Revista de Salud Pública
Print version ISSN 0124-0064
ZAPATA-OSSA, Helmer de Jesús et al. Cluster sampling in health surveys. Rev. salud pública [online]. 2011, vol.13, n.1, pp. 141-151. ISSN 0124-0064. http://dx.doi.org/10.1590/S0124-00642011000100012.
Objective Describing how the cluster sampling statistical technique can be applied to health surveys. Methods The article describes the most important aspects for each methodological step, emphasising cluster sampling's foundations. It was applied to a population of around 100,000 inhabitants from commune 18 in Cali. The sample covered 30 city blocks and 3,026 people were interviewed. The information was gathered using the Community-based Information System for Primary Health Attention. Results The sampling technique produced precise and reliable information for estimating population health event prevalence in a cross-sectional study. Sampling error was 2.5 % in the general population, 10% in children aged less than 5 years, 5.7% in children aged 5 to 14, around 3.0% in those aged 10 to 59 and 7.2% in the population aged more than 60. Conclusion The method guaranteed a 2.5% sampling error in estimating general population prevalence and less than 10% sampling errors for established population groups.
Keywords : Cluster sampling; health survey; prevalence; cross-sectional.