Hello Mary,
at the risk of not having understood your question: wouldn't a cross-sectional survey do i.e assuming you want to detect clustering? You'd assess for the deficiency using clinical signs (xerophthalmia) and or/other biomarkers say serum retinol. From the literature, your survey subjects would be children in this case.
This sounds too simplistic, though and you may be inquiring about something else: ifso, please clarify what your question.

The question was as simple as it was....just clustering. Studies by FAO have shown that the occurrence of forms of VAD are usually clustered and not evenly distributed in a population. This may mean that there are epidemiological traits that cause this clustering. Given that the traits may be very dynamic too (keeps changing), how then can these traits which I believe are the risk factors be determined?

Hello Mary

If I have understood the question correctly think the only way to answer it would be to construct your hypothesis regarding the potential risk factors and then do a series of different aetiological studies in the locations of interest; including longitudinal cohort studies.

However, donor would be unlikely to put money in to such studies as the causes are thought to be well established and it might be better spent doing operational research to improve the capsule distribution system, and WASH and FSL interventions.

I hope that addresses your question adequately?