What model should I use?

#1
I have a longitudinal survey (5 waves) that collects a health index (from 0 to 100) and some risk factors and socioeconomic characteristics. I want to develop a model that will predict this health index (with consistent biographies, meaning that the predicted health index should not be independent of the past health index). What kind of regression model should I use? My first idea was simply to run a logit regression on the index including in predictors the value of the index of the previous wave, but I don’t think it would be the most theoretically appropriate way to do it.
 
#2
This may be something that you have to use HLM for (hierarchical linear modeling, nested models). I say HLM because you it sounds as if you are trying to predict change over time (change at each time period, measured from baseline).

Note: SPSS does not do hierarchical linear modeling/nested models, but there is a program called HLM that is free for limited use or for student use I believe. R does HLM if you know programming.

Hope this helps.