-INLA and Shiny.

Both models have been fitted to the same spatial points but have used data from two different years. I am interested in knowing whether the prevalence of a disease changed between the two time points. Hence I have opted to do the difference of the posteriors of the marginals.fitted.values from the two models.

I thought this would be a simple subtraction but the structure of the marginals.fitted.values does not allow this.

Any guidance will be appreciated and sorry I may have used wrong terminology I am new in the field. You can also point me to the resources that explain this in detail.

My naive approach that did not work, below:
model1$marginals.fitted.values[index] - model2$marginals.fitted.values[index]

I have asked the same question on the inla project platform but the answers were still too abstract for me to understand. Here is the link to that discussion.

Let me know if I need to add further details.