PT - JOURNAL ARTICLE AU - Tracey Timmerman AU - Tanya Verrall AU - Lisa Clatney AU - Helena Klomp AU - Gary Teare TI - Taking a closer look: using statistical process control to identify patterns of improvement in a quality-improvement collaborative AID - 10.1136/qshc.2008.029025 DP - 2010 Dec 01 TA - Quality and Safety in Health Care PG - e19--e19 VI - 19 IP - 6 4099 - http://qualitysafety.bmj.com/content/19/6/e19.short 4100 - http://qualitysafety.bmj.com/content/19/6/e19.full SO - Qual Saf Health Care2010 Dec 01; 19 AB - Background Published reports suggest that there is considerable variation in improvement capacity and capability among participants in quality improvement collaboratives. Generating knowledge about why these complex initiatives do or do not work in different contexts requires both qualitative and quantitative approaches. Time-series analysis using line graphs and statistical process control is a rigorous quantitative approach with relatively unexplored potential in evaluating complex quality improvement interventions.Aim The purpose of this study was to apply and illustrate the use of line graphs and statistical process control to identify variation in improvement among practices participating in the Saskatchewan Chronic Disease Management Collaborative.Methods The authors used line graphs and regression analysis to determine whether improvement occurred at the aggregate level, and small multiples, rational ordering and rational subgrouping to examine differences in the level and rate of improvement among practices.Results Small multiples allowed us to observe qualitative differences in patterns of improvement among practices. Stratifying data for all subgroups on one control chart using rational subgrouping provided quantitative evidence for these differences. Stratification by administrative health region using rational ordering showed consistent differences, indicating that the health region in which the practice was located may play a role in these differences. Results from the study are being used to inform a purposive sampling strategy for interviews with participants to explore why these differences occurred.