Hello,
I have a study where I am evaluating the relationship between the total number of health care visits and geographic access (measured by travel time categories).
Initially I thought this would be a mixed-effects Poisson regression - total visits + individual level fixed effects & random intercept at geographic area level -- and using robust SE.
However, I have a significant amount of no-visits (25.7%) for my dependent variable.
Therefore I'm looking at a mixed effect zero-inflated negative binomial model. I am wondering if this is logical or if others have recommendations.
Thanks,
Devin
I have a study where I am evaluating the relationship between the total number of health care visits and geographic access (measured by travel time categories).
Initially I thought this would be a mixed-effects Poisson regression - total visits + individual level fixed effects & random intercept at geographic area level -- and using robust SE.
However, I have a significant amount of no-visits (25.7%) for my dependent variable.
Therefore I'm looking at a mixed effect zero-inflated negative binomial model. I am wondering if this is logical or if others have recommendations.
Thanks,
Devin
