I have a set of about five variables of interest (continuous and categorical), collected at four time points: baseline; six months; 12-months and 24 months. All groups got the same questions and therapy but stayed in treatment for four different time intervals (Group 1: baseline only; Group 2: baseline and six moths; Group 3: baseline, six months, 12 months; Group 4: baseline, six months, 12 months, 24 months. I am trying to find an analysis that would identify significant variations at different points in time but am hampered by the fact that people dropped out at each stage. Have run analyses with just the 24 month group but would like to compare all those who completed baseline, etc. as well. Does anyone work with data where there is no experimental condition -- all treatment the same -- but the duration changes? Would love to hear what analysis you use. Thanks.