Here is my suggestion: Consider the context of interest. What would be the basic health information you expect people to know to be considered health literate in the given context? Come up with a list of 10-20 items. Test the list in a pilot to validate and adapt it further. You can then use it in a cross-sectional survey to determine to what extent it would predict "healthy behavior", keeping in mind that health literacy may not necessarily result in healthy behavior depending on the health behavior in question. A good example of this is smoking. Knowing it is unhealthy does not always result in people not smoking.
Here are some examples of papers about adapting health literacy measures for different countries and cultures:
• Ndahura, N. B. (2012). Nutrition literacy status of adolescent students in Kampala district, Uganda.
• Patel, P., Panaich, S., Steinberg, J., Zalawadiya, S., Kumar, A., Aranha, A., & Cardozo, L. (2013). Use of nutrition literacy scale in elderly minority population. The journal of nutrition, health & aging, 17(10), 894-897.
• Sampaio, H. A. D. C., Silva, D. M. D. A., Sabry, M. O. D., Carioca, A. A. F., & Chayb, A. P. V. (2013). Letramento nutricional: desempenho de dois grupos populacionais brasileiros; Nutrition literacy: performance of two Brazilian population groups. Nutrire Rev. Soc. Bras. Aliment. Nutr, 38(2).
Or look at the ways Newest Vital Sign (NVS) has been modified for use in different countries or language groups. The NVS is available in both English and Spanish versions as well as a validated UK-specific version, NVS-UK. Available from: http://www.pfizer.com/health/literacy/public_policy_researchers/nvs_toolkit
I agree with Rimmer. NVS is easy to translate and to adapt. You can analyse its possibilities for your country. In Brazil there is NVS in portuguese, named NVS-BR. Our group adapted a scale to measure nutrition literacy, an instrument developed by Diamond (2007), Nutritional Literacy Scale. We translated it and modified it according our cultural reality.