In Reply Dr Rahman and Mr Gilmour raise concerns regarding the hypertension prevalence estimates in the Prospective Urban Rural Epidemiology (PURE) study, particularly for Bangladesh. We cautioned in the article that the sampling framework in each country was not nationally representative and therefore may not be representative of each country. The sample size in low-income countries was 31 685 and the response rate was 55%.
The analyses were not weighted because population census data were not available from all communities. Sampling was not identical across communities because of variation in the availability of population lists used for sampling. Hence, for practical reasons, investigators in each country, in consultation with the project office, identified the best method of sampling to obtain a representative sample of households in a community.
Chow CK, Yusuf R, Kelishadi R. Prevention and Control of Hypertension in Different Countries—Reply. JAMA. 2014;311(4):419-420. doi:10.1001/jama.2013.285327