estimate_year_1_2_diff_pooled: Estimate pooled year 1 to 2 usage differencs
In evanjflack/polypharmacy: Behavioral Hazard and Patient Decision Making

Estimate the first stage of likelihood that people with a year 2 usage amount had a specific usage amount in year 1

estimate_year_1_2_diff_pooled(
  DT,
  diff_var,
  year2_var,
  u_diff_var,
  x_main,
  x_int,
  exclude = TRUE
)

`DT`	data.table
`diff_var`	character, name of the column with the year 1 - year 2 amount
`year2_var`	character, name of the column of the year 2 usage amount
`u_diff_var`	numeric vecotr with all unique amounts of the year 1/2 difference
`x_main`	character,
`x_int`	character vector,
`exclude`	logical (default = TRUE), if TRUE, then excludes people with year 2 amoutns that are not compatible with the year 1 difference we are trying to estimate