My study comprises of repeatedly measured maternal smoking (secondhand smoking) variable (binary 0,1) at different time points of 2months, 6 months, 5years, 9years and 13years and outcomes as childhood dental caries (dmft index, count data) measured at 5y and 13y. My preferred choice of analysis was negative binomial hurdle model as most of my values are zero inflated and calculate odds ratio. However as i have repeated measures of dmft index is it possible to combine a repeated masurement model with a hurdle model. What kind of complexity can I expect if this is possible?
Secondly, as the dentition at 5y(primary teeth) and 13y (permanenet dentition) are two completely different dentitions so would the measures be corelated with each other, although in the same person?
If they may not be correlated then, would it be more appropriate to carry out analysis at two different time points i.e no repeated measures.
