04: Generating Estimates: Age-standardization
In RSTr: Gibbs Samplers for Discrete Bayesian Spatiotemporal Models

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
library(RSTr)
is_cran <- !identical(Sys.getenv("NOT_CRAN"), "true")

Overview

In the previous vignette, we discussed the model setup process in-depth. But how do we get our estimates once we've run our model? In this vignette, we discuss extracting estimates from our model object with the get_estimates() function, and how to age-standardize those estimates with age_standardize().

The `get_estimates()` function

In the RSTr introductory vignette, we generated age-standardized estimates for lambda based on our example Michigan dataset. To extract rates from an RSTr object, we can simply run get_estimates():

mod_mst <- mstcar(name = "my_test_model", data = miheart, adjacency = miadj)

# For computational reasons, full model fitting is not run during CRAN checks.
# When building on CRAN, this vignette loads a pre-fitted example model included with the package.
# The pkgdown website shows the full model-fitting workflow.
example_dir <- system.file("extdata", package = "RSTr")
mod_mst <- load_model("mstcar_example", example_dir)

estimates <- get_estimates(mod_mst, rates_per = 1e5)
head(estimates)

The `age_standardization()` function

In many cases, we will want to age-standardize our estimates based on some (or all) age groups in our dataset. In our Michigan dataset, we have six ten-year age groups over which we can standardize; let's age-standardize from ages 35-64. For RSTr objects, age_standardize() takes in four arguments:

RSTr_obj: The RSTr model object created with *car();
std_pop: A vector of standard populations associated with the age groups of interest. Since our Michigan data is from 1979-1988, we can use 1980 standard populations from NIH. It is recommended that you use the standard population that is most closely associated with your dataset;
new_name: The name of your new standard population group; and
groups: A vector of names matching each group of interest. To age-standardize by all groups in a dataset, leave this argument blank.

Once we have our std_pop vector, we can age-standardize our estimates:

std_pop <- c(113154, 100640, 95799)
mod_mst <- age_standardize(mod_mst, std_pop, new_name = "35-64", groups = c("35-44", "45-54", "55-64"))
mod_mst

Notice now that the mod_mst object indicates we have age-standardized our estimates and the names of our age-standardized group. We can also add on to our list of age-standardized estimates by simply specifying a different group:

std_pop <- c(68775, 34116, 9888)
mod_mst <- age_standardize(mod_mst, std_pop, new_name = "65up", groups = c("65-74", "75-84", "85+"))
mod_mst

If we want to generate estimates for all groups, i.e. 35 and up, we can omit the groups argument and expand std_pop to include all of our populations:

std_pop <- c(113154, 100640, 95799, 68775, 34116, 9888)
mod_mst <- age_standardize(mod_mst, std_pop, new_name = "35up")
mod_mst
mst_estimates_as <- get_estimates(mod_mst)
head(mst_estimates_as)

Now, get_estimates(mod_mst) shows the age-standardized estimates as opposed to our non-standardized estimates. Should you want to see the non-standardized estimates instead, you can set the argument standardized = FALSE.

Final thoughts

In this vignette, we explored the get_estimates() function and investigated age-standardization with the age_standardize() function. Age-standardization is one of the most important features of the RSTr package; using just a few arguments, we can easily generate estimates across our population groups.