Heather Allore, PhD, Core Leader for the IMPACT Design and Statistics Core, recently appeared on the JAMAevidence podcast to discuss latent class analysis to identify hidden clinical phenotypes. A latent variable is an unobserved variable that investigators don’t have a construct or measure for, but believe it exists and could impact outcomes. Dr. Allore uses frailty as an example of latent variables in the discussion.
The podcast was hosted by JAMA Statistical Editor Roger Lewis, MD, PHD and is a supplement to the JAMA Guide to Statistics and Methods.