Trending Data Over Time: Run Charts

Metric data can be presented in multiple formats including tables, bar graphs, pie graphs, and run charts. Although tables of average performance may be the easiest to build, it may not be the most useful way of representing the data. Most people gain a better overall understanding of data shown in different graphical presentations. For data associated with changes over time, strongly consider using run charts. Run charts have several advantages over before-and-after summaries: it’s easier to see the effect of different aspects of your interventions on specific measures as they occur, you get a quicker picture of whether an intervention is working (although you have to be careful not to jump to conclusions too quickly with a single data point), it is easier to separate out the impact of your intervention from secular trends, and it is intuitively easier to interpret data graphically displayed as a run chart rather than as a table. Important components of run charts include: (1) well-labeled axes, with time usually on the X axis and performance on the Y axis; (2) a clearly marked line identifying the goal rate this is easy to see; And (3) points connected with a line that demonstrate actual performance over time. Time stamps, often using a vertical line, should highlight different components of the improvement project or external changes that may affect the metric of interest. In the example below, minutes to PCI is on the Y axis, and time is on the X axis. The green bar marks the target and the points, and the blue line designates performance.

Bar graphs may be most helpful when looking at data at a more refined level than the entire medical center. If, for instance, you want to see physician-to-physician variation, you may want to create bar graphs in which each physician of interest is represented by a bar. In these graphs, the Y axis should be performance and the X axis should be physician group. Each bar should have confidence intervals associated with it to better describe how precise the estimates are. It is helpful to use a horizontal line across all the bars to demonstrate average performance.