stracts, EconLit, Cochrane Library, Dissertation Abstracts on Disc, ABI, and NTIS (Devereaux et al, 2451). A review of their own files was also conducted, and direct contact was made with experts from June 2001 to January, 2002. Searches were also made on PubMed using the ôrelated articlesö feature, SciSearch and all studies that met their criteria for inclusion in their analyses.
The study criteria included published and unpublished observational studies and randomized trials which looked at the difference in mortality rates in private for-profit and private not-for-profit dialysis centers (Devereaux et al, 2451). The process identified 7045 citations on the subject, which were further narrowed to 779 full-text publications identified by pairs of screeners for full review. The screeners then evaluated the articles, in which the results had been masked, to determine if they fit the study criteria for examining health outcomes, appropriatness of quality care, and costs in private for-profit and not-for profit centers. Data abstracted from the studies included sampling methods, data source, type of facility (free-standing, hospital-based), dates of data collection, duration of patient followup, number of patients each type of facility evaluated, and mortality rates.
The study was adjusted for age, sex, race, educational level, income, cause of renal disease, and comorbid conditions (Devereaux et al, 2451). It also considered the number of years of dialysis treatment for each patient, the facilityÆs market share, and whether the facility was part of a multinational chain corporation. It looked at whether measures of comorbid factors, such as hypertension and other complications which could affect treatment outcomes, were taken before treatment began, and whether measures controlled by the facility administrators, such as the use of the anemia-fighting drug erythopoietin II, which could affect mortality rates and also pro...