Dear all,
I have a set of cohort dat including about 2000 children 0-9 years old
followed-up for 5 years. We carried out a series of bimonthly
cross-sectional surveys (a total of 22 surveys) to determine the infection
status in each child. Those with complete follow-up for instance would have
yielding about 22 blood results. Since it was a dynamique cohort, those who
entered the cohort late or those who miss certains surveys may have less.
I first would like to determine the age trend of the incidence of malaria
infection (age-specific) from cross-sectional surveys. It should be notice
that an infection may last more than 3 months. A child can change infection
status several times during the 5 years period.Only 1% of the children
remain constantly infected during the study period, while no child remain
constantly negative over five year period.
I would like some suggestion about which statistical model might be relevant
for this matter and which can allow also to estimate the average duration of
infection.
The reversible catalytic model (which is based on the Markov process)has
been used, but I have no idea how that could be apply using S-Plus.
I would like to model the incidence rate using the markov model you know
pretty well. Some authors have used the reversible catalytic model which is
based on the markovian principle (2 states: infected, non infected). I would
like your advise on this matter.
____________________________________________
Department of Epidemiology,
Tulane School of Public Health &Tropical Medicine
SL 2000, 1440 Canal St New Orleans, LA 70112
Tel: Off: (504) 988 8255
|