查看文章

Methods

A population-based survey for non-communicable diseases risk factors started in September 2010 by adopting the World Health Organization (WHO) STEPs Instrument (12), and the data collection was completed in March 2011. At that time the census of population in Kosova wasn’t conducted, therefore the data for households according to the settlements from Statistical Agency of Kosova for 2008 were used (13), in total seven regions, 30 municipalities and 1464 settlements. The two-stage cluster random sampling was designed. Firstly, 120 enumeration areas were selected using probability proportional to size as the primary sampling units, followed by randomly selecting households from them as the secondary sampling units, using the proportion of households in urban and rural areas. Respondents aged 15-64 years old were selected randomly within each gender and 10-year age-group. One resident aged 15 to 64 years within each of the households was recruited for the survey using the Kish method, which provides tool for random selection of one individual from a household (12). The total sample size consisted of 6,400 men and women. The following assumptions for this crosssectional study were used for sample size calculation: level of confidence 95%, margin of error 5%, baseline level of risk factors 50%, expected response rate 90% and the design effect of 1.5.