Creating effective interrupted time series graphs: Review and recommendations

Abstract

Introduction: Interrupted Time Series (ITS) studies may be used to assess the impact of an interruption, such as an intervention or exposure. The data from such studies are particularly amenable to visual display and, when clearly depicted, can readily show the short- and long-term impact of an interruption. Further, well-constructed graphs allow data to be extracted using digitizing software, which can facilitate their inclusion in systematic reviews and meta-analyses.

Aim: We provide recommendations for graphing ITS data, examine the properties of plots presented in ITS studies, and provide examples employing our recommendations.

Methods and results: Graphing recommendations from seminal data visualization resources were adapted for use with ITS studies. The adapted recommendations cover plotting of data points, trend lines, interruptions, additional lines and general graph components. We assessed whether 217 graphs from recently published (2013-2017) ITS studies met our recommendations and found that 130 graphs (60%) had clearly distinct data points, 100 (46%) had trend lines, and 161 (74%) had a clearly defined interruption. Accurate data extraction (requiring distinct points that align with axis tick marks and labels that allow the points to be interpreted) was possible in only 72 (33%) graphs.

Conclusion: We found that many ITS graphs did not meet our recommendations and could be improved with simple changes. Our proposed recommendations aim to achieve greater standardization and improvement in the display of ITS data, and facilitate re-use of the data in systematic reviews and meta-analyses.

Keywords: data visualization; display of data; graph; interrupted time series; meta-analysis; systematic review.

FIGURE 1

Interrupted time series graphs showing key components. The data points are plotted, and the combination of trend lines and counterfactual allow visualization of the change in level from counterfactual to post‐intervention trend as well as the change in slope. The timing of the interruption is indicated by a vertical line in, A, and a shaded area in, B. The legend is included in the first example (A) but could easily be removed to increase the size of the graph if appropriate text was used in the caption of the graph (B) [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

Multiple interrupted time series in a single graph. The different series can be distinguished by different colors, symbols, or ideally both, A. Care must be given to not overly clutter the graph. The legend could be omitted with extra labels on the graph, as in, B, or with text in the figure caption [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

Interrupted time series graphs with additional lines. Neutral colors should be used, such as gray for the seasonality curves, A, and confidence intervals for the modeled trends, B. Note that with appropriate text in the figure caption, the legend can be removed, placing greater emphasis on the data points, B [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 4

Interrupted time series graphs with additional text. By removing the legend and restricting the scale of the y‐axis to the range of data, the space available to plot the data is maximized, A. Including text showing the level and slope change with confidence intervals in (A) enables the reader to quickly access the analysis results. The horizontal grid lines and y‐axis labels in (A) have been shown at the minimum, maximum and level change points. Leaving the box around the legend in, B, unnecessarily clutters the image. Rotating the text for the y‐axis labels from vertical (B) to horizontal (A) also aids in interpretation without unduly reducing the space available to plot the data [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 5

A, Published version of the graph. Reproduced with permission from JAMA Internal Medicine 2017;177(1):44‐50. doi: 10.1001/jamainternmed.2016.6811. Copyright©(2017) American Medical Association. All rights reserved. B, Revised graph using the proposed graphing recommendations [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE 6

A, Published version of the graph. Reprinted from The American Journal of Medicine, 2016;129(7):706‐14.e2, Hospital Readmissions Following Physician Call System Change: A Comparison of Concentrated and Distributed Schedules, Copyright©(2016) with permission from Elsevier. B, Revised graph using the proposed graphing recommendations [Colour figure can be viewed at wileyonlinelibrary.com]

FIGURE A1

terminology used for some components of graphs [Colour figure can be viewed at wileyonlinelibrary.com]