Participating sites included 9 public community hospitals or chronic care facilities, 9private community hospitals, 6 state agencies,2 state universities, and 1 private health organization. Worksites ranged in size from approximately 250 to 2800 employees. Worksites were blocked on size of the worksite and type of agency. We had 4 blocks: small health organizations (<800 employees and 1 building)9 sites; large health organizations (more than 800 employees or multiple buildings)—9 sites; state agencies—6 sites; university campuses 2 sites. Within each block, we used a random number generator to assign worksites to intervention so that half the worksites in each block were assigned to each condition. Because of the odd number of sites in 2 of the groups, therandomization algorithm also ensured that the total number of worksites was divided equally into intervention and control conditions.
การแปล กรุณารอสักครู่..
